WeiDU DocumentationValerio Bigiani, AKA The Bigg |
The main home page for WeiDU is: http://weidu.org/. I encourage you to download the latest version.
WeiDU is designed to make it easier to write and distribute modifications to Infinity Engine games. It can load and modify Infinity Engine resources according to instructions you provide. WeiDU is ideal for packaging modifications that include dialogue or that want to be compatible with other modifications.
I’ll be honest with you up front: WeiDU is initially harder to use than some of its alternatives. However, most users report that (1) the alternatives are insufficient because they lack features that only WeiDU provides and (2) WeiDU grows on you over time.
You are welcome to use these utilities to make and distribute your own Infinity Engine mods. This utility is covered by the GNU General Public License, but you are also allowed to distribute an unmodified binary copy of WeiDU.EXE (without the source code) with your mod if you like.
I decided to write my own Infinity Engine DLG and TLK utilities because I was unable to get the TeamBG DLG Editor and Mass Converter to work properly. Either they wouldn’t parse the strings or they would mangle the text or they would randomly crash ... it was bad all around. Also, they were all GUIs. As a unix weenie I’m in love with command line utilities and as a PL doctoral student I love making little languages and compilers. WeiDU was originally a family of small programs with unimaginative names like DC, DD and TP. The more appealing term “WeiDU” (which rhymes with “IDU”, Eye-Dee-You) was coined by Jason Compton and Ghreyfain, noted BGII mod authors.
Step-By-Step Beginner’s Guide to WeiDU:
This section is a gentle introduction to how Infinity Engine DLG files are structured. First, let’s use WeiDU to create SCSARLES.D and take a look at the dialogue of Sir Sarles.
You may install WeiDU.exe anywhere on your system. However, I recommend that you put it in your Baldur’s Gate 2 installation directory. However, WeiDU will use the Windows Registry to attempt to locate your BG2 game files.
To run the effect described, open up a DOS prompt window and change directories to get to your BGII directory. Then just type in the text in red at the DOS Prompt.
C:\Program Files\Black Isle\BGII - SoA\> weidu SCSARLES.DLG
This will create a text file called SCSARLES.D in the current directory. Open it up with Notepad or Microsoft Word or something. It’s just a text file that describes the game dialogue.
It will look something like:
// creator : c:\bgate\weidu\weidu.exe // argument : SCSARLES.DLG // game : C:\Program Files\Black Isle\BGII - SoA // source : C:\Program Files\Black Isle\BGII - SoA\data\Dialog.bif // dialog : C:\Program Files\Black Isle\BGII - SoA\DIALOG.TLK // dialogF : (none) BEGIN ~SCSARLES~ IF ~NumTimesTalkedTo(0)~ THEN BEGIN 0 // from: SAY #28655 /* ~Who is it? Might I ask why you have disturbed my meditations? My creative muse must be gently awakened, and your stomping about is simply not conducive to this.~ [SARLES02] */ IF ~~ THEN REPLY #28656 /* ~My apologies. I will leave you to your thinking.~ */ GOTO 1 IF ~~ THEN REPLY #28657 /* ~I apologize, but I have come to request your talent on a commissioned artwork.~ */ DO ~SetGlobal("TalkedToSarles","GLOBAL",1)~ GOTO 2 END IF ~~ THEN BEGIN 1 // from: 0.0 SAY #28661 /* ~Then I shall forget you were ever here. Actually, it is an astoundingly easy thing to do.~ */ IF ~~ THEN DO ~SetNumTimesTalkedTo(0)~ EXIT ENDDialogues in Infinity Engine games behave like finite state machines. If you aren’t familiar with the concept of a finite state machine, see http://whatis.techtarget.com/definition/0,,sid9_gci213052,00.html or http://www.c3.lanl.gov/mega-math/workbk/machine/mabkgd.html. Each block of the form:
IF ~Initial Condition~ THEN BEGIN state1 SAY ~Something~ IF ~Reply Condition~ THEN REPLY ~Reply Text~ GOTO state2 ENDrepresents a state (more details below). When the player starts a conversation with an NPC, the game engine scans through all of the states in that NPC’s DLG file in a special WEIGHTed order and picks the one with a non-empty and true "Initial Condition". If no state has a non-empty and true "Initial Condition" then you get that “Bob - has nothing to say to you.” message. Don’t worry about the weighting process for now.
The speaker (in this case, Sir Sarles) then says whatever appears after SAY. The REPLY lines represent responses the PC can say back. If the "Reply Condition" is true, the player is given the option of saying the "Reply Text" and moving to another state in the dialogue (where Sarles will probably say something else).
Remember: SAY is for what the NPC says, REPLY is for what the player says back. If you think carefully, you’ll notice that all dialogue in Infinity Engine games is structed in this manner.
Conditions use the same syntax as triggers do in Infinity Engine BCS scripting. You will need to learn Infinity Engine scripting before too long. Strings are delineated by tildes or %% or "" (your choice, but WeiDU uses the tilde by default). After SAY or REPLY or JOURNAL you may give two Strings instead of one. The first is used with DIALOG.TLK, the second is used with DIALOGF.TLK (foreign language version for when the main character is female). If you do not give two Strings, the one String you gave is used for both.
You may also use raw numbers prefaced with a number sign (like #1234) to specify a strref inside DIALOG.TLK directly. This is useful when modifying existing dialogues (say, the Fate Spirit in ToB) so that you if a foreign user installs your dialogue they will retain all of the foreign versions of the strings you didn’t change or add. Normally the string reference numbers are put right after the SAY keyword and the string text is put in comments. The --text command-line option causes string text to be emitted with the string number in comments.
You may also indicate that a sound file (WAV/WAVC) should be associated with a given String by including its up-to-8-letter resource name in [brackets] after the string, as in:
SAY ~Hello~ [HELLO]Comments are C/C++ style: everything from // to the end of the line is a comment, as is /* everything in these star-slash things */. Comments are ignored by WeiDU. They are there for your benefit. Example:
SAY ~Hello~ [HELLO] // this is a comment ... way out to here! IF /* this is also a comment */ ~~ THEN EXITReplies can also contain actions (using the DO keyword) which behave just like Infinity Engine BCS script actions. They can also add Journal entries, end the dialogue or transfer to another speaker.
Examples:
IF ~~ THEN BEGIN 2 // from: 0.1This line marks the beginning of state 2 in a dialogue. The comment tells you that it can be reached by the first reply transition from state 0.
IF ~~ THEN REPLY ~My apologies. I will leave you to your thinking.~ // #28656 GOTO 1This REPLY can always be chosen and involves the spoken text "My apologies...". That text is string reference number 28656. If the PC chooses that reply, it transitions to state 1.
Finally, a transition may also take the form:
COPY_TRANS filename labelDuring processing, COPY_TRANS will be replaced by all of the transitions from state "label" of file "filename". The copying takes place before all other D actions.
The D file format is a way of describing Infinity Engine dialogues and modifications to Infinity Engine Dialogues in a portable, easy-to-understand format. It supports foreign language translations and allows you to describe extensions to existing game dialogues without forcing you to describe their content. This allows you to write mods that work with mods written by others.
The D file format is presented here in an extended context-free grammar notation. If you are unfamiliar with CFGs, take a look http://www.wikipedia.com/wiki/Context-free_grammar, http://cs.wpi.edu/~kal/PLT/PLT2.1.2.html or http://www.cs.rochester.edu/users/faculty/nelson/courses/csc_173/grammars/cfg.html. You don’t really need to understand a CFG formally, though.
To get a real idea of how they work, use WeiDU to create JAHEIRA.D for yourself and look at it in a text editor. You can also browse the examples and test directories that come with WeiDU.
All of the syntax keywords are given in a UPPERCASE COURIER. All other keywords are symbolic. Notes:
D File | A D file is a text file that contains a number of D Actions. D Files tell WeiDU how to create and modify Infinity Engine DLG files. | |
is | D Action list | A D File is a list of D Actions. Typically the first and only one is BEGIN, which defines the content of a new dialogue. Other D Actions can be used to modify existing dialogues. |
D Action | A D Action tells WeiDU how to create or modify Infinity Engine DLG files. | |
is | BEGIN filename [ nonPausing ] state list | BEGIN tells WeiDU that you are creating a new DLG file from scratch. Any existing DLG file with the same name will be overwriten. The new DLG file contains exactly the states in the list. If you set nonPausing to a non-zero integer, the game will not “stop time” while the conversation takes place. By default time stops during conversations. |
or | APPEND [ IF_FILE_EXISTS ] filename state list END | This tells WeiDU to place the given states at the end of the already-existing dialogue filename.DLG. If there is IF_FILE_EXISTS and the file doesn’t exists, this action is skipped. |
or | APPEND_EARLY [ IF_FILE_EXISTS ] filename state list END | Works like APPEND, but the states are added early on in the compilation timeline (just after BEGIN is processed). Thus they can be the targets for INTERJECT_COPY_TRANS and friends. |
or | CHAIN [ IF [ WEIGHT #weight ] stateTriggerString THEN ] [ IF_FILE_EXISTS ] entryFilename entryLabel chainText list chainEpilogue | This instructs WeiDU to make a long conversation in which the PC can say nothing. This is useful when you want the NPCs to talk among themselves for a long time. It and its friends, INTERJECT and INTERJECT_COPY_TRANS can incredible time-savers when you’re writing non-trivial dialogue. See the examples for ideas. CHAIN will only append to existing dialogues. You cannot use CHAIN to create a new DLG. If the entryFilename file is missing and there is IF_FILE_EXISTS, the whole CHAIN is not compiled. |
or | INTERJECT entryFilename entryLabel globalVariable chainText list chainEpilogue | Behaves like CHAIN except that all of the chainText is additionally guarded by the transition predicate Global("globalVariable","GLOBAL",0) and accompanied by the action SetGlobal("globalVariable","GLOBAL",1). If you pick globalVariable to be unique, this will ensure that the chainText is only ever seen once per game. This is useful for making interjections. |
or | INTERJECT_COPY_TRANS entryFilename entryLabel globalVariable chainText list | This behaves just like INTERJECT except that the exitFilename and exitLabel are not present. Instead, whenever the dialogue would pass out of the chainText it follows a copy of the transitions that were at the state with stateLabel originally. This is convenient for making quick interjections from NPCs that do not actually change the true flow of the conversation. See the transition COPY_TRANS and the INTERJECT_COPY_TRANS tutorial for more information about this idea. |
or | INTERJECT_COPY_TRANS2 entryFilename entryLabel globalVariable chainText list | This works just like INTERJECT_COPY_TRANS, except that any actions taken in the transitions of the state specified by entryFilename and entryLabel are preserved and kept with the speaker associated with entryFilename (rather than being mistakenly performed by your new speaker). See the INTERJECT_COPY_TRANS2 tutorial. |
or | INTERJECT_COPY_TRANS3 entryFilename entryLabel globalVariable chainText list | This works just like INTERJECT_COPY_TRANS, except that all states in chainText get a link in the entry state, rather than only the first one. Expecting documentation. |
or | INTERJECT_COPY_TRANS4 entryFilename entryLabel globalVariable chainText list | This is either INTERJECT_COPY_TRANS3 with INTERJECT_COPY_TRANS2-style action handling, or INTERJECT_COPY_TRANS2 with the extended INTERJECT_COPY_TRANS3 state creation rules, depending at how you look at it. |
or | EXTEND_TOP filename stateLabel list [ #positionNumber ] transition list END | This instructs WeiDU to add the transitions in list to the top of
the transition list for the specified states in filename.DLG
(which must already exist). If a positionNumber is given, WeiDU to insert the transitions just between already-existing transitions #positionNumber and #positionNumber+1 in the given states for the given file. The first transition is number 1. |
or | EXTEND_BOTTOM filename stateNumber list [ #positionNumber ] transition list END | Behaves just like EXTEND_TOP but adds the transitions to the bottom of the list instead. |
or | ADD_STATE_TRIGGER filename stateNumber stateTriggerString [ stateNumber list ] | This instructs WeiDU to add the stateTriggerString to all of the states with the given stateNumbers in the file filename.DLG (which must already exist). This is handy for adding extra conditions to an existing dialogue state. |
or | ADD_TRANS_TRIGGER filename stateNumber transTriggerString [ moreStateNumbers list ] [ DO transNumber list ] | This instructs WeiDU to add the transTriggerString to all of the transitions in all of the states with the given stateNumbers in the file filename.DLG (which must already exist). This is often used in conjunction with EXTEND_BOTTOM to make a new branch in an existing state. Use ADD_TRANS_TRIGGER to add the negation of some predicate to all of the existing transitions, then use EXTEND_BOTTOM to add a transition with that predicate to that state. If a list of transNumbers is specified, only those transitions will have transTriggerString added to them. If such a list is not specified, every transition in every specified state will be modified. Note that the “first” transition is number 0. |
or | ADD_TRANS_ACTION filename BEGIN stateNumber list END BEGIN transNumber list END transActionString | This instructs WeiDU to add the transActionString to all of the actions in all the transitions in all of the states specified by the stateNumber list and the transNumber list. You may use state labels in the stateNumber list. If the transNumber list is empty, the text added to all transitions on all listed states. Note that the BEGIN and END keywords must be present, even if you specify an empty list of transNumbers. The “first” transition is number 0. Any out-of-bounds transNumbers are silently ignored. The transActionString is prepended to any existing action text on a per-transition, per-state basis. |
or | REPLACE_TRANS_ACTION filename BEGIN stateNumber list END BEGIN transNumber list END oldText newText | This instructs WeiDU to replace all instances of oldText in newText to all of the actions in all the transitions in all of the states specified by the stateNumber list and the transNumber list. You may use state labels in the stateNumber list. If the transNumber list is empty, the text added to all transitions on all listed states. Note that the BEGIN and END keywords must be present, even if you specify an empty list of transNumbers. The “first” transition is number 0. Any out-of-bounds transNumbers are silently ignored. |
or | REPLACE_TRANS_TRIGGER filename BEGIN stateNumber list END BEGIN transNumber list END oldText newText | This instructs WeiDU to replace all instances of oldText in newText to all of the triggers in all of the transitions in all of the states specified by the stateNumber list and the transNumber list. You may use state labels in the stateNumber list. If the transNumber list is empty, the text added to all transitions on all listed states. Note that the BEGIN and END keywords must be present, even if you specify an empty list of transNumbers. The “first” transition is number 0. Any out-of-bounds transNumbers are silently ignored. |
or | ALTER_TRANS filename BEGIN stateNumber list END BEGIN transNumber list END BEGIN changeWhat changeInto list END | Fine altering of a single transition. See the ALTER_TRANS tutorial. |
or | REPLACE filename state list END | This instructs WeiDU to load filename.DLG and replace some of its states with the new ones described in the state list. All of the states should have numeric stateLabels (e.g., "5" or "67"). A new state with label X will replace the old state number X. |
or | SET_WEIGHT filename stateLabel #stateWeight | This instructcs WeiDU to destructively change the WEIGHT of the given state in filename.DLG (which must exist). This should only be used to patch or workaround existing dialogues. Never use SET_WEIGHT if you can help it. |
or | REPLACE_SAY filename stateLabel sayString | This instructs WeiDU to destructively change the sayString of the given state in filename.DLG (which must exist). This should only be used to patch or workaround existing dialogues. Never use REPLACE_SAY if you can help it. |
or | REPLACE_STATE_TRIGGER filename stateNumber stateTriggerString [ stateNumber list ] | This instructs WeiDU to destructively set the stateTriggerString of all of the states with the given stateNumbers in the file filename.DLG (which must already exist). It should be used with caution. |
or | REPLACE_TRIGGER_TEXT filename oldText newText | This instructs WeiDU to destructively replace every occurrence of oldText (which may be a regexp) in the stateTriggerStrings and transTriggerStrings of filename.DLG (which must exist). This should only be used to patch or workaround existing dialogues. Never use this if you can help it. |
or | REPLACE_TRIGGER_TEXT_REGEXP filenameRegexp oldText newText | Just like REPLACE_TRIGGER_TEXT but the filename is a regexp. The .DLG is implied. Do not use this. |
or | REPLACE_ACTION_TEXT filename oldText newText [ moreFilenames ] | This instructs WeiDU to destructively replace every occurrence of oldText (which may be a regexp) in the stateActionStrings of filename.DLG (which must exist). This should only be used to patch or workaround existing dialogues. Never use this if you can help it. |
or | REPLACE_ACTION_TEXT_REGEXP filenameRegexp oldText newText [ moreFilenameRegexps ] | Just like REPLACE_ACTION_TEXT but the filenames are regexps. The .DLG is implied, do not include it in your regexps. Do not use this. |
or | REPLACE_ACTION_TEXT_PROCESS filename oldText newText [ moreFilenames ] | This instructs WeiDU to destruveily replace every occurrence of oldText
(which may be a regexp) in the stateActionStrings
of filename.DLG (which must exist) with newText. However,
newText is first compiled as a BAF action list. In particular,
this means that replacing with commands like:
~DisplayString(Myself,@123)~... will do what you expect. This should only be used to patch or workaround existing dialogues. Never use this if you can help it. |
or | R_A_T_P_R filenameRegexp oldText newText [ moreFilenameRegexps ] | Just like REPLACE_ACTION_TEXT_PROCESS, but the filenames are regexps. The .DLG is implied. Do not use this. R_A_T_P_R is shortand for REPLACE_ACTION_TEXT_PROCESS_REGEXP (to avoid undue scrollbars in the readme on a 1024x768 monitor). |
chainEpilogue | Determines where the dialogue should flow at the end of the CHAIN. | |
is | END filename stateNumber | Transfer to the given state in the given dialogue file. |
or | EXTERN filename stateNumber | Transfer to the given state in the given dialogue file. |
or | COPY_TRANS filename stateNumber | At the end of the CHAIN text, copy all transitions from the given state in the given file. This is useful for interjections (see INTERJECT). |
or | COPY_TRANS_LATE filename stateNumber | Same as above, but only copy after all the other D Actions have been applied. |
or | EXIT | At the end of the CHAIN text, exit the dialogue. |
or | END transition list | Execute the given transitions after the final state in the CHAIN. |
state | In Infinity Engine games, this is the fundamental unit of dialogue. | |
is | IF [ WEIGHT #weightNumber ] stateTriggerString [ THEN ] [ BEGIN ] stateLabel SAY sayText [ = sayText ... ] transition list END | When you start conversing with a creature that uses a DLG file, the
Infinity Engine searches through all of the states in that file
in order of increasing WEIGHT and selects the first one it finds
for which the stateTriggerString is both true and not empty.
The creature then says all of the associated sayText. Finally,
the transitions are evaluted in bottom-up (i.e., reverse) order.
If a transition is found with a transTriggerString that
evaluates to True and no replyText, that transition is
immediately executed. Otherwise, all of the transitions are
presented as options to the PC. If a stateLabel is an integer it is called a stateNumber. All of the states in the DLG files that come with the original game use stateNumbers. Only D files use symbolic strings for stateLabels. Including more than one bit of sayText here is often called Multisay. Finally, once you are familiar with the syntax you may omit the THEN and BEGIN keywords if you like. |
or | APPENDI filename state list END | This is legacy syntax that behaves just like the D Action APPEND but is considered a state. Avoid it. |
or | CHAIN2 entryFilename entryLabel chain2Text list exitFilename exitLabel | This is legacy syntax that behaves somewhat like the D Action CHAIN but is considered a state. In addition, chain2Text is slightly different from chainText. Avoid this construction. |
sayText | sayText and replyText are displayed to the user as part of a dialogue. | |
is | text | sayText and replyText are both text. |
transition | Transitions determine how dialogue flows from one state to another. | |
is | IF transTriggerString [ THEN ] transFeature list transNext | If the transTriggerString evaluates to true or is empty, this transition is viable. If it contains no replyText within its transFeature list, it is immediately taken. Otherwise, the replyText is presented as an option to the user. If the transition is taken, any actions in the transFeature list are performed and the dialogue flows to the point indicated by the transNext. transitions are evaluated in "reverse order". That is, the "bottom" or "last" response for a state is checked first. If its transTriggerString evaluates to true and it has no REPLY text, that transition is immediately taken. See SAREV25A state 1 for an example of a state with all kinds of transitions. |
or | + [ transTriggerString ] + replyText transFeature list transNext | This abbreviated syntax for transitions that would contain REPLY (which is by far the most common case) allows you to save yourself some time and typing. It behaves like the full form above. |
or | COPY_TRANS filename stateLabel | This instructs WeiDU to copy all of the transitions from the
state with the given stateLabel in filename.DLG. This
copying takes place before all other D Actions. For example,
this is a valid transition list:
IF ~Before()~ THEN GOTO my_state COPY_TRANS PLAYER1 33 IF ~After()~ THEN EXTERN SOLA 55 |
or | COPY_TRANS_LATE filename stateLabel | Same as COPY_TRANS, but is invoked after all other D Actions. |
transFeature | These are features or actions associated with taking a transition. | |
is | REPLY replyText | If this transition is taken, the PC says the replyText. |
or | DO stateActionString | If this transition is taken, the stateActionString is executed. |
or | JOURNAL text | If this transition is taken, the text is added to the PC’s journal. |
or | SOLVED_JOURNAL text | If this transition is taken, the text is added to the “solved” section of the PC’s journal. |
or | UNSOLVED_JOURNAL text | If this transition is taken, the text is added to the “unsolved” section of the PC’s journal. |
or | FLAGS integer | This allows you to set the features associated with a transition directly using the binary format of DLG files. Do not use this! |
transNext | This determines where dialogue flows after a transition has been taken. | |
is | GOTO stateLabel | The dialogue continues at the state with label stateLabel in the same DLG file as the current state. |
or | EXTERN [ IF_FILE_EXISTS ] filename stateLabel | The dialogue continues at the state with label stateLabel in the file filename.DLG. The whole transition is not compiled if there’s IF_FILE_EXISTS and the file filename doesn’t exist. |
or | EXIT | The conversation ends. |
or | + stateLabel | This is a synonym for GOTO. |
chainText | This is a rapid shorthand for chaining together many little bits of dialogue when the PC is not saying anything. | |
is | [ IF transTriggerString THEN ] sayText = sayText ... | |
followed by | [ == [ IF_FILE_EXISTS ] fileName [ IF transTriggerString THEN ] sayText = sayText ... ] | The == (that’s two consecutive equal signs) marks the beginning of a new speaker (indicated by fileName). If the transTriggerString is true or if it is not present, this new speaker says all of its sayText in order. If the IF_FILE_EXISTS part is present, these lines are not compiled at all if the current file is not missing. |
or followed by | BRANCH transTriggerString BEGIN [ == [ IF_FILE_EXISTS ] fileName [ IF transTriggerString THEN ] sayText = sayText ... ] END | As above, except that the first transTriggerString is appended to all existing dialogue units. |
text | This represents strings that are shown to the player, rather than strings that the game uses internally for predicates and actions. | |
is | String [ [WAVEFILE] ] | The given string is used for both male and female players. The optional [WAVEFILE] is the associated sound. |
or | String [ [WAVEFILE] ] String [ [WAVEFILE] ] | The first string and sound file are used if the PC is male, the second string and sound file are used if the PC is female. This is useful mainly for international versions of Infinity Engine games. |
or | #integer | The string with reference number #integer from DIALOG.TLK should be used unchanged. |
or | @integer | The last definition of the translation string @integer given in any TRA file should be used. |
or | !integer text | Forced String Reference. As with text in general, but rather than being assigned a new, previously-unused DIALOG.TLK string entry (or merging with an existing one that has the same text), this text is written over DIALOG.TLK string entry #integer. Do not use this feature. |
String | This is how you tell WeiDU what text you want shown to the player. For international mods or international translations, you may use any encoding you like (that is, you are not restricted to 7-bit characters or Latin-1 or anything like that). | |
is | "abcdef" | A string can be any sequence of characters not including a " that is enclosed in ""s. |
or | ~ abcdef~ | A string can be any sequence of
characters not including a ~ that is enclosed in
~~ s. |
or | %abcdef% | A string can be any sequence of characters not
including a % that is enclosed in %%s. This is handy for Big5
translations, since " and ~ can be part of Big5-encoded characters. |
or | ~~~~~ abcdef~~~~~ | That’s five consequtive
tildes on each side. A string can be any
sequence of characters not including ~~~~~ that is enclosed in
~~~~~ s. For example,
string #8750 is
~!@#$\%^&*()_+-=[]{}\|;:'",<.>/?
and can be given to WeiDU as
~~~~~~!@#$\%^&*()_+-=[]{}\|;:'",<.>/?~~~~~
(the content of the string is shown in red for clarity). |
or | String ^ String | String literal concatenation. The second string is appended to the first string. No whitespace is added. Thus "hello" ^ "World" is the same as "helloWorld". |
WeiDU is a command-line utility. GUIs are available, but this document only describes command-line invocation. Use the DOS Shell ("command" or "cmd") to run WeiDU. You control its behavior by passing arguments to it on the command line.
You invoke WeiDU by typing WeiDU and then any number of options and files, as described below.
Note: since it may not be obvious, if an option accepts ’X’ and says to be cumulative, if you want to apply it multiple times you have to use the following hideousness:
weidu --string 1 --string 2 --string 3
This does not apply to *-rest commands like --biff-get-rest and --force-install-rest.
Moreover, *-rest commands must be the last, EG
weidu --biff-get-rest sw1h01.itm --out foo
will try to extract sw1h01.itm, --out and foo from the biffs; you have to express the above as
weidu --out foo --biff-get-rest sw1h01.itm
*-list commands work like the -rest variants, except that they stop parsing for the current switch once they find an option starting with ’-’. Basically, this works like you’d expect:
weidu --biff-get-list sw1h01.itm --out foo
Compiling And Decompiling | |
FILE.D | Compile FILE to a DLG (dialogue file). |
FILE.DLG | Decompile FILE to a D (dialogue text file). |
number | When decompiling a DLG file to a D file, emit only state number. You may specify this multiple times and only the states you specify will be emitted. |
numberA-numberB | When decompiling a DLG file to a D file, emit only states between numberA and numberB (inclusive). You may specify this multiple times and only the states you specify will be emitted. |
FILE.BAF | Compile FILE to a BCS (script file). |
FILE.BCS | Decompile FILE to a BAF (script text file). |
--script-style X | Use the given BAF/BCS scripting style. X must be BG or BG1 or BG2 or PST or IWD or IWD1 or IWD2. |
--transin X | Use FILE as a source of translation strings when processing D and BAF files. |
FILE.TRA | Equivalent to --transin FILE.TRA. |
Module Packaging Input And Control | |
FILE.TP or FILE.TP2 | Read FILE and ask the user whether to install, reinstall or uninstall its TP2 Components. |
--yes | Answer all TP2 questions with ’Yes’ and do not prompt for a key to be pressed at the end of TP2 processing. |
--uninstall | Answer all TP2 questions with ’Uninstall’ and do not prompt for a key to be pressed at the end of TP2 processing. |
--reinstall | Re-install all TP2 components that are already installed and do not promtp for a key to be pressed at the end of TP2 processing. |
--languagem X | Sets the TP2 Language to the one passed here. Has no effect if the value is bigger than the language count (I.E. you’ll get asked for the language, unless you gave --uninstall, --reinstall or --force-uninstall). |
--force-install X | installs component number X, skips the others (cumulative). |
--force-uninstall X | uninstalls component number X, skips the others (cumulative). If there is no --force-install, you don’t get asked for the language. |
--force-install-rest X Y... | installs component number X Y..., skips the others (cumulative). |
--force-uninstall-rest X Y... | uninstalls component number X Y..., skips the others (cumulative). |
--force-install-list X Y... | installs component number X Y..., skips the others (cumulative). |
--force-uninstall-list X Y... | uninstalls component number X Y..., skips the others (cumulative). |
--skip-at-view | AT_* VIEW this actions (and the extra chromosome versions like NOTEPAD this ) aren’t processed, while still processing batch files and similia. |
--safe-exit | Save weidu.log every time a component installation is begun. This makes it impossible to uninstall components (don’t ask), but allows the user to kill the weidu process (E.G. closing the DOS console via the X button) without leaving the game in an unrecoverably inconsistent state (you can [R]e or [U]ninstall the component and the game will be consistent again). This might break with mods using overly complicated AT_* actions (basically don’t use this with Big Mods). |
--save-components-name | Reprints weidu.log. Useful if you install with --quick-log. |
--ask-every | Behave as if ASK_EVERY_COMPONENT were present for all TP2 components. |
--continue | Continue TP2 processing despite TP2 Action errors. |
--args X | X will be stored in the tp2 variable %argv[n]% , where
n is 0 for the first argument, 1 for the second, etc. If the installation is non-interactive, said variables will be loaded from the last interactive session. |
--args-rest X Y... | X Y Z will be stored in the tp2 variables %argv[n]% ,
where n is 0 for the first argument, 1 for the second, etc.If the installation is non-interactive, said variables will be loaded from the last interactive session. |
--args-list X Y... | X Y Z will be stored in the tp2 variables %argv[n]% ,
where n is 0 for the first argument, 1 for the second, etc.If the installation is non-interactive, said variables will be loaded from the last interactive session. |
--debug-assign | Print out all values assigned to TP2 variables (even implicit ones created by WeiDU). |
--debug-value | Print out all values encountered in TP2 processing and the results they evaluate to. Among other things, this is useful for catching parenthesis errors in your values. |
--clear-memory | calls CLEAR_MEMORY after every TP2 action. |
Automatic Updating Options | |
--update-all | Auto-update all WeiDU setup files (e.g., Setup-MyMod.exe) in the current directory. |
--noautoupdate | If you are running WeiDU as Setup-MyMod.exe, do not attempt to update yourself or any other mod. |
--noselfupdatemsg | If you are running WeiDU as Setup-MyMod.exe and it automatically updates itself, do not display a message or ask the user to press return. |
Infinity Engine Game Location Options | |
--game X | Set main game directory to X. WeiDU looks for CHITIN.KEY and DIALOG.TLK and the override directory in the main game directory (but see --tlkin and --search). WeiDU will look in the current directory and use the registry to find your game. If this fails, you will need to run WeiDU using the --game switch to define the full path to the BG2 directory. WeiDU will also search for BG1, IWD and PST. |
--nogame | Do not load any default game files. Unless you also specified --tlkin, no DIALOG.TLK will be loaded. Unless you also specified --search, no override directory will be used. |
--search X | Look in X for input files (cumulative). X is treated as an override directory and is given priority over the default override directory. |
Game Text (TLK) File Input | |
--tlkin X | Use X as DIALOG.TLK (instead of looking for DIALOG.TLK in the game directory). |
--ftlkin X | Use X as DIALOGF.TLK (instead of looking for DIALOGF.TLK in the game directory). |
FILE.TLK | Equivalent to --tlkin X. |
--tlkmerge X | Merge strings from X over the strings from any other loaded DIALOG.TLK. |
General Output Options | |
--out X | Emit most output files generated by command-line options (e.g., D, DLG, kits, --biff-get, BAF, BCS, --automate, --traify-tlk, --extract-kits, --list-biff, --cmp-from, --dcmp-from, etc.) to file X. If X is a directory, certain commands (e.g., D, DLG, --biff-get, etc.) will place their output there. Does not affect TP2 processing. |
--append X | Like --out, but if X is an existing file then the result will be appended to it instead of overwriting it. |
--backup X | Backup files to directory X before overwriting. Does not affect TP2 processing. |
--tlkout X | If any strings were added to or changed in the loaded DIALOG.TLK, emit X as an updated version that reflects those changes. Many operations (e.g., compiling D files, --tlkmerge, STRING_SET) can add to or modify DIALOG.TLK. |
--ftlkout X | If any strings were added to or changed in the loaded DIALOGF.TLK, emit X as an updated version that reflects those changes. |
Dialogue Text File (D) Options | |
--noheader | Do not emit D header comments. |
--nofrom | Do not emit D // from: comments. |
--full-from | Generate complete // from: comments with a slower two-pass process. |
--nocom | Do not emit ANY D or BAF comments. |
--text | Emit string text with string references in comments. |
--transitive | Follow EXTERN links when making D files. See the tutorial on --transitive. |
--toplevel | Emit only top-level dialogue states -- that is, states with non-empty triggers. |
Translation Options | |
--traify X | Convert X (which should be a D or TP2 or BAF) so that it uses translation references instead of literal strings. Use --out Y to specify a name for the transformed version of X and its new TRA file. |
--traify# X | Use with --traify and --traify-tlk. Start the newly-created TRA file at translation string @X instead of @0. |
--traify-old-tra X | Assumes file X to contain strings already in .tra format. Use this over --traify# for merging new strings in an already traified D or TP2 or BAF file. |
--untraify-d X | convert .D file X to use hardcoded strings... |
--untraify-tra X | ...from TRA file X. Please note that the outcoming file
could not work properly (if baf code embedded in a d file contains @ references, or if a string contains
the~ character). |
--trans | Emit coupled D and TRA files when decompiling a DLG. |
--transref | Emit string reference numbers in TRA files when using --trans. |
--traify-tlk | Emit a TRA file for the loaded TLK file (see --tlkin, --out, --min and --traify#). |
--make-tlk X | Create a TLK file from TRA file X (cumulative, see --tlkout). |
--testtrans | Test all specified TRA translation files to see if any of them use text that is already in the loaded DIALOG.TLK. If they do you can save translation effort by using string references instead. |
--forceify X | Convert the given D file to use forced strrefs (see --out, SAY, Forced String Reference). |
Game Text Repository (TLK) Options | |
--string X | Display string reference #X (cumulative). If you also specify --min or --max, all string references between --min (or 0) and --max (or infinity) will be displayed. |
--strfind X | Display strings that contain X (cumulative, regexp allowed). |
--strapp X | Append string X to DIALOG.TLK (cumulative). |
Game Archive (BIFF) Options | |
--list-biffs | Enumerate all BIFF files in CHITIN.KEY. |
--list-files | Enumerate all resource files in CHITIN.KEY |
--biff X | Enumerate contents of BIFF file X (cumulative). |
--biff-get X | Extract resource X from game BIFFs (cumulative, regexp allowed). |
--biff-get-rest X Y ... | Every argument given on the command line after --biff-get-rest is treated as if it were preceeded by --biff-get. Use this command to extract multiple different files (or regexps) at once. |
--biff-get-list X Y ... | Every argument given on the command line after --biff-get-rest is treated as if it were preceeded by --biff-get, with the -list exception. Use this command to extract multiple different files (or regexps) at once. |
--biff-str X | Search all game BIFFs for files containing X (regexp allowed). |
--biff-value X | Search all game BIFFs for files containing value X at offset ADDR. Must be used with --biff-value-at. |
--biff-value-at ADDR | Gives the offset address for a --biff-value search. |
--biff-type X | Limit --biff-str or --biff-value searches to resources of type X (cumulative). |
--biff-name X | When a --biff-str or --biff-value search finds a matching file, assume it has a strref name at offset X and print that name out as well. |
--make-biff X | Create data/X.bif from all files in folder X and destructively update CHITIN.KEY. Do not use this feature. Do not even think of using this feature without backing up CHITIN.KEY. |
--remove-biff X | Remove references to BIFF X and all of its resources from CHITIN.KEY. Do not use this feature. |
Comparison Options (see --out) | |
--cmp-from X | Emit WRITE_BYTEs to turn this file ... |
--cmp-to X | ... into this one. |
--dcmp-fromX | Emit REPLACEs to turn this DLG file ... |
--dcmp-to X | ... into this one. |
--tcmp-fromX | Compare this TRA file (or directory of TRA files)... |
--tcmp-to X | ... with this one (or this directory). |
--tlkcmp-fromX | Emit STRING_SETs to convert this TLK file ... |
--tlkcmp-toX | ... into this one. |
--tlkcmp-use-strings | When using --tlkcmp-from, emit commands of the form STRING_SET "Hello" @1 instead of STRING_SET #1 @1. |
--bcmp-from X | Emit APPLY_BCS_PATCH to turn this BCS file ... |
--bcmp-to X | ... into this one. |
--bcmp-orig X | Original file X to apply ... |
--bcmp-patch X | ... this patch to. |
Range Options | |
--min X | Lower range for some commands. See --traify-tlk, --tlkcmp-from and --string. |
--max X | Upper range for some commands. |
Automatic Module Packaging Options | |
--automate X | Automatically create a TP2 file for resources in folder X. See --out. |
--automate-min X | Only --automate string references above X. If not specified, it is assumed as 62169 (that is, the number of strings in unmodded SoA). |
--extract-kits X | Extract all kits starting with kit #X and create TP2 actions to install those kits as part of a module. |
Resouce Exploration Options | |
--list-eff X | List effects in resource X. See --out. |
F.ITM or F.EFF or F.SPL | Equivalent to --list-eff F.EXT. |
Logging Options | |
--log X | Log output and details to X. |
--autolog | Log output and details to WSETUP.DEBUG. |
--logapp | Append to log file instead of overwriting it. |
Finally, note that WeiDU will not add duplicate strings to DIALOG.TLK. If you instruct WeiDU to make use of the string “Imoen” (via SAY or --strapp or whatever) it will re-use any existing definition of “Imoen” instead. Two strings are equivalent in this sense only if they have the same text and the same associated sounds (normally strings have no associated sounds).
C:\Program Files\Black Isle\BGII - SoA\> weidu bodhi.dlg
[C:\Program Files\Black Isle\BGII - SoA\chitin.key] 182 BIFFs, 41793 resources [C:\Program Files\Black Isle\BGII - SoA\DIALOG.TLK] 84458 string entries [C:\Program Files\Black Isle\BGII - SoA\data\Dialog.bif] 2729 file entries [BODHI.DLG] loaded [.\BODHI.D] created from [BODHI.DLG]
This loads BODHI.DLG from the standard search path (i.e., the current directory, your override directory, then the game BIFFs) and creates BODHI.D from it.
C:\Program Files\Black Isle\BGII - SoA\> weidu bodhi.dlg --trans
... [.\BODHI.TRA] created as translation file [.\BODHI.D] created from [BODHI.DLG]
This creates BODHI.D as above and also the translation file BODHI.TRA (listing all of the strings in BODHI.D in an easy-to-traslate or spell-check format). BODHI.D will be created with special references to those strings.
This is particularly useful if you are converting existing modifications you may have created with another tool, such as IDU, into WeiDU format. It allows you to both create the WeiDU D code and the translation-friendly string labels at the same time.
C:\Program Files\Black Isle\BGII - SoA\> weidu --nofrom bodhi.dlg --out foozle.d --text
... [.\foozle.d] created from [BODHI.DLG]
This creates foozle.d (instead of BODHI.D) and does not put any "// from:" comments in foozle.d. It will include states with SAYs of the form
SAY ~Hello~ /* #1 */instead of
SAY #1 /* ~Hello~ */
C:\Program Files\Black Isle\BGII - SoA\> weidu bodhi.dlg jaheira.dlg --out test
... [test\JAHEIRA.D] created from [JAHEIRA.DLG] [test\BODHI.D] created from [BODHI.DLG]
This loads BODHI.DLG and JAHEIRA.DLG and creates BODHI.D and JAHEIRA.D. The optional --out test argument instructs WeiDU to put the resulting D files in the test directory.
C:\Program Files\Black Isle\BGII - SoA\> weidu bodhi.d
... [bodhi.d] parsed [BODHI.DLG] saved 135 states, 259 trans, 16 strig, 66 ttrig, 54 actions
This loads and parses bodhi.d and then executed all instructions in it. This bodhi.d file just defines BODHI.DLG, which is created. If bodhi.d contains strings that do not occur in DIALOG.TLK, BODHI.DLG will be created with invalid string references.
C:\Program Files\Black Isle\BGII - SoA\> weidu bodhi.d --tlkout new-DIALOG.TLK
... [bodhi.d] parsed [BODHI.DLG] saved 135 states, 259 trans, 16 strig, 66 ttrig, 54 actions [new-DIALOG.TLK] created, 84459 string entries
This loads and parses bodhi.d and then executed all instructions in it. This bodhi.d file just defines BODHI.DLG, which is created. If there are any new strings a new version of DIALOG.TLK is written to new-DIALOG.TLK.
C:\Program Files\Black Isle\BGII - SoA\> weidu ppworker.d bodhi.d --out test
... [bodhi.d] parsed [ppworker.d] parsed [BODHI.DLG] saved 135 states, 259 trans, 16 strig, 66 ttrig, 54 actions [PPWORKER.DLG] saved 33 states, 81 trans, 4 strig, 12 ttrig, 10 actions
This creates test/BODHI.DLG and test/PPWORKER.DLG based on the instructions in bodhi.d and ppworker.d. If these D files include new text, use --tlkout to make a new DIALOG.TLK.
C:\Program Files\Black Isle\BGII - SoA\> weidu examples/sola/solae1.d
OR
C:\Program Files\Black Isle\BGII - SoA\> weidu examples\sola\solae1.d
... [examples/sola/solae1.d] parsed [SOLA.DLG] loaded [SOLA.DLG] saved 336 states, 401 trans, 64 strig, 18 ttrig, 125 actions [SOLAE1.DLG] saved 36 states, 49 trans, 1 strig, 11 ttrig, 1 actions [SOLAE2.DLG] saved 3 states, 3 trans, 0 strig, 0 ttrig, 0 actions [SOLAE3.DLG] saved 2 states, 2 trans, 0 strig, 0 ttrig, 0 actions [SOLAE4.DLG] saved 3 states, 3 trans, 1 strig, 0 ttrig, 0 actions [SOLAE5.DLG] saved 2 states, 2 trans, 0 strig, 0 ttrig, 0 actions [SOLAE6.DLG] saved 4 states, 5 trans, 0 strig, 2 ttrig, 0 actions
It just so happens that solae1.d APPENDs text to SOLA.DLG and creates SOLAE1.DLG, SOLAE1.DLG, SOLAE3.DLG, ..., SOLAE6.DLG. You could have put them all in the override directory with --out override. You may use the forward slash (/) or the backslash (\) for directories. Use --tlkout to make a new DIALOG.TLK if these contain new text.
C:\Program Files\Black Isle\BGII - SoA\> weidu examples/sola/solafoe.d --transin examples/sola/solafoe.tra
OR
C:\Program Files\Black Isle\BGII - SoA\> weidu examples/sola/solafoe.d examples/sola/solafoe.tra
... [examples/sola/solafoe.tra] parsed (15 translation strings) [examples/sola/solafoe.d] parsed [SOLA.DLG] loaded [SOLA.DLG] saved 336 states, 401 trans, 65 strig, 18 ttrig, 124 actions [SOLAFOE.DLG] saved 11 states, 14 trans, 1 strig, 2 ttrig, 1 actions
It happens that solafoe.d uses 15 strings from a translation file, APPENDs to SOLA.DLG and creates SOLAFOE.DLG. You may use --transin to specify a translation file or (if it ends in TRA) just throw it on the command line. If you include multiple TRA files, the last one to define a particular string index wins for that string. They need not all cover the same set. Use --tlkout if there is new text involved.
C:\Program Files\Black Isle\BGII - SoA\> weidu --string 123 --strfind understudy --strfind acid.*rows
... [C:\Program Files\Black Isle\BGII - SoA\chitin.key] 182 BIFFs, 41793 resources [C:\Program Files\Black Isle\BGII - SoA\DIALOG.TLK] 84458 string entries String #123 is ~Haer' Dalis, all of you, stop them!~ String #6763 is ~Acid Arrows~ String #11662 is ~Biff The Understudy~ ...
This displays string #123 and all strings that contain the string "understudy" and all strings that match the regular expression (regexp) "acid.*rows". Note that case does not matter.
C:\Program Files\Black Isle\BGII - SoA\> weidu --strapp ANewString --tlkout happy.tlk
[C:\Program Files\Black Isle\BGII - SoA\DIALOG.TLK] 84458 string entries [.\happy.tlk] created, 84459 string entries
Not much to say here. String reference #84459 in happy.tlk is now “ANewString”.
C:\Program Files\Black Isle\BGII - SoA\> weidu --biff data/dialog.bif
... [data\Dialog.bif] contains ABELA.DLG at index 0 [data\Dialog.bif] contains ACHEN.DLG at index 1 ...
This shows all of the resources (e.g., ACHEN.DLG is a resource) that are contained in data/Dialog.bif.
C:\Program Files\Black Isle\BGII - SoA\> weidu --biff-get dragred.cre
[C:\Program Files\Black Isle\BGII - SoA\chitin.key] 182 BIFFs, 41793 resources [C:\Program Files\Black Isle\BGII - SoA\DIALOG.TLK] 84458 string entries [C:\Program Files\Black Isle\BGII - SoA\data\Creature.bif] 3194 file entries [.\dragred.cre] 1776 bytes, created from [C:\Program Files\Black Isle\BGII - SoA\data\Creature.bif]
This grabs Firkraag’s dragon-form CRE creature file from the game BIFFs and saves it in the current directory.
C:\Program Files\Black Isle\BGII - SoA\> weidu --biff-get sper.*itm
[.\chitin.key] loaded, 590551 bytes [.\chitin.key] 182 BIFFs, 41793 resources [.\DIALOG.TLK] loaded, 10154904 bytes [.\DIALOG.TLK] 77666 string entries [.\data\Items.bif] loaded, 659688 bytes [.\data\Items.bif] 1990 file entries [.\SPER01.ITM] 266 bytes, created from [.\data\Items.bif] [.\SPER02.ITM] 314 bytes, created from [.\data\Items.bif] [.\SPER03.ITM] 362 bytes, created from [.\data\Items.bif] [.\SPER04.ITM] 322 bytes, created from [.\data\Items.bif] [.\SPER05.ITM] 266 bytes, created from [.\data\Items.bif] [.\SPER06.ITM] 266 bytes, created from [.\data\Items.bif] [.\SPER07.ITM] 554 bytes, created from [.\data\Items.bif] [.\SPER08.ITM] 314 bytes, created from [.\data\Items.bif] [.\SPER09.ITM] 314 bytes, created from [.\data\Items.bif] [.\SPER10.ITM] 362 bytes, created from [.\data\Items.bif] [.\data\25Items.bif] loaded, 222370 bytes [.\data\25Items.bif] 479 file entries [.\SPER11.ITM] 314 bytes, created from [.\data\25Items.bif] [.\SPER12.ITM] 1610 bytes, created from [.\data\25Items.bif] [.\SPERMEL.ITM] 890 bytes, created from [.\data\25Items.bif]
This one assumes that the game is in the current directory and asks for every spear item in the game. Note that --biff-get uses regular expressions (regexp), not DOS-style wildcards. Note also that --biff-get does not look in the override directory. Finally, if you are using a Mac (or otherwise running unix) you’ll want to put the regular expression in double quotes, like so: C:\Program Files\Black Isle\BGII - SoA\> weidu --biff-get "sper.*itm"
C:\Program Files\Black Isle\BGII - SoA\> weidu --biff-type CRE --biff-str SPWI911
... LICH01.CRE in [data\Creature.bif] matches HLKANG.CRE in [data\Creature.bif] matches ...
This finds all CRE files that contain the string "SPWI911", which is equivalent to finding all enemy mages that know the spell Meteor Swarm (which has resource name "SPWI911"). You could also try something like:
C:\Program Files\Black Isle\BGII - SoA\> weidu --biff-type BCS --biff-str Terminsel
... AR0300.BCS in [data\Scripts.bif] matches AR0308.BCS in [data\Scripts.bif] matches JAHEIRA.BCS in [data\Scripts.bif] matches ...
to find all of the game scripts that include a variable that includes the substring "Terminsel". As you would expect, Jaheira shows up. Note that these searches are moderately time-consuming (e.g., searching all scripts takes about 20 seconds).
C:\Program Files\Black Isle\BGII - SoA\> weidu --tlkcmp-from DIALOG.TLK --tlkcmp-to dialog-asc.tlk
... [DIALOG.TLK] loaded, 8692747 bytes [DIALOG.TLK] 74107 string entries [dialog-asc.tlk] loaded, 10211578 bytes [dialog-asc.tlk] 82805 string entries WARNING: DIALOG.TLK has 74107 entries, dialog-asc.tlk has 82805 entries STRING_SET 70866 ~Babau~ [] STRING_SET 70867 ~Babau~ []
This compares all strings in common between two DIALOG.TLK files and generates a list of STRING_SET TP2 entries to convert the TLK file named in --tlkcomp-from into the TLK file named in --tlkcomp-to. In this case, WeiDU indicates there are two differences in the strings shared between a standard ToB TLK file and an Ascension Classic TLK file: strings 70866 and 70867 were changed to "Babau". Also note that the Ascension Classic TLK file has more entries (82805 compared to 74107).
If you have made a large number of manual changes to a TLK file (such as grammar/spelling corrections, or other dialogue tweaks), this is a handy way to generate install-ready scripting to apply those changes to an end user’s version of BG2.
Note that the use of --out will be helpful for a long list.
C:\Program Files\Black Isle\BGII - SoA\> weidu --tlkcmp-from DIALOG.TLK --tlkcmp-to dialog-asc.tlk --out mylist.txt
This will make a new file mylist.txt file that contains the STRING_SET parts of the output, which can then be put into a TP2 file.
C:\Program Files\Black Isle\BGII - SoA\> weidu --automate MyMod/SomeFolder --append MyMod.tp2
WeiDU’s --automate feature can save you oodles of time, so you probably will want to learn how to use it. It’s rather simple when you get the hang of it (but everything is right?).
Suppose you have just created some items, some spells and some creatures and some areas for you mod and you want to distribute them to others using WeiDU. You could manually write out string patching code by hand for each resource. Or you can get WeiDU to do it automatically.
WeiDU will scan every item, spell and creature inside the given folder (in this example, the folder is MyMod/SomeFolder) and emit TP2 commands to COPY those resources from that folder into the override folder. In addition, each resource’s current strings (like item descriptions and monster names) will be loaded from your TLK file and used to patch that resource as it is copied.
For example, the output of --automate on a folder that contains a potion of extra healing looks like this:
COPY ~MyMod/SomeFolder/potn52.itm~ ~override/potn52.itm~ SAY NAME ~Potion~ SAY NAME2 ~Potion of Extra Healing~ SAY UNIDENTIFIED_DESC ~Potions are typically found in ceramic, crystal, glass, or metal flasks or vials. Flasks or other containers generally contain enough fluid to provide one person with one complete dose to achieve the effects of the potion.~ SAY DESC ~When wholly consumed, this potion restores 27 hit points to the person. The effect is instantaneous and the potion is destroyed in the process.~ SAY 0xde ~Gulp!~ [GULP]
And there you have it, apparently there is a substitute for hard work.
Some translators who are using WeiDU to translate non-WeiDU mods find it handy to be able to convert between TLK and TRA files.
First, let’s create a TRA file:
C:\Program Files\Black Isle\BGII - SoA\> weidu --traify-tlk --min 2000 --max 2002
... @2000 = ~Indeed! It's been quite tasty so far. Listen, we're not here to devour everything. In fact, we'd like to help a little girl named Jaella.~ @2001 = ~No, we haven't. We will devour you if you don't tell us what we need to know.~ @2002 = ~Let us stop this charade. I'm only here to ask you a few questions.~ ...
You may also extract only those strings matching a regexp:
C:\Program Files\Black Isle\BGII - SoA\> weidu --traify-tlk --strfind lawyer
... @36568 = ~Honor-bound and honor-branded, then, is it? Very well, lawyer, you have set me free and for that I thank you.~ ...
Finally, you may redirect the output to a file using --out and read from a different TLK file by adding it on the command line.
Once you have a TRA file with a few entries you can create a TLK file from it:
C:\Program Files\Black Isle\BGII - SoA\> weidu --make-tlk my.tra --tlkout new.tlk
[c:\src\weidu\weidu.exe] WeiDU version 109 [C:\Program Files\Black Isle\BGII - SoA/chitin.key] 182 BIFFs, 41793 resources [C:\Program Files\Black Isle\BGII - SoA/dialog.tlk] 82405 string entries [my.tra] parsed [my.tra] has 100 translation strings New TLK will have 200 entries [new.tlk] created, 200 string entries
String @1 in your TRA file will become string reference #1 in the TLK file. If your TRA file has “holes” the new TLK file will have blank entries. You may specify --make-tlk multiple times: the last TRA file to define a translation string determine that string reference.
This tutorial was thoughtfully provided by Japheth.
Using --make-biff is dead easy. Here’s a quick and dirty example:
Assuming WeiDU is in your BGII directory, grab a bunch of files from your override folder and stick them into a folder inside your BGII directory. We’ll call the folder mybiff for this example.
Once the files are inside the folder, this is all you would have to do to make WeiDU create a biff of the files:
C:\Program Files\Black Isle\BGII - SoA\> weidu --make-biff mybiff
[c:\src\weidu\weidu.exe] WeiDU version 122 [./chitin.key] 182 BIFFs, 41793 resources [./dialog.tlk] 82479 string entries [data\mybiff.bif] will contain 20 resources totalling 211096 bytes [data\mybiff.bif] incorporating [mybiff/udsola02.cre] ... [data\mybiff.bif] incorporating [mybiff/sola10.cre] KEY saved (183 biffs, 41813 resources)
WeiDU will add the files into the BIFF mybiff.bif and stick that file in your data directory. It will also update the CHITIN.KEY.
Before using this feature, make sure you backup your CHITIN.KEY.
You can use --remove-biff to remove a BIFF file from CHITIN.KEY and destructively replace it with a new, smaller CHITIN.KEY. Before using this feature, make sure you backup your CHITIN.KEY. Do not use this feature.
C:\Program Files\Black Isle\BGII - SoA\> weidu --make-biff mybiff data\25ArMisc.bif
[c:\src\weidu\weidu.exe] WeiDU version 157 [./chitin.key] 182 BIFFs, 41793 resources [./dialog.tlk] 83772 string entries Removing references to 16 resources in [data\25ArMisc.bif] KEY saved (181 biffs, 41777 resources)
Note that the specified BIFF is not deleted from your computer’s file system, it is merely removed from CHITIN.KEY. When specifying the BIFF case does not matter but the folder separator, usually “\” or “:”, does. Use --list-biffs to get the official list of possible BIFFs to remove. Do not use this feature.
C:\Program Files\Black Isle\BGII - SoA\> weidu --nocom --text --transitive banomen.dlg
The --transitive flag tells WeiDU to follow EXTERN references when emitting D files. So the resulting BANOMEN.D file has lines like this:
IF WEIGHT #31 ~InParty("Edwin") See("Edwin") Gender("Edwin",FEMALE) !StateCheck("Edwin",STATE_SLEEPING) Global("BAnomen1","LOCALS",0)~ THEN BEGIN 10 SAY ~Hey, Edwina! I shall be your champion at the next tournament that we come to if only you give me a piece of your robe, uh, that is, dress to adorn my shield.~ [ANOMEN49] IF ~~ THEN DO ~SetGlobal("BAnomen1","LOCALS",1)~ EXTERN ~BEDWIN~ 104 END IF ~~ THEN BEGIN BEDWIN 104 SAY ~(My condition draws fools like flies to honey). Silence, you idiot! You've a death wish that is larger than your swollen head.~ [EDWINW39] IF ~~ THEN GOTO 11 END IF ~~ THEN BEGIN 11 SAY ~Fair Edwina, I am truly bereft by your non-acceptance. It is tragic when a knight has no fair maiden to moon over. Heh he he...~ IF ~~ THEN EXIT END
Note that both lines from both Edwin and Anomen are presented. The resulting “D” file is not valid in that it cannot be fed back to WeiDU directly, but it should make it easier for you to read all of the jokes offline.
Note that WeiDU may well go into an infinite loop when you use --transitive. Sorry about that.
This section includes tutorials on specific parts of WeiDU. Many of them were contributed by users like you.
This tutorial was thoughtfully provided by Jason Compton.
Although a single SAY line can be of any length, for style purposes (particularly in BG2) it is considered good form to break up very large lines into smaller chunks.
One can easily create a series of simple SAY blocks, one doing GOTO to the next, but if there are no special conditions being checked or actions being taken, you can very easily string several lines together.
Let’s say you have a scenery NPC teaching a lesson about the Bill of Rights to the US Constitution.
BEGIN TEACHER IF ~NumTimesTalkedTo(0)~ THEN BEGIN constitution_1 SAY ~On September 25, 1789, the First Congress of the United States therefore proposed to the state legislatures 12 amendments to the Constitution that met arguments most frequently advanced against it. The first two proposed amendments, which concerned the number of constituents for each Representative and the compensation of Congressmen, were not ratified. Articles 3 to 12, however, ratified by three- fourths of the state legislatures, constitute the first 10 amendments of the Constitution, known as the Bill of Rights.~ IF ~~ THEN EXIT ENDThis is a perfectly valid block of dialogue, but it is extremely long, and would likely scroll out of the text window for players with lower resolution.
Rather than break each sentence up into a new explicit state, we can use Multisay and save a lot of typing. Multisay is invoked with the = (equals) sign, which tells WeiDU that, "the current speaker should say another line here."
Here’s how that D state would look with Multisay:
IF ~NumTimesTalkedTo(0)~ THEN BEGIN constitution_1 SAY ~On September 25, 1789, the First Congress of the United States therefore proposed to the state legislatures 12 amendments to the Constitution that met arguments most frequently advanced against it.~ = ~The first two proposed amendments, which concerned the number of constituents for each Representative and the compensation of Congressmen, were not ratified.~ = ~Articles 3 to 12, however, ratified by three-fourths of the state legislatures, constitute the first 10 amendments of the Constitution, known as the Bill of Rights.~ IF ~~ THEN EXIT ENDThis will create three dialogue states, separated by simple "click to continue" transitions.
And that’s Multisay in a nutshell. Note that (as always with WeiDU) the line break and spacing before and after the = are totally optional, and used here only for illustration.
SAY ~One~ = ~Two~ = ~Three~is perfectly valid as well.
You may Multisay inside almsot any state, so you may use it within an APPEND D Action. This is valid:
APPEND J#KLSYJ IF ~~ THEN BEGIN Renal1_1 SAY ~All right, CHARNAME. I can accept that... you are right, there are bigger issues to consider.~ = ~But I hope you do understand why I said something, why it would be upsetting to have someone so close to me, in a role like that.~ IF ~~ THEN EXIT END // end of state Renal1_1 END // end of APPEND J#KLSYJHowever, you cannot use Multisay inside REPLACE, because the nature of REPLACE is to change (that is, replace) a single state, while Multisay’s nature is to create multiple states.
This tutorial was thoughtfully provided by Jason Compton.
CHAIN is an extension of the Multisay concept, simply with multiple participants. If you have two NPCs bantering back and forth for a prolonged period of time, and you do not need to do any special condition checks or actions as they babble, it can get very tedious to set up a separate IF/THEN/BEGIN/SAY block for each line.
Imagine a conversation like this:
If you wanted to add this witty little banter to your game, you could do it with 11 APPEND blocks for each line, or save a little time doubling up the back-to-back Imoen and Kelsey lines with Multisay inside their APPENDs, so you’d only need 9. In fact, you could get away with 2 APPEND blocks (remember that the states don’t have to appear in the order they are spoken, so you could mention all of Kelsey’s lines first and then all of Immy’s as long as the labels thread up the conversation correctly), one with 5 state declarations and one 4. But there’s an even better way, and that’s to use CHAIN. It works very much like Multisay. You use = to indicate that the current speaker should speak again, and == (that’s two consecutive equal signs) to indicate that a new speaker should take over.
Note that as of WeiDU 82, CHAIN can now define state triggers, perform DO actions, and end with an EXIT or COPY_TRANS. This means that for most simple NPC/NPC banters, where the PC does not have an opportunity to speak, you no longer need to use anything else.
Watch and see how this dialogue works using a CHAIN.
CHAIN IF ~Global("KelseyImoenPizza","LOCALS",0) InParty("Imoen2") See("Imoen2") !StateCheck("Imoen2",STATE_SLEEPING)~ THEN BJKLSY pizzachain ~Imoen, what do you like on your pizza?~ DO ~SetGlobal("KelseyImoenPizza","LOCALS",1)~ == IMOEN2J ~Oregano.~ = ~Oh, and maybe with a little basil mixed in.~ == BJKLSY ~Well, yeah, but anything else?~ == IMOEN2J ~Sauce is good.~ == BJKLSY ~(laughs) You're not being very helpful, Imoen.~ == IMOEN2J ~Crust. I like crust on my pizza. Cooked crust is better.~ == BJKLSY ~Do you want me to make you this pizza or not?~ = ~It WAS your idea.~ == IMOEN2J ~I can't decide. Never mind, I'm just gonna have yogurt.~ == BJKLSY ~(sigh)~ EXITNote how this dialogue works.
We use the CHAIN statement to define the state trigger (the starting conditions that must be true) and assign it to BJKLSY.DLG. The "pizzachain" label is mostly just for internal reference. Kelsey delivers the first line, then we use == IMOEN2J to allow her to answer, "Oregano." Then, we can use the single = to indicate that the current speaker (Imoen) has two consecutive lines.
Then it’s Kelsey’s turn to speak, so we use == BJKLSY to tell WeiDU to tell WeiDU to switch to the other speaker (which is Kelsey, since we specified BJKLSY. They banter back and forth for a while, and then when it is Kelsey’s turn to have back-to-back lines Do you want me to make you this pizza or not? and It WAS your idea., we separate with a single = to indicate that the current speaker (Kelsey) has two consecutive lines.
After Kelsey’s final, exasperated sigh, we use the EXIT command to terminate the CHAIN, and exit the dialogue.
And that’s all you need to know to use CHAIN. It saves a tremendous amount of time over setting up individual APPEND blocks, even Multisay blocks, for each NPC.
Advanced CHAINing:
You may include DO actions and conditionals inside chainText, as in:
CHAIN IF ~Global("KelseyImoenPizza","LOCALS",0) InParty("Imoen2") See("Imoen2") !StateCheck("Imoen2",STATE_SLEEPING)~ THEN BJKLSY pizzachain ~Imoen, what do you like on your pizza?~ DO ~SetGlobal("KelseyImoenPizza","LOCALS",1)~ == IMOEN2J ~Oregano.~ = ~Oh, and maybe with a little basil mixed in.~ == BJKLSY ~Well, yeah, but anything else?~ == IMOEN2J ~Sauce is good.~ == BJKLSY IF ~PartyHasItem("pepperoni")~ THEN ~Look, we HAVE pepperoni. Why don't I just use that? I'll eat it, anyway. If you don't like it, have yogurt instead.~ == IMOEN2J IF ~!PartyHasItem("pepperoni")~ THEN ~Crust. I like crust on my pizza. Cooked crust is better.~ == BJKLSY IF ~!PartyHasItem("pepperoni")~ THEN ~Do you want me to make you this pizza or not?~ = ~It WAS your idea.~ == IMOEN2J IF ~!PartyHasItem("pepperoni")~ THEN ~I can't decide. Never mind, I'm just gonna have yogurt.~ == BJKLSY IF ~!PartyHasItem("pepperoni")~ THEN ~(sigh)~ EXITIn this case, the dialogue changes if the party has the pepperoni item. If it does, Kelsey says the we HAVE pepperoni and then (sigh) and then the dialogue ends. If not, the dialogue works as before. The chainText lines with IFs in them are only spoken if their conditionals are true.
This tutorial was thoughtfully provided by Jason Compton.
There are some complex branching dialogues in Infinity Engine games that you, as a mod creator, may wish to add to. Consider Baldur’s Gate 2 and the "Arrival In Hell" dialogue. There is a brief internal dialogue as the protagonist comes to terms with the fact that he/she is now in Hell, and then all of the companions coded by Bioware have a chance to speak. (PLAYER1.DLG state 25. It will help the rest of this explanation if you go use WeiDU to decompile PLAYER1.DLG into PLAYER1.d, and/or open up PLAYER1.DLG in Near Infinity and look at state 25.)
After the PC’s internal voice says You doubt they will be pleased with their present circumstance, when you don’t even know why you are here yourself., every Bioware NPC has the opportunity to speak. If you are creating a new NPC and want it to have that full, rich Bioware flavor, you may wish to let your character speak here as well. For that, a simple EXTEND_BOTTOM will do the job.
Here’s the example of how Weimer does this with Solaufein:
EXTEND_BOTTOM PLAYER1 25 IF ~IsValidForPartyDialogue("Sola") Global("SolaWelcomeHell","GLOBAL",0)~ THEN DO ~SetGlobal("SolaWelcomeHell","GLOBAL",1)~ EXTERN SOLA inHell1 ENDThis puts a new transition at the bottom of PLAYER1 25 that tells the game to branch to Solaufein’s observation about your arrival in Hell, if he is present (IsValid) and if we have not already seen his comment once before (the check for SolaWelcomeHell=0, then setting it to 1. This ensures that this path can only happen once, which is important for a reason I will explain later.) SOLA inHell1, which we will define later, contains Solaufein’s comment.
Once Solaufein makes his comment, it would be very thoughtful of us to allow the other Bioware NPCs to have their say as well, as Bioware intended and as the experienced players out there expect. You could simply use the decompiled PLAYER1.D and copy and paste the transition list out. But there are some good reasons not to do that. On a trivial level, it’s a big waste of space in your D file.
The most important reason is this: If another mod NPC has come along and done their own EXTEND_BOTTOM, you would have no way of knowing that. By putting the Bioware stock transition list in, you would ensure that only your mod NPC got to have their say. The rest would be silent. So if you were the developer of Solaufein, and Solaufein were installed after Kelsey and Tashia in a game, Kelsey and Tashia would be skipped, because you only copied the Bioware transition list. That’s a heavy responsibility.
That’s why COPY_TRANS exists. COPY_TRANS pulls the entire transition list from a specified state and makes it the transition list for your new state.
To illustrate, look at SOLA inHell1 :
APPEND SOLA IF ~~ THEN BEGIN inHell1 SAY @2 = @3 // use strings @2 and @3 from translation file COPY_TRANS PLAYER1 25 END ENDInstead of copying and pasting that huge list of IF "" THEN EXTERN transitions from the PLAYER1.D, we let WeiDU do it for us. COPY_TRANS PLAYER1 25 tells WeiDU to grab the current list of transitions from PLAYER1 state 25, and use it as the transition list for SOLA inHell1. This ensures that Solaufein will be able to properly branch out to Imoen, Aerie, Minsc, and the rest of the gang, as well as grabbing the transitions that may have been added by other NPCs such as Kelsey, Valen, or Tashia.
COPY_TRANS can form all of your new state’s transition list, or only part of it. This would be valid, for example:
IF ~~ THEN BEGIN commentary SAY ~Hey, I think I might like to run the transition list from TOLGER 75... or I might want to do something else, if I'm in chapter six!~ COPY_TRANS TOLGER 75 IF ~Global("Chapter","GLOBAL",6)~ THEN GOTO chapter6commentary ENDThis would make the GOTO commentary2 transition show up at the bottom of the transition stack (below the list copied from TOLGER 75). Remember that transition triggers are read bottom to top, so it would be the first transition evaluated. If you want it to be evaluated after the list of transitions in the COPY_TRANS, put it above. Note, however, that Bioware usually structures its transition lists so that the topmost trigger will always be true (in fact, sometimes it is "True()") so it is somewhat unlikely you would ever want to put a new transition trigger above the COPY_TRANS.
Now, that explanation for why the SolaWelcomeHell variable check is important: if a user accidentally installs the same mod more than once, and it employs COPY_TRANS, the list the second time around will include our new trigger:
IF ~IsValidForPartyDialogue("Sola") Global("SolaWelcomeHell","GLOBAL",0)~ THEN DO ~SetGlobal("SolaWelcomeHell","GLOBAL",1)~ EXTERN SOLA inHell1If there was no flag being set to ensure that the transition could only run once, the user would get stuck in a loop. This can and has happened with mods in the wild. The end result would be Solaufein or Kelsey or whomever constantly offering their commentary, over and over again.
Important note: The WeiDU D compiler runs COPY_TRANS before other actions that you might take to affect a transition list within the same D file (like EXTEND_TOP and EXTEND_BOTTOM). This is a good thing.
This tutorial was thoughtfully provided by Jason Compton.
Interjections, the little comments party members make, are a great way to spice up a new NPC or a new quest you create. It shows that the characters are paying attention to their game world, and that they have an opinion about what goes on around them.
Through interjections, an NPC can advise a course of action, complain about a decision, force your hand... fun things.
The traditional way to do an interjection is to find a state of a dialogue where another NPC might comment, and use EXTEND_BOTTOM and APPEND in conjunction.
Here’s an old-school example:
EXTEND_BOTTOM SAHPR4 7 IF ~IsValidForPartyDialog("J#Kelsey")~ THEN EXTERN J#KLSYJ KelseySAHPR4 END APPEND J#KLSYJ IF ~~ THEN BEGIN KelseySAHPR4 SAY ~Urk. Who was the lucky donor?~ IF ~~ THEN EXTERN SAHPR2 10 END ENDThis works, but it’s also more work than it needs to be since the introduction of INTERJECT. (Incidentally, for you WeiDU historians out there, there are two main reasons advanced functions have been added to WeiDU since the first versions of Solaufein and the original CHAIN command: Either Weimer needed them for his own modding goals, or a mod project, usually Kelsey, requested it. INTERJECT came about when Westley finally decided that Solaufein and Valen should comment about quests. On the other hand, COPY_TRANS and INTERJECT_COPY_TRANS were my idea.)
INTERJECT simplifies this process considerably. To run an INTERJECT, you need to know the source state (the line after which you want one or more NPCs to interject), and the destination state (where you want the dialogue tree to go after the interjection.) Typically, but not necessarily, the destination state will be wherever the dialogue was originally planning to go, but if the NPC takes the conversation in a new direction, that may change. We’ll stick with the simpler cases for illustration.
INTERJECT is a specialized form of CHAIN. So if you’re familiar with CHAIN, this will look familiar.
Consider the dryads in the Irenicus start dungeon. Perhaps Minsc should say something to them. IDRYAD1.DLG state 1 offers a good opportunity.
IF ~~ THEN BEGIN 1 // from: SAY #11080 /* ~We are his possessions.~ */ IF ~~ THEN EXTERN ~IDRYAD2~ 1 ENDMinsc is outraged. Here’s how he can express it. The idea is that we want Minsc to comment, but for the dialogue to continue just as it would have if he was not there.
INTERJECT IDRYAD1 1 MinscDryad == MINSCJ IF ~IsValidForPartyDialog("Minsc")~ THEN ~Boo is outraged that the strange wizard would own these lovely ladies! Can Minsc and Boo help you nice girls?~ END IDRYAD2 1Here’s what’s going on here.
Invoking INTERJECT requires three arguments: the dialogue name and state we’re interjecting into, plus a unique global variable name. This variable will be set from 0 to 1 after the INTERJECT runs, to ensure that it can only happen once. (This is important in case players accidentally install your mod twice, as it could create a looping problem similar to the one described in COPY_TRANS.)
So INTERJECT IDRYAD1 1 MinscDryad tells WeiDU "Put this dialogue after IDRYAD1 state 1. This dialogue will run if MinscDryad is 0. After it runs, we will set MinscDryad to 1."
Then we need to define who is speaking. == (that’s two consecutive equal signs) is CHAIN-style notation for a new speaker, and MINSCJ is the proper dialogue to use for Minsc’s "joined-party" commentary. If Minsc is in the party and valid for dialogue, he will say his line. After that, we transition to IDRYAD2 state 1 (END IDRYAD2 1), which is where the dialogue was heading in the first place. We can make this more complicated and let the dryad reply to his interruption before proceeding.
INTERJECT IDRYAD1 1 MinscDryad == MINSCJ IF ~IsValidForPartyDialog("Minsc")~ THEN ~Boo is outraged that the strange wizard would own these lovely ladies! Can Minsc and Boo help you nice girls?~ == IDRYAD1 IF ~IsValidForPartyDialog("Minsc")~ THEN ~Large mortal, we are having a dramatic scene. Please do not interrupt.~ END IDRYAD2 1Note that we repeat the
IF ~IsValidForPartyDialog("Minsc")~ THENcheck for IDRYAD1. That is to ensure that she only says that line if Minsc is there, and by extension has already made his little comment. So now, Minsc interrupts, Dryad #1 scolds him, and then we proceed to the second dryad’s line.
One general piece of advice: while it’s not necessary to pick a state that has only a single transition to another state, unless you’re willing to experiment (or you’re intentionally trying to remove player choices from the equation, by making the NPC say something that forces immediate action, for instance), don’t INTERJECT into a state that has player options (REPLYs).
This tutorial was thoughtfully provided by Jason Compton.
INTERJECT is good for creating interjections where none already exist. However, many of the really good opportunities for interjections in game dialogue already have interjections in them. If you use standard INTERJECT, chances are you’ll skip right over them.
Consider TOLGER.DLG state 75. After Tolgerias lays down the law by saying This is a sensitive matter, and I cannot tell all to every curious soul. I must have your commitment that you agree to the task, four NPCs (Edwin, Jaheira, Yoshimo, Korgan) will tell you what they think of that rotten arrangement. What’s more, Bioware structured the dialogue transitions so that they all can get their comment in, if all four of them are in the party.
However, using a standard INTERJECT for a new NPC line would skip over those four comments, which is rather impolite. INTERJECT_COPY_TRANS works much like regular INTERJECT, but instead of defining a state to transition to after END, WeiDU will COPY_TRANS the transition list from the state you are INTERJECTing into.
This is not as confusing as it sounds. Watch as hypothetical new NPC Aqualung responds to Tolgerias’s terms:
INTERJECT_COPY_TRANS TOLGER 75 AquaTolger == AQUALUNJ IF ~IsValidForPartyDialogue("Aqualung")~ THEN ~Hey, that's a really crummy offer! Where did those little girls go? I could be sitting on a park bench, I don't need this aggravation! Who are you, anyway?~ == TOLGER IF ~IsValidForPartyDialogue("Aqualung")~ THEN ~You poor old sod, you see it's only me. Now, did anyone else have a smart remark they wanted to make?~ ENDSo, if Aqualung is around, we’ll hear from him and then Tolgerias will respond to him. After that, the game will look for the presence of Edwin, Jaheira, Yoshimo, and Korgan and we’ll get their responses to Tolgerias as well.
Hint: INTERJECT_COPY_TRANS is fine to use even if there were no other interjections in the source state, i.e. if there’s just a single IF "" THEN GOTO blah. That is, as long as you plan to proceed to the original destination. It saves you the trouble of having to look up and input the destination.
This tutorial was thoughtfully provided by Rastor.
There seems to be a great deal of confusion among the members of the modding community regarding the purpose and overall function of INTERJECT_COPY_TRANS2. Allow me to take a moment to dispel the rumors that you may have heard.
INTERJECT_COPY_TRANS2 is not intended as a replacement for INTERJECT_COPY_TRANS. INTERJECT_COPY_TRANS is a great function but it has a flaw in certain special cases. That’s what INTERJECT_COPY_TRANS2 is intended to remedy. This difference between the two is best illustrated with an example.
Let us suppose that you want to add an NPC response to PPSAEM2 8. Here is some of what that state looks like before any mods have been applied:
IF ~~ THEN BEGIN 8 // from: 7.0 SAY #44931 /* ~Blah blah blah~ */ IF ~!IsValidForPartyDialog("Jaheira") !IsValidForPartyDialog("Anomen") !IsValidForPartyDialog("Edwin") IsValidForPartyDialog("Viconia")~ THEN DO ~SetGlobal("WackoArmy","GLOBAL",1) OpenDoor("DOOR12") EscapeArea()~ UNSOLVED_JOURNAL #7045 EXTERN ~VICONIJ~ 129 END
Note that the DO actions include EscapeArea(). If you use INTERJECT_COPY_TRANS to add an interjection here, the DO actions will be performed by your interjector (usually a party member). The result is that after encountering this dialogue in the game the party member will leave the area (and Saemon will stay where he is!) promptly after performing the interjection. Instead, we want to keep the DO action associated with Saemon.
INTERJECT_COPY_TRANS2 allows modders to remedy this problem. INTERJECT_COPY_TRANS2 does exactly the same thing as INTERJECT_COPY_TRANS except that DO actions will be kept with their original actor and not transferred to the interjector.
To code your interjection using INTERJECT_COPY_TRANS2, you would do this:
INTERJECT_COPY_TRANS2 PPSAEM2 8 ~Blah blah blah~ END
The dialogue state that WeiDU will create from this INTERJECT_COPY_TRANS2 statement looks something like:
IF ~~ THEN BEGIN 133 // from: SAY #78199 /* ~Blah blah blah~ */ IF ~!IsValidForPartyDialog("Jaheira") !IsValidForPartyDialog("Anomen") !IsValidForPartyDialog("Edwin") IsValidForPartyDialog("Viconia")~ THEN UNSOLVED_JOURNAL #7045 EXTERN ~VICONIJ~ 129 END
Note that INTERJECT_COPY_TRANS2 is not intended as a universal replacement for INTERJECT_COPY_TRANS. This is most obvious when using the command to interject into a state that starts a cutscene. Here is an example of an interjection into PPIRENI2 27.
The original, unmodded state:
IF ~~ THEN BEGIN 27 // from: 28.0 26.0 SAY #44869 /* ~I bid you farewell, child of Bhaal. We shall not meet again.~ [IRENIC52] */ IF ~~ THEN DO ~EraseJournalEntry(7252) EraseJournalEntry(7253) EraseJournalEntry(22952) EraseJournalEntry(23306) SetGlobal("AsylumPlot","GLOBAL",40) StartCutSceneMode() StartCutScene("Cut41j")~ SOLVED_JOURNAL #7255 EXIT END
Using INTERJECT_COPY_TRANS2 to code your interjection into this state will cause the game to crash when your interjection plays (because the special StartCutSceneMode() action cannot occur in the middle of a dialogue, loosely). INTERJECT_COPY_TRANS will work properly, however.
stateTriggerStrings, the conditions that determine what state should be used for the beginning of a dialogue, may have WEIGHTs. These WEIGHTs are used by the Infinity Engine to choose which state to pick if multiple state triggers evaluate to "true". [ In reality, the WEIGHTs are just the offsets within the state trigger table in the DLG file, but this detail is not important unless you are writing your own tool. ] WEIGHTs only make sense for stateTriggerStrings that are not empty.
If multiple stateTriggerStrings evaluate to true, the Infinity Engine will pick the state with the lowest WEIGHT. Usually the weighting follows the order of state declaration in the D file. That is, the first state mentioned has the lowest weight (i.e., will be picked first in case of a tie) and the last state mentioned has the highest weight (i.e., will be picked last in case of a tie). However, you may include an explicit WEIGHT directive to change things around. For example, consider this D file:
BEGIN foozle IF ~True()~ THEN BEGIN a SAY ~Jason~ END IF ~True()~ THEN BEGIN b SAY ~yada~ END IF ~True()~ THEN BEGIN c SAY ~Compton~ END IF ~True()~ THEN BEGIN d SAY ~kelsey~ ENDIf you talk to foozle, it will always say Jason. However, you may explicitly alter the weights so that the third state is picked first, as in:
BEGIN foozle IF WEIGHT #10 ~True()~ THEN BEGIN a SAY ~Jason~ END IF ~True()~ THEN BEGIN b SAY ~yada~ END IF WEIGHT #2 ~True()~ THEN BEGIN c SAY ~Compton~ END IF ~True()~ THEN BEGIN d SAY ~kelsey~ ENDWith this D file, foozle will always say Compton. All states with WEIGHT directives come before all states without them. States without WEIGHT directives are ranked in order of appearance. So the state order for foozle.DLG is c-a-b-d. Yes, this is complicated.
Strong Style Suggestion: do not use the WEIGHT directive in your hand-made D files. Just use the implicit ordering.
The WEIGHT directive was introduced to facilitate handling of Bioware-created DLG files (e.g., BJAHEIR.DLG) that include tricky weighting. Only states with non-empty triggers are given implicit weights. If you create a D file from a DLG that features non-trivial weighting, WeiDU will emit comments like this:
IF WEIGHT #8 /* Triggers after states #: 11 12 24 25 26 36 58 even though they appear after this state */ ~True()~ THEN BEGIN 10 // from: SAY #52190 /* ~Please do not interrupt our thoughts. We must prepare carefully if we are to see a weakness in the illithid web. ~ */ IF ~~ THEN EXIT ENDto remind you that the order is not what you think.
All non-empty state triggers in DLG files are given weights counting up from 0 to the maximum number of state triggers in the DLG file. You may use any number you like (even a negative one): WeiDU will simply sort them. ADD_STATE_TRIGGERdoes not change the weight associated with that trigger. APPEND can be used to give a non-trivial weight to a state, as in:
APPEND BJAHEIR IF WEIGHT #-999 ~MyCondition()~ THEN BEGIN mystate SAY ~My Stuff~ END ENDSince BJAHEIR will have implicit WEIGHTs in the range from #0 to about #50, this causes mystate to have priority over all states that already exist in BJAHEIR. Without such drastic action, APPENDed states will have use the implicit ordering, and will thus have the lowest priority (because they appear at the end of the file). Multisay and CHAIN also append states, but since they always append states with empty stateTriggerStrings, WEIGHTs are not relevant.
Consider the following example:
BEGIN foozle IF WEIGHT #10 ~True()~ THEN BEGIN a SAY ~Jason~ END IF ~~ THEN BEGIN b SAY ~yada~ END IF WEIGHT #2 ~True()~ THEN BEGIN c SAY ~Compton~ END IF ~True()~ THEN BEGIN d SAY ~kelsey~ END ADD_STATE_TRIGGER foozle 1 /* state b */ ~MyCondition()~The resulting foozle dialogue will still have the c-a-b-d weighting order.
Here’s another example:
BEGIN foozle IF ~True()~ THEN BEGIN a SAY ~Jason~ END IF ~~ THEN BEGIN b SAY ~yada~ END IF ~True()~ THEN BEGIN c SAY ~Compton~ END IF ~True()~ THEN BEGIN d SAY ~kelsey~ END ADD_STATE_TRIGGER foozle 1 /* state b */ ~MyCondition()~The resulting foozle dialogue will have the (expected) a-b-c-d ordering.
However, consider this evil example:
////// // foozle.DLG contents, assume it has already been created and is // sitting on your hard drive somewhere // IF ~True()~ THEN BEGIN a SAY ~Jason~ END // IF ~~ THEN BEGIN b SAY ~yada~ END // IF ~True()~ THEN BEGIN c SAY ~Compton~ END // IF ~True()~ THEN BEGIN d SAY ~kelsey~ END ////// // new D file ADD_STATE_TRIGGER foozle 1 /* state b */ ~MyCondition()~This will update foozle and the resulting order will be a-c-d-b (because when foozle.DLG was loaded from the disk, a c and d were given weights but b was not (because it had an empty trigger)). Thus, you should avoid using ADD_STATE_TRIGGER on states with empty triggers unless you know what you are doing.
If you are writing a mod and you would like to make it easier to translate it into another language you can use "translation files" (much like BGII itself uses DIALOG.TLK) to separate your dialogue structure and content. A translation file basically lists the string texts in order. For example,
C:\Program Files\Black Isle\BGII - SoA\> WeiDU --trans SCSARLES.DLG
This creates scsarles.D and scsarles.tra. scsarles.D now contains:
IF ~NumTimesTalkedTo(0)~ THEN BEGIN 0 // from: SAY @1 /* ~Who is it? Might I ask why you have disturbed my meditations? My creative muse must be gently awakened, and your stomping about is simply not conducive to this.~ [SARLES02] #28655 */ IF ~~ THEN REPLY @2 /* ~My apologies. I will leave you to your thinking.~ #28656 */ GOTO 1 IF ~~ THEN REPLY @3 /* ~I apologize, but I have come to request your talent on a commissioned artwork.~ #28657 */ DO ~SetGlobal("TalkedToSarles","GLOBAL",1)~ GOTO 2 ENDNote that all of the strings have been replaced by @number and the texts have been put in comments.
The translation file scsarles.tra contains all of those strings:
// SCSARLES translation file @1 = ~Who is it? Might I ask why you have disturbed my meditations? My creative muse must be gently awakened, and your stomping about is simply not conducive to this.~ [SARLES02] @2 = ~My apologies. I will leave you to your thinking.~ @3 = ~I apologize, but I have come to request your talent on a commissioned artwork.~You may then ask someone who speaks another language to write a new translation file by translating every string in scsarles.tra. This prevents the string text and the structure from getting out of sync and simplifies translation; non-technical players can translate raw text files easily.
When compiling a D file that contains @number translation references you must supply (at least) one translation file. For example, you might say:
C:\Program Files\Black Isle\BGII - SoA\> WeiDU SCSARLES.D italian.tra
You may specify multiple translation files. The last one to define a string wins. This is useful if one language is more up to date than the others. In this example:
C:\Program Files\Black Isle\BGII - SoA\> WeiDU SCSARLES.D english.tra italian.tra
Strings will be taken from the italian translation whenever possible, but if they are not available it will fall back on the english versions.
You may use WeiDU to check and make sure that translations are up to date. WeiDU will automatically generate a text file listing all of the strings that are present in one translation (usually your native one) that are missing in another. You can then send this file to your translators so that they know what to do. This example command compares all of the TRA files in the american and french directories and creates a file called MISSING.
C:\Program Files\Black Isle\BGII - SoA\> weidu --tcmp-from american --tcmp-to french --out MISSING
This tutorial was thoughtfully provided by Jason Compton.
D and TP2 files allow programmers to describe text either literally:
// Greeting.d SAY ~Hello.~or with a companion TRA (translation) file that supports multiple languages:
// Greeting.d SAY @1 // Greeting.tra @1 = ~Hello.~The idea being that one can make a French version of Greeting.tra which contains
// French-Greeting.tra @1 = ~Bonjour.~However, some WeiDU users, for reasons of convenience or simply never anticipating the opportunity to translate, may have originally chosen the "hard-coded" approach but now regret that decision. --traify will break out all the text in SAY, REPLY, and JOURNAL entries into translation-ready format. --traify may also be used on TP2 files.
To turn the hard-coded D file FWKI.d into a new D/TRA combo, use --traify and --out to specify the input and output filenames, respectively:
C:\Program Files\Black Isle\BGII - SoA\> weidu --traify fwki.d --out fwki-new.d
After a brief pause, fwki-new.d and fwki-new.tra will be created.
The --traify process turns fwki.d’s
APPEND J#KLSYJ IF ~~ THEN BEGIN KelseySAHPR4 SAY ~Urk. Who was the lucky donor?~ IF ~~ THEN EXTERN SAHPR2 10 END ENDinto fwki-new.d’s
APPEND J#KLSYJ IF ~~ THEN BEGIN KelseySAHPR4 SAY @0 IF ~~ THEN EXTERN SAHPR2 10 END ENDand in the newly created fwki-new.tra, you will find
@0 = ~Urk. Who was the lucky donor?~and this dialogue file is now ready for translation.
--traify works from the top of the D down, starting at @0. It will NOT skip over any existing @x translation references it finds, so if your D contains any translation support at all, it is best to use --traify-old-tra as well.
Because standard --traify starts at @0 and is unaware of any any existing @x entries in the D, if you have begun to convert a D to a D/TRA pair by hand, you may have @x entries that clash with --traify’s results.
In other words, if you already have a state that says
IF ~~ THEN BEGIN blah SAY @0 IF ~~ THEN EXIT END--traify will not skip @0 automatically, you will have two locations where @0 is used but you intended to use two different strings, and this will be bad.
To avoid this problem, add the --traify-old-tra argument to specify the file containing the already traified strings:
C:\Program Files\Black Isle\BGII - SoA\> weidu --traify fwki.d --traify-old-tra fwki.tra --out fwki-new
will put the existing and new ones @x references in fwki-new.tra.
Finally, note that --traify works for BAF files as well.
Note: We were previously suggesting to use --traify# here. Unless you want for some reason to have new references starting at a given number, you should use --traify-old-tra instead.
In all cases, you might also find it beneficial to put the --traify-comment argument in your command line, so that you’ll also have the content of the @references in your file (which will make it easier to edit it later):
C:\Program Files\Black Isle\BGII - SoA\> weidu --traify fwki.d --traify-old-tra fwki.tra --out fwki-new --traify-comment
In particular, using --traify-comment and --traify-old-tra together will add the missing comments without doubling the existing ones.
This tutorial was thoughtfully provided by Japheth.
REPLACE_ACTION_TEXT can be used with regular expressions (regexp). So, some actions such as CreateCreature("blah",[0.0],1) won’t be matched by WeiDU because when you say [0.0] it is looking for a character set.
Here’s one dialogue that I’m fixing and how to do it correctly.
Arghai.dlg has this action trigger before REPLACE_ACTION_TEXTing it.
CreateCreature("OGREHA",[1351.1078])}
There’s no point given, which makes WeiDU and NI angry. To fix it using REPLACE_ACTION_TEXT this is what you have to do:
REPLACE_ACTION_TEXT arghai ~CreateCreature("OGREHA",\[1351.1078\])~ ~CreateCreature("OGREHA",[1351.1078],0)~
So, all you have to remember to do is escape the square brackets with a backslash so WeiDU doesn’t confuse them with a regular expression.
This tutorial was thoughtfully provided by CamDawg.
ALTER_TRANS is a more compatibility-friendly way of changing transitions in an existing dialogue state. This is best illustrated with an example; let’s use state 9 from doghma.dlg in BG2:
IF ~~ THEN BEGIN 9 SAY #45751 IF ~~ THEN REPLY #45752 GOTO 2 IF ~GlobalLT("chapter","GLOBAL",4)~ THEN REPLY #45753 GOTO 5 IF ~~ THEN REPLY #45754 GOTO 4 IF ~Global("RevealUmar","GLOBAL",1) PartyHasItem("miscbl") PartyHasItem("miscbm") PartyHasItem("miscbn") PartyHasItem("miscbo")~ THEN REPLY #57922 GOTO 10 IF ~~ THEN REPLY #45755 GOTO 1 END
The problem lies in transition 3, which should have an OR(4) before those four PartyHasItem checks. Before ALTER_TRANS, the least destructive way to fix it would be
ADD_TRANS_TRIGGER DOGHMA 9 ~False()~ DO 3 EXTEND_BOTTOM DOGHMA 9 IF ~Global("RevealUmar","GLOBAL",1) OR(4) PartyHasItem("miscbl") PartyHasItem("miscbm") PartyHasItem("miscbn") PartyHasItem("miscbo")~ THEN REPLY #57922 GOTO 10 END
This presents some compatibility issues though. If another mod has altered this transition (for example, a new action via ADD_TRANS_ACTION) then those changes are lost via this method. It also reorders the replies as they’re displayed on-screen--not a big issue for transitions involving replies, but it is a big issue for non-reply transitions where they’re evaluated from the bottom-up. It’s also an issue for other mods that target this transition, as the False() essentially eliminates it. Enter ALTER_TRANS:
ALTER_TRANS DOGHMA // file name BEGIN 9 END // state number (can be more than one) BEGIN 3 END // transition number (can be more than one) BEGIN // list of changes, see below for flags "TRIGGER" ~Global("RevealUmar","GLOBAL",1) OR(4) PartyHasItem("miscbl") PartyHasItem("miscbm") PartyHasItem("miscbn") PartyHasItem("miscbo")~ END
There are eight flags you can use in the list of changes. If you do not use a particular flag, the previous value of the transition will be retained (i.e. not specifying an ACTION flag will preserve the current action of the transition).
"TRIGGER" ~Global("foo","GLOBAL",0)~
"ACTION" ~SetGlobal("foo","GLOBAL",1)~
"REPLY" ~#57922~
"JOURNAL" ~@100~
EPILOGUE" ~EXTERN BVICONI 0~
Let’s try one last example. This is bviconi.dlg, state 103:
IF WEIGHT #22 ~Global("LoveTalk","LOCALS",46)~ THEN BEGIN 103 SAY #10537 IF ~~ THEN REPLY #10538 GOTO 367 IF ~~ THEN REPLY #10539 EXTERN ~~ 0 END
As you can see the transition destination for the second transition is broken. In the olden days this would be fixed with
ADD_TRANS_TRIGGER BVICONI 103 ~False()~ DO 1 EXTEND_BOTTOM BVICONI 103 IF ~~ THEN REPLY #10539 GOTO 368 END
With ALTER_TRANS, it beomes much easier:
ALTER_TRANS BVICONI BEGIN 103 END BEGIN 1 END BEGIN "EPILOGUE" ~GOTO 368~ END
Addendum: please note that it’s important to wrap the target (for example, REPLY or EPILOGUE) in quotes, tildas, or percentage signs.
Further addendum: REPLY, JOURNAL and friends will remove the feature (rather than add an empty journal line or whatever) if the associated string is empty.
At some point you will be done with your mod (a collection of CRE, ITM, D, etc., files) and you will want to package it up so that other users can install it (and then perhaps uninstall it later) easily. WeiDU can handle this task for you (and you may freely distribute WeiDU.exe with your module).
A TP2 describes how to install components of your module. WeiDU will read the file, ask the user questions, and then perform the installation. Uninstallation and upgrading are also handled.
See the file examples/mymod.tp2 for a commented example of how this all works.
Here is the context-free grammar syntax for the TP2 file format:
TP2 File | A TP2 File is a text file that contains a number of mod Components. TP2 Files tell WeiDU how to install various parts of your mod on an end-user’s computer. | |
is | BACKUP directoryName TP2 Flag list Language list Component list | A TP2 File is basically a prologue and then a list of Components. The BACKUP declaration tells WeiDU where to put backed-up versions of files that would be overwritten so that they can be uninstalled later. This directory will be created automatically if it is missing. The AUTHOR directive gives an email address for users to send bugs to if there are problems during the installation. The TP2_AUTHOR variable is set to the “emailAddress” value. The TP2_FILE_NAME variable is set to the name of the tp2 file. TP2 Flags set global options. Languages are the various languages in which your mod is available. The Finally, the Components make up the actual meat of your mod. Different Components can be installed or uninstalled separately, but all of the parts within a Component are treated as a unit. |
TP2 Flag | A TP2 Flag declaration tells WeiDU to apply some global option to your TP2 file. | |
is | AUTO_TRA path | The AUTO_TRA flag is used with the COMPILE TP2 Action. It automatically loads TRA files that match your D files. |
or | ALLOW_MISSING file list | ALLOW_MISSING directive allows you to specify files that can be missing (when you try to copy them or reference them from D files). Empty versions of those files will be created on demand. Try to use ACTION_IF instead of this. |
or | ASK_EVERY_COMPONENT | This flag instructs WeiDU to ask about installing every component in this TP2 file individually, rather than asking questions like "Would you like to install them all?" |
or | ALWAYS TP2 Action list END | This flag specified a TP2 Action that is executed at the beginning of each Component, before the component-specific TP2 Action. |
or | README file list | The given file is displayed to the user after choosing the language, but before installing any component. If more than one file name is provided, they are checked for existance in the order they’re provided, and the first available one will be opened. If you wish to open multiple readme files, use multiple README statements. |
or | UNINSTALL_ORDER String list | Don’t use without a real reason. Specifies the order in which uninstallation operations are handled. The operations are STRSET (uninstall STRSET and ALTER_TLK), MOVE (uninstall MOVE), AT (perform AT_UNINSTALL and AT_INTERACTIVE_UNINSTALL) and COPY (uninstall all various file operations). If not specified, the order is MOVE - STRSET - COPY - AT. If this is specified, then ALL operations must be specified (if any isn’t, a warning is printed). Please note that this means that your mod might start printing warnings three years down the line if a new uninstall action is coded. |
or | MODDER String list | Enable additional debug info. By default, these messages are verbose but do not halt the installation. Some options are configurable, see the MODDER in-depth description. |
or | VERSION String | "String" (variables evaluated) will be appended to every component’s name in the weidu.log file. |
or | SCRIPT_STYLE style | This flag determines how WeiDU will read in BAF and BCS files and write out BAF and BCS files. Possible options for “style” include BG (the default), IWD1, IWD2, and PST. See the Scripting Styles tutorial. |
or | NO_IF_EVAL_BUG | this action solves a long-standing bug with the IF_EVAL action. Since solving it is not directly possible without damaging backwards compatibility, you have to use this action (preferably in an ALWAYS statement) to solve the bug for yourself. IF_EVAL will not work with this action in the TP2 - use PATCH_IF and BUT_ONLY_IF_IT_CHANGES instead. |
Language | A Language declaration tells WeiDU where to find TRA files. | |
is | LANGUAGE languageName languageDirectory defaultLanguageTRA list | The languageName is the name of the language as it is presented to the user. "American English" and "Italiano" are examples. The languageDirectory is the name of the subdirectory in which you have stored the TRA files for that language. Examples include "american" and "spanish". The variable named LANGUAGE is set to languageDirectory if the user selects this language.. Finally, all of the TRA files in the defaultLanguageTRA list are loaded as soon as the user selects a language. |
Component | A Component is a contiguous group of files and actions that a user can install, uninstall or upgrade. | |
is | BEGIN componentName Component Flag list TP2 Action list | Basically, if componentName is "Foo", the user will be asked: "Do you want to install Foo?". If so, all of the associated TP2 Actions are performed. If not, they are skipped. |
Component Flag | A Component Flag determines how WeiDU treats a component. | |
is | DEPRECATED String | Mark the given component as deprecated. If it is currently installed, it will be uninstalled and the given String will be displayed. The user will never be asked to install the given component -- it will be silently skipped in all listings. However, it will still take up a “component number”. |
or | REQUIRE_COMPONENT modToUninstall modComponent String | Make this component so that it can only be installed if another component is installed. If that other component is not installed, the String will be displayed and the user will not get a chance to install this component. This is in some sense the opposite of the UNINSTALL TP2 Action. For example, REQUIRE_COMPONENT "setup-ease.tp2" "0" "You must have infinite stacking installed!" prevents a component from being installed unless the infinite stacking part of the Ease-of-Use mod is installed. |
or | FORBID_COMPONENT modToUninstall modComponent String | Make this component so that it can only be installed if another component is not installed. This does the opposite of REQUIRE_COMPONENT. |
or | REQUIRE_PREDICATE value String | This component can only be installed if the value evaluates to true (non-zero). |
or | SUBCOMPONENT String [ value ] | At most one component of the given subcomponent group can be installed at any time. All subcomponents of the same group are listed together for the user. See the SUBCOMPONENT tutorial. |
or | FORCED_SUBCOMPONENT String [ value ] | See the SUBCOMPONENT tutorial. |
or | GROUP String | Used for grouping components; the user will be asked to display these or not. See the GROUP tutorial. |
or | INSTALL_BY_DEFAULT | If WeiDU would ask the user whether to install this component or not, and this component is not already installed, WeiDU will instead install it by default (without asking the user). If there is an error or the component is already installed, WeiDU will ask the user. The --uninstall command-line argument overrides this. See also REQUIRE_COMPONENT and ALWAYS. |
or | DESIGNATED forcedNumber | Normally module components are numbered based on their order in the TP2 file (starting from 0). This flag sets the current component number to forcedNumber. The next component (if it lacks a DESIGNATED flag) will be forcedNumber+1. You can easily shoot yourself in the foot by setting forcedNumber too low (e.g., so that multiple components have the same number). Do not use this flag. |
or | NO_LOG_RECORD | Normally all module components are recorded in WeiDU.log and can be uninstalled later. This component flag prevents this component from writing a log record when it is successfully installed. As a result it is “invisible” to WeiDU, can be installed multiple times, and cannot be uninstalled with WeiDU. Do not use this flag. |
TP2 Action | A TP2 Action tells WeiDU how to install a component. This usually involves copying files and writing in new string references. | |
is | COPY optNoBackup optGlob fromFile toFile ... patch list when list | You may specify as many fromFile-toFile pairs as you like. Each
fromFile is copied to its associated toFile. If there are any WeiDU
variables inside explicit %s in toFile or fromFile, they are
replaced by their values. All of the
patches are applied. If there are any when
conditions and any of them are false, the copy does not happen.
A typical example is COPY "mymod/sword.itm"
"override/m#sword.itm". COPY commands set the user-defined SOURCE_DIRECTORY, SOURCE_FILESPEC, SOURCE_FILE, SOURCE_RES, SOURCE_EXT, DEST_DIRECTORY, DEST_FILESPEC, DEST_FILE,DEST_RES and DEST_EXT variables based on fromFile and toFile as follows. If fromFile is mymod/cre/bigboss.cre, then SOURCE_DIRECTORY is mymod/cre, SOURCE_FILESPEC is mymod/cre/bigboss.cre, SOURCE_FILE is bigboss.cre, SOURCE_RES is bigboss and SOURCE_EXT is cre. The DEST_ variables are similarly based on toFile. In addition, SOURCE_SIZE is set to the size (in bytes) of the source file. This is generally only useful if you have enabled globbing. Any user-defined variables in toFile are replaced with their values. You may also reference these variables in patches. See the Module Distribution section for information about finding a good unique prefix for your mod-created resources. |
or | COPY_EXISTING optNoBackup fromFile toFile ... | Behaves like COPY except that the fromFiles are drawn from the game BIFFs or override directory. This is useful for making changes to files that other mods may have changed as well. |
or | COPY_EXISTING_REGEXP optNoBackup optGlob fromFileRegexp toDir ... patch list when list | Behaves like COPY_EXISTING except that fromFileRegexp may
contain regular expressions (regexps). All matching files in
the game BIFFs will be copied to the directory specified by
toDir. If GLOB is specified, matching files in override will also be patched and copied. If a file appears in both the BIFFs and the override folder, it will only be copied once. For example, if HARM.ITM is in the BIFFs and HARM2.ITM is in override, this code will copy and patch them both: COPY_EXISTING_REGEXP GLOB ~HARM.*.ITM~ ~override~ SAY // ... whatever |
or | COPY_LARGE optNoBackup optGlob fromFile toFile ... | Behaves like COPY except that the fromFiles can be of arbitrary size (the limit should be over a Gigabyte), but on the other side of the coin you can’t apply patches to the copying. |
or | COPY_RANDOM ( file1 list ) [ ( fileN list ) list ] patch list when list | This command works like COPY_EXISTING but the destination for any given souce file in the file1-list is some other different file in the file1-list. Similarly, the destination for any file in the fileN-list is some other file in the fileN-list. This allows you to randomly shuffle categories of game resources. |
or | COPY_ALL_GAM_FILES patch list when list | Copies Default.gam from the biff and the ones in the save/ and mpsave/ saved games, applying the list of patches. If there are any when conditions and any of them are false, the copy does not happen. NO BACKUP is applied to files in the save/ and mpsave/ directories. The savegame from the biffs is backuped as usual. |
or | MOVE fromFile toFile ... | You may specify as many fromFile-toFile pairs as you like. Each
fromFile is moved to its associated toFile. Wildcard support will
be added on request. Safety notes: when uninstalling, MOVE is restored first, then all generic actions are restored, then AT_*!UNINSTALL is handled. As such, do your AT_NOW, then do your COPY, then do your MOVEs (exception: if you MOVE for biffing purposes, it is save to call MAKE_BIFF after MOVE. Do not use the --make-biff command-line argument). |
or | COMPILE [ EVALUATE_BUFFER ] sourceFile list [ patch list ] [ USING traFile list ] | This command compiles D and BAF source files. If
sourceFile is a directory, all D and BAF files within
that directory are processed individually. If there is EVALUATE_BUFFER,
all %variables% in the files are substituted with their values.After the special EVALUATE_BUFFER is executed, all other patches are executed. First, this loads all of the traFiles presented. If any of their paths contain %s, the %s is replaced with the languageDirectory of from the Language the user selected. If you specified AUTO_TRA mymod/%s above, WeiDU will also attempt to load mymod/languageDirectory/sourceFile.tra for every sourceFile in the list. Once all of the TRA files are loaded, the D and BAF files are compiled. Any DLGs or BCSs they create or modify are placed in the override directory. |
or | CLEAR_MEMORY | removes all variables from the memory, then reloads the automatic ones (TP2_AUTHOR, TP2_FILE_NAME, LANGUAGE, WEIDU_ARCH, WEIDU_OS, COMPONENT_NUMBER, all numeric constants such as NAME1, the soundsets, or SCRIPT_OVERRIDE). INTERACTIVE is automatically set when you call the next component. Don’t use this feature. |
or | CLEAR_IDS_MAP | force WeiDU to remove all loaded IDSes in memory (otherwise, if you compile a script, append to an IDS and compile another script, WeiDU will not handle the new IDS entry). |
or | ACTION_CLEAR_ARRAY String | ‘Forgets’ that the array string exists until its values are recalled. The actual variables’ values are kept, the only effect is that *PHP_EACH will not consider old values. |
or | SILENT | makes WeiDU skip all messages; it’s reenabled once a component is installed (or fails to), or a PRINT, PATCH_PRINT, VERBOSE or PATCH_VERBOSE is found. |
or | VERBOSE | undoes the latest SILENT command. |
or | MKDIR dirName list | Instructs WeiDU to create all of the directories in the list. |
or | RANDOM_SEED someInteger | Sets the random number generator seed to someInteger. This allows you to get reproducible results when using random functions. If you specify a string that is not a valid integer the system initializes itself (e.g., by using the current time). |
or | ACTION_READLN variable | Waits for the user to provided an enter-terminated string and store it in variable. Said string will be stored and re-used at reinstall time. See the READLN tutorial. |
or | APPEND filename newText when list [ KEEP_CRLF ] | If there are no when conditions or they are all true, the
ASCII text newText is appended to the existing file filename (which
is read from the game BIFFs or the override folder).
Any variables in newText are replaced by their values. If KEEP_CRLF is not present, double newlines in the file (but not in newText) will be removed; otherwise, they’re kept. In any event, if the file ends with CRLF only newText and an ending CRLF will be written; otherwise, if the file does not end with CRLF, it will be written before newText. |
or | APPEND_OUTER filename newText when list [ KEEP_CRLF ] | If there are no when conditions or they are all true, the
ASCII text newText is appended to the existing file filename (which
is treated as a path to a file).
Any variables in newText are replaced by their values. If KEEP_CRLF is not present, double newlines in the file (but not in newText) will be removed; otherwise, they’re kept. In any event, if the file ends with CRLF only newText and an ending CRLF will be written; otherwise, if the file does not end with CRLF, it will be written before newText. |
or | APPEND_COL filename newText [ prepend ] when list | If there are no when conditions or they are all true, the
string newText is appended column-wise to the existing file filename.
If filename was:
A B C D E F X Y Z P Q R and newText was "0 1 2 3", the result would be: A B C 0 D E F 1 X Y Z 2 P Q R 3 You must have the same number of whitespace-separated words in newText as there are columns in filename. prepend empty cells are assumed to be at the beginning of newText. A cell containing exactly $ is printed as empty. |
or | EXTEND_TOP existingBCS newFile patch list | Loads existingFile (which may be BAF or BCS), prepends all of newBCS to the top of it, applies all of the patches, and then copies it to the override folder. User variables in the filenames existingFile and newFile are replaced by their values. Use EVALUATE_BUFFER if you want to evaluate variables inside the body of newFile before parsing it. |
or | EXTEND_BOTTOM existingBCS newFile patch list | As EXTEND_TOP, but the newFile file is put at the bottom of the existingBCS file. User variables in the filenames existingFile and newFile are replaced by their values. |
or | EXTEND_TOP_REGEXP existingBCSregexp newFile patch list | As EXTEND_TOP, but the newFile file is put at the bottom of the every BCS file that matches the regexp existingBCSregexp. |
or | EXTEND_BOTTOM_REGEXP existingBCSregexp newFile patch list | See EXTEND_TOP_REGEXP. |
or | ACTION_IF value THEN BEGIN TP2 Action list END [ ELSE BEGIN TP2 Action list END ] | If value evaluates to true (non-zero), the TP2 Actions in the THEN-branch are executed. Otherwise, if an ELSE-branch is present, its commands are executed. Otherwise nothing happens. |
or | AT_EXIT commandToRun [ EXACT ] | Whenever this component attemps to be installed, commandToRun
is executed. Variables (e.g., %LANGUAGE%) in the string
commandToRun are replaced by their values. Note that the command is
executed even if the component does not install correctly, so
AT_EXIT should probably be the last command in a component.
EXACT is specified, /
and \ will instead be preserved as they are. The most common usage is:
AT_INTERACTIVE_EXIT ~VIEW mymod\README.txt~This causes your README file to be displayed using a system appropriate viewer. Here’s a more complicated example that pulls up a language-specific README if one is available: ACTION_IF FILE_EXISTS ~mymod\README.%LANGUAGE%.txt~ THEN BEGIN AT_INTERACTIVE_EXIT ~VIEW mymod\README.%LANGUAGE%.txt~ END ELSE BEGIN AT_INTERACTIVE_EXIT ~VIEW mymod\README.txt~ END |
or | AT_INTERACTIVE_EXIT commandToRun [ EXACT ] | As AT_EXIT, but the command is only executed if the user specifically asked for the component to be installed or upgraded. |
or | AT_UNINSTALL commandToRun [ EXACT ] | As AT_EXIT, but when this component is removed, commandToRun is executed. |
or | AT_INTERACTIVE_UNINSTALL commandToRun [ EXACT ] | As AT_EXIT, but whenever the user specifically asks for this component to be removed, commandToRun is executed. Only the %LANGUAGE% variable is guaranteed to be replaced, so do not count on any others. |
or | AT_NOW commandToRun [ EXACT ] | As AT_EXIT, but commandToRun is executed when found, rather than at the end of the installation. Use with oggdec, tisunpack, wav2acm... but not with readmes (you’ll create dozens of windows), nor with --make-biff (you’ll crash the installation). |
or | AT_INTERACTIVE_NOW commandToRun [ EXACT ] | As AT_NOW, but commandToRun is executed only if the installation is run interactively. |
or | MAKE_BIFF name-of-biff BEGIN directory-file-regexp list END | Create a biff Data/name-of-biff.bif from the files that can be matched by directory-file-regexp. The chitin.key file is restored as a normal file, and reloaded whenever the component is installed or uninstalled (which means that its contents will be available exactly when the biff exists), while the .bif file itself is NOT deleted on uninstall. |
or | LOAD_TRA traName list | Loads all traName files (variable substitution is done, the file cannot be inlined) to be used in the following tp2 actions, as if you had declared them in LANGUAGE. |
or | UNINSTALL modToUninstall modComponent | If the given component of the given mod is currently installed, uninstall it before proceeding. Do not use this action. This should only be used if you release a new version of a component under a new name. For example, many Tactics Mod components replace old Solaufein mod components. In order to prevent such a component from being installed twice, the Tactics version uninstalls the Solaufein version. |
or | COPY_KIT oldKit newKit ( diffName diffValue list ) | Copies a copy of the oldKit kit called newKit. This copy is non-player
selectable, while all other fields can be edited by setting diffName to
the field to be changed and diffValue to the new value. Valid values for
diffName are:
clasweap weapprof abclasrq abclsmod abdcdsrq abdcscrq dualclas alignmnt clab lower mixed help luabbr 25stweap unusabilitiesCurrently, lower, mixed and help can only be numbers referring to a TLK string. Unusabilities requires to be both kit and class unusables, there isn’t a way to BAND or BOR them. |
or | ADD_KIT internalKitName manyComplexArguments | This command allows you to add new kits to the BGII. See the example file mymod.tp2 or the tutorial at http://forums.gibberlings3.net/index.php?showtopic=584 for information on how to do this. |
or | ADD_MUSIC internalMusicName newMUSFile | No documentation yet! |
or | ADD_PROJECTILE modpath/PROName.PRO | Appends an entry for PROName to PROJECTL.IDS and assigns it the next
available ProRef number. Then copies the file modpath/PROName.PRO
to the override folder. The new ProRef number can be accessed
through the variable %PROName% and used to updated the
Projectile type field of an ITM or SPL file’s Item Ability or
Spell Ability sub-structures. (The hexadecimal offsets for these
fields can be found using NearInfinity.) In the following example, a
new PRO file and an ITM file that will use it are added to the game:
ADD_PROJECTILE ~MyMod/MYDXP.PRO~ COPY ~MyMod/MYDART.ITM~ ~override/MYDART.ITM~ WRITE_SHORT 0x09c ~%MYDXP%~ |
or | ADD_SPELL newSPLfile type level idsName patch list when list | adds newSPLfile as a spell of type and level into the game (IE, SPPR102 or
whatnot), in the first empty slot possible, appends the idsName to
spell.ids, and sets %idsName%
to the 4-digit code associated with idsName in Spell.ids. While you copy
newSPLfile to the override, you also apply the patches listed. |
or | STRING_SET indexOrString newValue list [ USING traFile ] | This command replaces each given string in the user’s TLK file with the associated newValue. Do not use this command. If a traFile is given, that file is is read once before all of the replacements take place and its contents are forgotten after. Do not use this command. |
or | STRING_SET_EVALUATE value newValue list [ USING traFile ] | This command replaces the string at index value in the user’s TLK file with the associated newValue. Do not use this command. If a traFile is given, that file is is read once before all of the replacements take place and its contents are forgotten after. Do not use this command. |
or | STRING_SET_RANGE #min #max USING traFile | For every integer i between min and max (inclusive) we do STRING_SET i @i USING traFile (except that this command should be executed more rapidly). The command will fail if @i is not defined (either by traFile or by some other tra file in scope) for some i between min and max. Do not use this command. |
or | ALTER_TLK_RANGE value value BEGIN patch list END | Executes the given patch list on all the strings *currently* in dialog.tlk between the two given values (both inclusive). This means that strings added in the current mod installation will not be patched, and strings already altered via ALTER_TLK* and/or STRING_SET* will be patched from the pre-patch condition. DNUT. |
or | ALTER_TLK_LIST BEGIN value list END BEGIN patch list END | Executes the given patch list on all the strings *currently* in dialog.tlk from the given list. This means that strings added in the current mod installation will not be patched, and strings already altered via ALTER_TLK* and/or STRING_SET* will be patched from the pre-patch condition. DNUT. |
or | ALTER_TLK BEGIN patch list END | Executes the given patch list on all the strings *currently* in dialog.tlk. This means that strings added in the current mod installation will not be patched, and strings already altered via ALTER_TLK* and/or STRING_SET* will be patched from the pre-patch condition. DNUT. Consider using ALTER_TLK_RANGE with the base game’s boundaries instead. |
or | REQUIRE_FILE filename warningString | If filename does not exist (or the bif file is not referenced inside the chitin.key), warningString is displayed and this component cannot be installed. This is checked before any actions are executed. |
or | FORBID_FILE filename warningString | If filename does exist (or the bif file is referenced inside the chitin.key), warningString is displayed and this component cannot be installed. This is checked before any actions are executed. |
or | FAIL warningString | If this TP2 Action is execution, warningString is displayed and the component fails to install. |
or | PRINT displayString | The string DisplayString is echoed to the user. Useful for debugging or status reports. If displayString contains %variable% references, their values will be displayed. |
or | OUTER_SPRINT variable stringWithVars | Any WeiDU variables (enclosed in %s) inside stringWithVars are replaced by their values and the resulting string (constructed at mod-installation time!) is assigned to the variable variable. This works somewhat like sprintf() but it not as cool. |
or | OUTER_SET variable = value | Update variable so that it is equal to value. |
or | OUTER_SET variable += value | Equivalent to OUTER_SET variable = variable + value. |
or | OUTER_SET variable -= value | Equivalent to OUTER_SET variable = variable - value. |
or | OUTER_SET variable *= value | Equivalent to OUTER_SET variable = variable * value. |
or | OUTER_SET variable /= value | Equivalent to OUTER_SET variable = variable / value. |
or | OUTER_SET variable &= value | Equivalent to OUTER_SET variable = variable BAND value. |
or | OUTER_SET variable |= value | Equivalent to OUTER_SET variable = variable BOR value. |
or | OUTER_SET variable <<= value | Equivalent to OUTER_SET variable = variable BLSL value. |
or | OUTER_SET variable >>= value | Equivalent to OUTER_SET variable = variable BLSR value. |
or | OUTER_INNER_PATCH buffString BEGIN patch list END | Any WeiDU variables inside %s within buffString are replaced by their values. All of the patches given are evaluated as if the contents of the current file were buffString. Any modifications to buffString are thrown away. This is considered an action, not a patch. |
or | OUTER_PATCH buffString BEGIN patch list END | Equal to OUTER_INNER_PATCH, if you feel that OUTER_INNER sounds wrong. |
or | OUTER_INNER_PATCH_SAVE savevar buffString BEGIN patch list END | As with OUTER_INNER_PATCH, except that any modifications to buffString are stored inside savevar. |
or | OUTER_PATCH_SAVE savevar buffString BEGIN patch list END | Equal to OUTER_INNER_PATCH_SAVE, if you feel that OUTER_INNER sounds wrong. |
or | OUTER_WHILE value BEGIN TP2 Action list END | If value is non-zero, execute the given TP2 Action list and then repeat, re-evaluating the value. Be very careful when using this command. You can easily describe an infinite loop. See the WHILE loop tutorial for more information. |
or | OUTER_FOR ( patch list ; value ; patch list ) BEGIN TP2 Action list END | The TP2 Action OUTER_FOR (init;pred;inc) BEGIN body END is equivalent to init WHILE pred BEGIN INNER_ACTION BEGIN body END inc END, except that it’s considered an action. Note that the predicate value cannot be empty. |
or | ACTION_BASH_FOR directory-file-regexp BEGIN TP2 Action list END | for all files that match directory-file-regexp, sets a bunch of variables
and executes the actions for each file found. Assuming the file being copied
is somedir/yourfile.cre, the following variables will be set:
"%BASH_FOR_DIRECTORY%" = "somedir" "%BASH_FOR_FILESPEC%" = "somedir/yourfile.cre" "%BASH_FOR_FILE%" = "yourfile.cre" "%BASH_FOR_RES%" = "yourfile" "%BASH_FOR_SIZE%" = <size of somedir/yourfile.cre> |
or | ACTION_DEFINE_ARRAY String1 BEGIN String2 list END | Sets the array $string1(0) ,$string1(1) etc. to the various elements in string2. |
or | ACTION_FOR_EACH String1 IN String list BEGIN TP2 Action list END | will set the string1 variable to each value in string list and process each action. |
or | ACTION_PHP_EACH String1 AS String2 => String3 list BEGIN TP2 Action list END | For every value of the string1 array that has been created or read, sets
string2_0 to the first array parameter, string2_1 to the second array
parameter, etc. and string3 to the result of the array. string2 will also contain
the first array parameter (as a synonym of string2_0 ). then executes the
actions listed. More exaustive documentation will be provided by SConrad. |
or | INCLUDE String list | loads the file string as if it were a list of tp2 actions and executes it (as if it had been typed inside your tp2). These files have by convenience tph as an extension, due to no reason other than randomness (you can use .tp1 or .xyzsucks if you’re so oriented). Please note that, to speed up inclusions, tph files are stored in memory, rather than re-read each time from the hard-disk. If you need automorphing code (IE, you want to edit your tph file), you can use REINCLUDE instead. |
or | ACTION_INCLUDE String list | A synonym of INCLUDE. |
or | REINCLUDE String list | Works like INCLUDE, but loads each time the tph file from the hard disk. This is slower, but required if you want to edit your tph file between an inclusion and another. |
or | ACTION_REINCLUDE String list | A synonym of INCLUDE. |
or | ADD_GAM_NPC npcCRE npcARE xCoord yCoord | See the ADD_GAM_NPC tutorial for more information about this action, which is used when adding NPCs to Baldur’s Gate 1. BG2 mods should not use this command. |
or | <<<<<<<< fileName fileBody >>>>>>>> | (That’s eight (8) angle brackets on each side.)
For the purposes of copying and compiling files, WeiDU will pretend
that fileName is a real file with contents fileBody. This allows you to
define inlined files inside your TP2 file. Such definitions
must be executed before the inlined file is used. Other operations
on fileName (such as FILE_EXISTS, FILE_MD5 or
FILE_SIZE) are undefined. Inlined files will be skipped by
COPY_EXISTING_REGEXP and other “wildcard” approaches: you
must name them directly. Unlike most other WeiDU things,
spacing matters here. After the initial <<<<<<<< there can be
any number of spaces. The first non-space character after the
<<<<<<<< is the first character of fileName. All other
characters up to and excluding the newline on that line are part of
fileName (and thus fileName cannot start with a space or contain a
newline). All user variables (e.g., %foo%) in fileName will be
replaced by their values.
The fileBody contains all characters after that newline up to
and excluding the <<<<<<<<. Note that a single inlined filename
namespace is shared by WeiDU for all TP2 files it reads (it
might read other TP2 files in the process of installing yours
in order to uninstall or reinstall another mod’s components), so it is
criticially important that you use some sort of unique prefix.
I also suggest that you avoid using your mod’s actual directory
structure (if any) in order to avoid confusion with real files.
Consider using .../yourmod-inlined/fileName. Here is a concrete
working example:
BEGIN ~MyMod Component 1~ <<<<<<<< .../mymod-inlined/myfile.baf IF True() THEN RESPONSE #100 Kill("Anomen") END >>>>>>>> COMPILE ~.../mymod-inlined/myfile.baf~This inclusion method is eight-bit clean, so if you are very careful you can inline a binary file (e.g., a CRE or BAM) with this method. Be careful to shave off the newline before the >>>>>>>> in such cases. Finally, note that inlined files have the same maximum size as other (non-BIFF) WeiDU files and strings (usually about 16 megs). |
optNoBackup | A COPY command normally makes a backup copy of its target (if one exists) so that the mod can be uninstalled later by restoring the backup. | |
is | If you don’t say anything here, WeiDU will make a backup copy of the file and the COPY will be undone if the mod is uninstalled. | |
or | + | If you put a + here, WeiDU will not make a backup copy of the file and the COPY will not be undone if the mod is uninstalled. Do not use this feature. |
or | - | If you put a - here, WeiDU will not copy the file, but only store it as an inlined file (so you can patch a file and then use it for EXTEND_BOTTOM or whatever. This option is not available for COPY_LARGE. |
directory-file-regexp | This is a directory path and a regexp string. It’s used by MAKE_BIFF, ACTION_BASH_FOR and PATCH_BASH_FOR for listing patterns to be matched. Please note that the directory / filename distinction must be enforced. | |
is | directory regexp | get a list of all files in directory that match regexp. Case doesn’t matter. |
or | directory EVALUATE_REGEXP regexp | get a list of all files in directory that match regexp. Case doesn’t matter. This is the same as the above item. |
or | directory EXACT_MATCH name | get the file directory/name. Please note that the directory / filename distinction must be enforced. |
optGlob | A COPY command may use globbing to expand filename wildcards with respect to files on the host filesystem. Unlike COPY_EXISTING_REGEXP, glob wildcards do not range over game resources. Instead, they range over physical files actually on the disk. | |
is | Do not use local filesystem globbing. This is the default. | |
or | GLOB | Use local filesystem globbing. Globbing is
generally architecture specific! Do not use globbing if you can
help it. Here is a concrete example. Imagine that CHITIN.KEY contains two files: C1.ITM and C2.ITM. The override contains C2.ITM and C_MOD.ITM. COPY_EXISTING_REGEXP "C.*.ITM" "override" // catches C1.ITM, C2.ITM COPY_EXISTING_REGEXP GLOB "C.*.ITM" "override" // catches C1.ITM, C2.ITM, C_MOD.ITM COPY "override" "override" // catches C2.ITM, C_MOD.ITM To put it another way: if you do not specify GLOB with COPY_EXISTING, WeiDU pretends that the override directory contains 0 files that are not in CHITIN.KEY. Finally, GLOB has no effect on wildcards in the "destination" part of the COPY: COPY GLOB "mymod/foo" "music/*" // this is illegal: DO NOT DO THIS |
optcase | Allows you to decide whether regexp matching is case-sensitive or not. | |
is | Uses the WeiDU default (which is undocumented and varies with the action). Please specify case-sensitiveness manually. | |
or | CASE_SENSITIVE | matching is case-sensitive. |
or | CASE_INSENSITIVE | matching is case-insensitive. |
optexact | Allows you to decide whether regexp matching is exact. | |
is | Uses the WeiDU default (which is undocumented and varies with the action). Please specify matching policy manually. | |
or | EXACT_MATCH | only the given string is searched for. |
or | EVALUATE_REGEXP | matching is evaluated following the usual regexp conventions. |
patch | A patch tells WeiDU how to modify a file. | |
is | SAY offset String | The string-ref associated with String is written at offset. This is commonly used to change the name or description of a spell or item. |
or | PATCH_PRINT displayString | The string DisplayString is echoed to the user. Useful for debugging or
status reports. If displayString contains %variable%
references, their values will be displayed. See also PRINT.
Example:
COPY_EXISTING_REGEXP ~.*\.CRE~ ~override~ READ_BYTE 0x272 race READ_BYTE 0x273 class PATCH_IF class = 3 THEN BEGIN PATCH_PRINT ~%SOURCE_FILE% is a cleric with race = %race%.~ END |
or | SAY_EVALUATED offset stringWithVars | Any WeiDU variables (enclosed in %s) inside stringWithVars are
replaced by their values and the resulting string (constructed at
mod-installation time!) is added to DIALOG.TLK and its string
reference it written to the offset. Example:
COPY_EXISTING_REGEXP ~RING.*.ITM~ ~override~ READ_LONG 0x38 cost SAY_EVALUATED IDENTIFIED_DESC ~I Am %SOURCE_RES%, I Cost %cost%~Do not use this feature. |
or | TO_UPPER variable | Turns to upper-case the contents of %variable% and store the
results in variable. |
or | TO_LOWER variable | Turns to lower-case the contents of %variable% and store the
results in variable. |
or | SPRINT variable stringWithVars | Any WeiDU variables (enclosed in %s) inside stringWithVars are replaced by their values and the resulting string (constructed at mod-installation time!) is assigned to the variable variable. This works somewhat like sprintf() but it not as cool. Currently this is the only way to assign a string value to a variable. |
or | SNPRINT value variable stringWithVars | As SPRINT, but only the first N characters are stored in
the variable, where N is the result of evaluating the
value. This works somewhat like snprintf(). Thus:
SPRINT "author" "Jason" SNPRINT 3 "myvar" "1:%author%"... assigns 1:J to myvar. |
or | SPACES variable stringWithVars | %variable% will be inited to a string the same length as stringWithVars,
but made entirely of spaces. SPACES var ~123~ will create %var% = " " . |
or | QUOTE variable stringWithVars | %variable% will be inited to a string that will match exactly stringWithVars.
QUOTE var ~some\thing^~ will create %var% = "some\\thing\^" . |
or | REPLACE optcase optexact regexp text | All occurences of regexp in the file are replaced with the ASCII printing of the string reference for text. So if regexp is "FRED" and the text ends up being strref #1234, "FRED" will be replaced with "1234". This is usually used to replace string references in BCS files (where they are stored textually). Put a command like DisplayString(Myself,99999) in your BCS file and use something like REPLACE 99999 "Hello, World". |
or | REPLACE_TEXTUALLY optcase optexact regexp string [ ( sizeValue ) ] | All occurences of the given regexp in the file are replaced with the given string. variable substitution (e.g., kit and music names) is performed on both the string and the regexp. If you use the sizeValue field, optexact is automatically set to true (IE without regexp patterns). The regexp and the string are padded by zeros to be long exactly sizeValue bytes. |
or | EVALUATE_BUFFER_SPECIAL string | String must be one character long. Any WeiDU variables (like XmyvarX)
inside the current file are replaced by their values, where X is to be intended
as string. Example:
<<<<<<<< .../script.baf IF See($myvar$) THEN RESPONSE #100 Kill(%myvar%) END >>>>>>>> EXTEND_TOP ~sola.bcs~ ~.../script.baf~ SPRINT myvar = ~"Anomen"~ EVALUATE_BUFFER_SPECIAL ~$~myvar will be substituted with Anomen, but %myvar% will be not. DNUT. |
or | EVALUATE_BUFFER | Any WeiDU variables (like %myvar%) inside the current file (which
should probably be a plain text file in order for this to make much sense)
are replaced by their values. Example:
<<<<<<<< .../script.baf IF See(%myvar%) THEN RESPONSE #100 Kill(%myvar%) END >>>>>>>> EXTEND_TOP ~sola.bcs~ ~.../script.baf~ SPRINT myvar = ~"Anomen"~ EVALUATE_BUFFERThose two actions extend the top of sola.bcs with the script block IF See("Anomen") THEN RESPONSE #100 Kill("Anomen") END. You can also use EVALUATE_BUFFER in COMPILE actions or before strings in values. |
or | APPLY_BCS_PATCH patchFile | Applies patchFile to the current file. See --bcmp-from and similar command-line arguments for constructing these patches. |
or | APPLY_BCS_PATCH_OR_COPY patchFile copyFile | Applies patchFile to the current file, as APPLY_BCS_PATCH. However, if the patching fails the current file is replaced with copyFile instead. |
or | WRITE_BYTE offset value | The first argument is the offset at which the second argument (an 8-bit
byte value) is written. Performs READ_BYTE offset THIS and
READ_SBYTE offset STHIS automatically (you can now write
WRITE_BYTE 100 THIS 2). |
or | WRITE_SHORT offset value | The first argument is the offset at which the second argument (a 16-bit
short value) is written. Performs READ_SHORT offset THIS and
READ_SSHORT offset STHIS automatically. |
or | WRITE_LONG offset value | The first argument is the offset at which the second argument (a 32-bit
long word value) is written. Performs READ_LONG offset THIS and
READ_SLONG offset STHIS automatically. |
or | WRITE_ASCII offset ascString [ #requiredSize ] | The ASCII ascString is written to the file starting at offset. If you specify a requiredSize then exactly that many bytes are written (if ascString is smaller, it is padded with NULs; if ascString is larger, it is truncated). If you do not specify a requiredSize, the terminating NUL is not written. |
or | WRITE_ASCII_LIST offset ascString list | The given ASCII strings are written at offset, offset + 8, offset + 16... Automatically evaluates variables. |
or | WRITE_ASCII_TERMINATE offset ascString | The ASCII ascString is written to the file starting at offset. The terminating NUL is written. You may use WRITE_ASCIIT as a synonym. |
or | WRITE_EVALUATED_ASCII offset ascString [ #requiredSize ] | The ASCII ascString is evaluated (so %variable% is replaced by its value) and written to the file starting at offset (as in WRITE_ASCII). You may use WRITE_ASCIIE as a synonym. |
or | WRITE_FILE offset filename | The entire contents of “filename” (which may contain variables) are loaded and copied over the current file starting at offset offset. “filename” must be a literal filename like mymod/data/file.bam. If there is not enough room between offset and the end of the file for the contents of “filename” the patch will fail with an error message. |
or | INSERT_FILE offset filename | Just like WRITE_FILE except that the entire contents of “filename” are inserted at offset, just as if you had done an INSERT_BYTES with the size of “filename” to that offset followed by a WRITE_FILE to that offset. |
or | APPEND_FILE filename | Just like INSERT_FILE except that the entire contents of “filename” are appended to the end of the current file. |
or | APPEND_FILE_EVALUATE filename | Just like APPEND_FILE except that %variables% in the contents of
“filename” are evaluated before the appending. |
or | REPLACE_BCS_BLOCK oldFile newFile | If the current file is a BCS file, the segment of it corresponding to oldFile is replaced with the contents of newFile. oldFile and newFile may be BCS or BAF files. If they are BAF files they will not get the benefit of AUTO_TRA. |
or | INSERT_BYTES offset value | The first argument is the offset, the second argument is the count. The file will be expanded at the given offset with count bytes worth of zeroes. |
or | DELETE_BYTES offset value | The first argument is the offset, the second argument is the count. The file will shrink as count bytes are deleted starting at the given offset. |
or | READ_BYTE offset variable [ ELSE value ] | An 8-bit value is read from the file at the given offset and is stored in the given variable. If offset is out-of-bounds and the ELSE is present, the ELSE-value is assigned to variable. If offset is out-of-bounds and the ELSE is not present, the patch fails with a visible error. |
or | READ_SBYTE offset variable [ ELSE value ] | As READ_BYTE, but the value is interpreted as signed. |
or | READ_SHORT offset variable [ ELSE value ] | A 16-bit value is read from the file at the given offset and is stored in the given variable. See READ_BYTE. |
or | READ_SSHORT offset variable [ ELSE value ] | As READ_SHORT, but the value is interpreted as signed. |
or | READ_LONG offset variable [ ELSE value ] | A signed 32-bit value is read from the file at the given offset and is stored in the given variable. See READ_BYTE. |
or | READ_SLONG offset variable [ ELSE value ] | As READ_SLONG, but the value is interpreted as signed. |
or | READ_ASCII offset variable [ ELSE string ] [ ( value ) [ NULL ] ] | A nul-terminated string is read from the file at the given offset and is stored in the given variable. The terminating nul is not stored. The default read size is 8 bytes. If an explicit size value is specified then that many bytes are read into the variable, even if some of them are nuls, except if NULL is present, in which case a nul-terminated string is read, without storing the terminating null. See READ_BYTE. If the offset is out-of-bounds and the ELSE clause is present, the string is evaluated as in WRITE_EVALUATED_ASCII and then assigned into variable. |
or | READ_STRREF offset variable [ ELSE string ] | A 32-bit Infinity Engine DIALOG.TLK string reference is read from the file at the given offset. The string reference is looked up in DIALOG.TLK and the (male) string value for it (without any quotes) is stored in the variable. In some sense this is the opposite of SAY_EVALUATED. |
or | GET_OFFSET_ARRAY String seven values | This is too complicated to explain with words. SConrad owes documentation. |
or | GET_OFFSET_ARRAY2 String eight values | This is too complicated to explain with words. SConrad owes documentation. |
or | GET_STRREF value variable | The string reference value is looked up in DIALOG.TLK and the (male) string (without any quotes) is stored in the variable. |
or | SET variable = value | Update variable so that it is equal to value. |
or | variable = value | Update variable so that it is equal to value. |
or | [ SET ] variable += value | Equivalent to SET variable = variable + value. |
or | [ SET ] variable -= value | Equivalent to SET variable = variable - value. |
or | [ SET ] variable *= value | Equivalent to SET variable = variable * value. |
or | [ SET ] variable /= value | Equivalent to SET variable = variable / value. |
or | [ SET ] variable &= value | Equivalent to SET variable = variable BAND value. |
or | [ SET ] variable |= value | Equivalent to SET variable = variable BOR value. |
or | [ SET ] variable <<= value | Equivalent to SET variable = variable BLSL value. |
or | [ SET ] variable >>= value | Equivalent to SET variable = variable BLSR value. |
or | WHILE value BEGIN patch list END | If value is non-zero, execute the given patch list and then repeat, re-evaluating the value. Be very careful when using this command. You can easily describe an infinite loop. See the WHILE loop tutorial for more information. |
or | FOR ( patch list ; value ; patch list ) BEGIN patch list END | The patch FOR (init;pred;inc) BEGIN body END is equivalent to init WHILE pred BEGIN body inc END. Note that the predicate value cannot be empty. |
or | PATCH_BASH_FOR directory-file-regexp BEGIN patch list END | for all files that match directory-file-regexp, sets a bunch of variables
and executes the patches for each file found. Assuming the file being copied
is somedir/yourfile.cre, the following variables will be set:
"%BASH_FOR_DIRECTORY%" = "somedir" "%BASH_FOR_FILESPEC%" = "somedir/yourfile.cre" "%BASH_FOR_FILE%" = "yourfile.cre" "%BASH_FOR_RES%" = "yourfile" "%BASH_FOR_SIZE%" = <size of somedir/yourfile.cre> |
or | DEFINE_ARRAY String1 BEGIN String2 list END | Sets the array $string1(0) ,$string1(1) etc. to the various elements in string2. |
or | PATCH_DEFINE_ARRAY String1 BEGIN String2 list END | Same as DEFINE_ARRAY. |
or | PATCH_FOR_EACH String1 IN String list BEGIN patch list END | will set the string1 variable to each value in string list and process each patch. |
or | PHP_EACH String1 AS String2 => String3 list BEGIN patch list END | For every value of the string1 array that has been created or read, sets
string2_0 to the first array parameter, string2_1 to the second array
parameter, etc. and string3 to the result of the array. string2 will also contain
the first array parameter (as a synonym of string2_0 ). then executes the
patches listed. More exaustive documentation will be provided by SConrad. |
or | PATCH_PHP_EACH String1 AS String2 => String3 list BEGIN patch list END | A synonym of PHP_EACH. |
or | CLEAR_ARRAY String | ‘Forgets’ that the array string exists until its values are recalled. The actual variables’ values are kept, the only effect is that *PHP_EACH will not consider old values. |
or | PATCH_CLEAR_ARRAY String | A synonym of CLEAR_ARRAY. |
or | PATCH_IF value [ THEN ] BEGIN patch list END [ ELSE BEGIN patch list END ] | If value is non-zero, execute the first patch list once. Otherwise, execute the second patch list (if any). As a convenient shorthand, you may omit the BEGIN-END in the ELSE branch if the ELSE branch contains exactly one patch. |
or | PATCH_INCLUDE String list | loads the file string as if it were a list of tp2 patches and executes it (as if it had been typed inside your tp2). These files have by convention tpp as an extension, due to no reason other than randomness (you can use .tp1 or .xyzsucks if you’re so oriented). Please note that, to speed up inclusions, tpp files are stored in memory, rather than re-read each time from the hard-disk. If you need automorphing code (IE, you want to edit your tpp file), you can use PATCH_REINCLUDE instead. |
or | PATCH_REINCLUDE String list | Works like PATCH_INCLUDE, but loads each time the tph file from the hard disk. This is slower, but required if you want to edit your tpp file between an inclusion and another. |
or | SET_2DA_ENTRY value value value value | The first value is the row, the second is the column and the third is the required column count. The entry on the given column of the given row is set to the fourth value, but only rows with at least as many columns as the required column count are considered. The fourth value, the new entry, is evaluated specially: if it can be evaluated like a value (e.g., ‘‘3+4’’) it will be evaluated and its integer result will be written as an ASCII string. Otherwise if it is a single string (that is not a variable in scope) that string will be written at the new value. See the SET_2DA_ENTRY tutorial for more information. |
or | SET_2DA_ENTRY_LATER string value value value | Similar to SET_2DA_ENTRY, memorizes the 2da changes in the RAM without flushing them to the file. See the SET_2DA_ENTRY_LATER tutorial. |
or | SET_2DA_ENTRIES_NOW string value | To be used with SET_2DA_ENTRY_LATER: flushes the file changes in one go. See the SET_2DA_ENTRIES_NOW tutorial. |
or | PRETTY_PRINT_2DA | Indents the 3rd line in the current file (which should be a 2da file). |
or | PRETTY_PRINT_2DA value | Indents the value’th line (counting from 0) in the current file (which should be a 2da file). |
or | INSERT_2DA_ROW value1 value2 String | String (after variable substitution) will be inserted as a new line in the table.
Value1 is the row count, value2 is the required column count. If row count is
the same as the number of long enough lines in the file, then the line will be inserted
at the end; if row count is lesser than the number of lines, it will be inserted just before
the nth line in the original file (starting count from 0); finally, if row count is greater
than the number of lines, the component will fail to install.
<<<<<<<< 2dafile asd foo a b c d e f g h i >>>>>>>> COPY ~2dafile~ ~2dafile~ INSERT_2DA_ROW 3 3 ~4 4 4~ INSERT_2DA_ROW 2 3 ~3 3 3~ INSERT_2DA_ROW 1 3 ~2 2 2~ INSERT_2DA_ROW 0 3 ~1 1 1~will result in asd foo 1 1 1 a b c 2 2 2 d e f 3 3 3 g h i 4 4 4 |
or | REMOVE_2DA_ROW value1 value2 | the Value1-th row from the 2da file containing at least value2 items will be removed. |
or | PATCH_READLN variable | Waits for the user to provided an enter-terminated string and store it in variable. Said string will be stored and re-used at reinstall time. See the READLN tutorial. |
or | PATCH_RANDOM_SEED value | See RANDOM_SEED. |
or | ADD_STORE_ITEM [ + ] itemName [ position ] numCharges ext2 ext3 itemFlag maxInStack [ unlimited ] | See the ADD_STORE_ITEM tutorial for more information. |
or | REMOVE_STORE_ITEM itmName list | Remove the listed items from the current store. Does nothing if an item is not for sale. Multiple instances are removed. |
or | READ_2DA_ENTRY value value value variable | The first value is the row, the second is the column and the third is the required column count. The variable specified is set to the the entry on the given column of the given row, but only column with at least as many columns as the required column count are considered. This is the reverse of SET_2DA_ENTRY. |
or | READ_2DA_ENTRIES_NOW string value | Reads in one go all the 2da in the RAM for READ_2DA_ENTRY_FORMER. See the READ_2DA_ENTRIES_NOW tutorial. |
or | READ_2DA_ENTRY_FORMER string value value variable | Similar to READ_2DA_ENTRY, but reads from a prepared READ_2DA_ENTRIES_NOW command rather than from the file. See the READ_2DA_ENTRY_FORMER tutorial. |
or | COUNT_2DA_ROWS value variable | The first value is the required column count. This command counts the number of rows in the current file (which should be a 2DA file) that have at least as many columns as the required column count and stores the result in the variable. |
or | COUNT_2DA_COLS variable | This command counts the number of columns in the current file (which should be a 2DA file) and stores the result in the variable. |
or | COUNT_REGEXP_INSTANCES optcase optexact regexp variable | This command counts the number of times regexp appears in the current file
and stores the result in the variable. Variable substitution is performed
on regexp prior to regexp handling (EG, if %var% = ~abc~ ,
~[%var%]~ is evaluated as ~[abc]~ . |
or | LOOKUP_IDS_SYMBOL_OF_INT variable idsFile value | The symbolic constant associated with value in idsFile (which may
contain user variables) is stored
in variable. If that doesn’t work, value is stored in variable.
Example:
LOOKUP_IDS_SYMBOL_OF_INT foo ~spell~ 1101 SPRINT myfile "SPELL" LOOKUP_IDS_SYMBOL_OF_INT bar ~%myfile%~ (0x44c + 1) LOOKUP_IDS_SYMBOL_OF_INT baz ~spell~ 77777Both foo and bar are CLERIC_BLESS while baz is 777777. |
or | COMPILE_BAF_TO_BCS | The current file, which must be a valid BAF script, is compiled to a BCS. In general you should use the COMPILE TP2 Action instead, unless you are using other patch commands to modify the file under consideration. |
or | DECOMPILE_BCS_TO_BAF | The current file, which must be a valid BCS script, is decompiled to a BAF. |
or | DECOMPILE_DLG_TO_D | The current file, which must be a valid DLG file, is decompile to a textual D file (with string refs and no comments). Once you have a D file you can use other patch commands to change the actions and triggers around. You should use D actions (like REPLACE_ACTION_TEXT instead whenever possible. |
or | COMPILE_D_TO_DLG | The current file, which must be a valid D file that defines a single DLG file (via an obvious BEGIN action) is compiled to a DLG. Typically this is only used after a DECOMPILE_DLG_TO_D. |
or | REPLACE_EVALUATE optcase findRegexp BEGIN patch list END replaceRegexp | For every instance of the regexp findRegexp found, the
patch list is evaluated (with the variable MATCHi set to
the ith matched group in findRegexp), variable substitute is performed
on replaceRegexp, and then findRegexp is replaced by replaceRegexp. Any
writes dones by the patch list (e.g., SAY or
WRITE_ASCII) are ignored: SET should be the main
component of the patch list. For example:
COPY ~nice.baf~ ~mean.baf~ REPLACE_EVALUATE ~Give(\([0-9]+\),\([0-9]+\))~ BEGIN SET "RESULT" = ("%MATCH1%" + "%MATCH2%") / 2 END ~Take(%RESULT%)~This COPY TP2 Action would replace Give(10,20) with Take(15). optcase allows you to decide if the matching is case-sensitive or not. |
or | ADD_MAP_NOTE xCoord yCoord color String | If the file currently being patched is an ARE area file, this
patch command adds a map note to it. Valid colors include: gray,
violent, green, orange, red, blue, darkblue, lightgray.
Example:
COPY_EXISTING ~ar0202.are~ ~override/ar0202.are~ ADD_MAP_NOTE #123 #777 ~violet~ ~This is my new map note! Yippee!~Special thanks to Japh for coding this feature. |
or | ADD_KNOWN_SPELL splName spellLevel spellType | When applied to a CRE file, this patch causes the given spell to
be known. Note that spellLevel counts from 0 (e.g., you should say
#2 for a third-level Fireball). Possible values for spellType are
priest, innate and wizard. Example:
COPY_EXISTING ~some.cre~ ~override/some.cre~ ADD_KNOWN_SPELL ~SPPR314~ #2 ~priest~ // Unholy Blight is now known as a 3rd level priest spellSpecial thanks to Japh for coding this feature. |
or | ADD_MEMORIZED_SPELL splName spellLevel spellType [ ( value ) ] | When applied to a CRE file, this patch causes the given spell to
be memorized one more time (or value times if specified).
Note that spellLevel counts from 0 (e.g., you
should say #2 for a third-level Fireball). Possible values for spellType
are priest, innate and wizard. Example:
COPY_EXISTING ~some.cre~ ~override/some.cre~ ADD_MEMORIZED_SPELL ~SPPR314~ #2 ~priest~ ( 5 ) // Unholy Blight is now memorized five times as 3rd priest |
or | REMOVE_KNOWN_SPELL splName list | When applied to a CRE file, this patch causes all of the
listed spells to be removed. Example:
COPY_EXISTING ~aerie.cre~ ~override/aerie.cre~ REMOVE_KNOWN_SPELL ~sppr101~ ~sppr102~Special thanks to Japh for coding this feature. |
or | REMOVE_MEMORIZED_SPELL splName list | When applied to a CRE file, this patch causes all of the
listed spells to be removed from their memory. Example:
COPY_EXISTING ~aerie.cre~ ~override/aerie.cre~ REMOVE_MEMORIZED_SPELL ~sppr101~ ~sppr102~ |
or | SET_BG2_PROFICIENCY String value | Removes all effects from the current CRE file that alter the proficiency specified by String (relative to Stats.ids). After that, if value is different from zero, add an effect to set said proficiency to value. |
or | ADD_CRE_ITEM itmName #charge1 #charge2 #charge3 flags slot [ EQUIP ] [ TWOHANDED ] | See the ADD_CRE_ITEM tutorial. |
or | REPLACE_CRE_ITEM itmName #charge1 #charge2 #charge3 flags slot [ EQUIP ] [ TWOHANDED ] | If there’s an item in the slot position, replace it; otherwise, process ADD_CRE_ITEM |
or | REMOVE_CRE_ITEM itmName list | Remove the listed items from the current creature. Does nothing if an item is not possessed. Multiple instances are removed. |
or | REMOVE_CRE_ITEMS | Remove all items from the current creature. |
or | REMOVE_CRE_EFFECTS | Remove all effects from the current creature. |
or | REMOVE_KNOWN_SPELLS | Remove all known spells from the current creature. |
or | REMOVE_MEMORIZED_SPELLS | Remove all memorized spells from the current creature. |
or | PATCH_SILENT | makes WeiDU skip all messages; it’s reenabled once a component is installed (or fails to), or a PRINT, PATCH_PRINT, VERBOSE or PATCH_VERBOSE is found. |
or | PATCH_VERBOSE | undoes the latest SILENT or PATCH_SILENT command. |
or | INNER_PATCH buffString BEGIN patch list END | Any WeiDU variables inside %s within buffString are replaced by
their values. All of the patches given are evaluated as if the
contents of the current file were buffString. Any modifications to
buffString are thrown away (making this mostly useful for reads).
Example:
INNER_PATCH "ABC" BEGIN READ_BYTE 0x2 "myvar" END PATCH_PRINT "myvar is %myvar%"This sequence will always print myvar is 67 (since 67 is the ASCII code for C). |
or | INNER_PATCH_SAVE savevar buffString BEGIN patch list END | As with INNER_PATCH, except that any modifications to buffString are stored inside
savevar.
Example:
SPRINT foo "ABC" INNER_PATCH_SAVE bar "%foo%" BEGIN WRITE_BYTE 0x2 65 END PATCH_PRINT "foo is %foo%, bar is %bar%"This sequence will always print foo is ABC, bar is ABA (since 65 is the ASCII code for A). |
or | INNER_PATCH_FILE resource BEGIN patch list END | Any WeiDU variables inside %s within resource are replaced by
their values. If the resulting resource is present in the game or in the
override folder, the patches given are evaluated as if the
current file were that resource. If not, nothing happens. Any
modifications to that resource are thrown away (making this mostly useful
for reads).
Example:
INNER_PATCH_FILE "SW1H01.ITM" BEGIN READ_BYTE 0x1 "myvar" END PATCH_PRINT "myvar is %myvar%"This sequence will always print myvar is 84 (since 84 is the ASCII code for T and SW1H01.ITM starts with ITM). |
or | INNER_ACTION BEGIN TP2 Action list END | See the INNER_ACTION tutorial, but loosely the current COPY is paused, the given TP2 Actions are executed, and then the current COPY is resumed. Note that an INNER_ACTION should never modify a file that is being modified by the current action. For example, never put APPEND foo.2da inside of COPY_EXISTING foo.2da . More formally, if the inner action and the outer action both modify the same file, the results are undefined. |
or | DECOMPRESS_REPLACE_FILE start length uncompressedlength | Decompress (ZLIB) the start...start+length-1 portion of the current
file (requires to specify the uncompressed length as well). The result overwrites
the current file. It’s a synonym for
DECOMPRESS_INTO_FILE start length uncompressedlength 0 BUFFER_LENGTH |
or | DECOMPRESS_INTO_FILE start length uncompressedlength overwritefrom overwriteto | Decompress (ZLIB) the start...start+length-1 portion of the current
file (requires to specify the uncompressed length as well). The result overwrites
the overwritefrom...overwriteto-1 portion of the current file. |
or | DECOMPRESS_INTO_VAR start length uncompressedlength varName | Decompress (ZLIB) the start...start+length-1 portion of the current
file (requires to specify the uncompressed length as well). The result is stored
into varName. |
or | COMPRESS_REPLACE_FILE start length level | Compress (ZLIB) the start...start+length-1 portion of the current
file (level is 0 to 9; 0 means don’t compress, 1 is fastest compression, 9 is best
compression). The result overwrites
the current file. It’s a synonym for
DECOMPRESS_INTO_FILE start length uncompressedlength 0 BUFFER_LENGTH |
or | COMPRESS_INTO_FILE start length level overwritefrom overwriteto | Compress (ZLIB) the start...start+length-1 portion of the current
file (level is 0 to 9; 0 means don’t compress, 1 is fastest compression, 9 is best
compression). The result overwrites
the overwritefrom...overwriteto-1 portion of the current file. |
or | COMPRESS_INTO_VAR start length level varName | Compress (ZLIB) the start...start+length-1 portion of the current
file (level is 0 to 9; 0 means don’t compress, 1 is fastest compression, 9 is best
compression). The result is stored into varName. |
when | A when clause gives you local control over when a COPY, COPY_EXISTING or APPEND_COL happens. If the COPY or COPY_EXISTING contains multiple files, each one is checked against the when clauses separately. | |
is | IF_SIZE_IS fileSize | True if the input file size is fileSize. |
or | IF regexp | True if the input file contains regexp. |
or | UNLESS regexp | False if the input file contains regexp. |
or | BUT_ONLY_IF_IT_CHANGES | True only if the file is actually changed by patching actions. Unlike all other when clauses, this one is evaluated just before the result would be written out to the disk. |
offset | A offset is a normal value. | |
value | An expression that evaluates to an integer. See --debug-value. | |
is | integer | An absolute location or amount. You may format your numbers in decimal, hex, octal or binary. Use 0x for hex, 0o for octal and 0b for binary. |
or | ( value ) | |
or | value + value | Addition. |
or | value - value | Subtraction. |
or | value * value | Multiplication. |
or | value / value | Division. Division by zero yields the value zero. Briefly, fractions are dropped and the result is an integer (so 11 / 6 = 1). More technically, “this division rounds the real quotient of its arguments towards zero” -- see http://caml.inria.fr/pub/docs/manual-ocaml/libref/Pervasives.html for more information. |
or | BYTE_AT offset | The 8-bit value from the file at the given offset. |
or | SBYTE_AT offset | The 8-bit signed value from the file at the given offset. |
or | SHORT_AT offset | The 16-bit value from the file at the given offset. |
or | SSHORT_AT offset | The 16-bit signed value from the file at the given offset. |
or | LONG_AT offset | The 32-bit value from the file at the given offset. |
or | SLONG_AT offset | The 32-bit signed value from the file at the given offset. |
or | value ** value | Exponentiation. The first value is raised to the power of the second. This is done by converting both to floating point numbers, calculating the exponent, and converting the answer back to a 32-bit integer. If anything goes wrong the value zero is returned or the result is undefined. |
or | value ** ( value value ) | Fractional exponentation. a ** (b c) yields a raised to the
power of (b/c). All of the internal operations are done in floating
point. If anything goes wrong, zero is returned or the result is
undefined. Example:
SET x = 100 ** 2 SET y = 100 ** (1 2) // square root of 100 = 10 SET z = 100 ** (1 3) // cube root of 100 = 4.64 PATCH_PRINT "w = %w%, x = %x%, y = %y%, z = %z%"yields x = 10000, y = 10, z = 4. |
or | value = value | Integer Equality. If the two values evaluate to equal integers, the result is 1. Otherwise, the result is 0. See STRING_COMPARE for comparing Strings. Synonym: ==. |
or | NOT value | Negation. If the value is 0, the result is 1. Otherwise the result is 0. |
or | ABS Value | Absolute value. If the value is >= 0, the result is x. Otherwise the result is -x. |
or | value != value | Integer Disequality. If the two values evaluate to equal integers, the result is 0. Otherwise the result is 1. See STRING_COMPARE for comparing Strings. |
or | value OR value | Disjunction. If either value is non-zero, the result is 1. Otherwise, the result is 0. Synonym: ||. |
or | value AND value | Conjunction. If both values are non-zero, the result is 1. Otherwise, the result is 0. Synonym: &&. |
or | value BAND value | Bitwise And. 0b101 BAND 0b110 = 0b100. Synonym: &. |
or | value BOR value | Bitwise Or. 0b101 BOR 0b110 = 0b111. Synonym: |. |
or | value BXOR value | Bitwise Exclusive Or. 0b101 BXOR 0b110 = 0b011. Synonym: ^^. |
or | BNOT value | Bitwise Not.
BNOT 0b111 = 0b1111111111111111111111111111000. Synonym: ` . |
or | value BLSL value | Bitwise Logical Shift Left. 0b101 BLSL 2 = 0b10100. Synonym: <<. |
or | value BLSR value | Bitwise Logical Shift Right. 0b101 BLSR 2 = 0b1. Synonym: >>. |
or | value BASR value | Bitwise Arithmetic Shift Right. This is an arithmetic shift: the sign bit of the first value is replicated and inserted in the vacated bits. 0b101 BASR 2 = 0b1. |
or | value > value | If the first value is greater than the second, the result is 1. Otherwise, the result is 0. |
or | value >= value | If the first value is greater than or equal to the second, the result is 1. Otherwise, the result is 0. |
or | value < value | If the first value is less than the second, the result is 1. Otherwise, the result is 0. |
or | value <= value | If the first value is less than or equal to the second, the result is 1. Otherwise, the result is 0. |
or | value ? value : value | An expression-valued conditional. If the first value is not 0 then the second value is evaluated and returned, otherwise the third value is evaluated and returned. |
or | String STRING_COMPARE String | This is legacy syntax for STRING_EQUAL: This expressione valuates to 0 if and only if its two string arguments are equal (have the same length and the same contents). Otherwise it will evaluate to a negative or positive integer, depending on whether the first string argument would sort lexicographically before or after the second string argument. variables within the strings (e.g., “%mykit%”) are replaced by their values. Note that variables that you want expanded must be put in %’s, otherwise the raw text will be used. You may use STR_CMP as a synonym for STRING_COMPARE. This function works just like C’s strcmp. Note also that STRING_EQUAL and STRING_COMPARE are similar, but STRING_EQUAL has more intuitive return values. |
or | String STRING_COMPARE_CASE String | This is legacy syntax for STRING_EQUAL_CASE: As STRING_COMPARE, but the comparison ignores case. That is, "ANOMEN" and "aNoMeN" are considered equal. |
or | String STRING_EQUAL String | This expressione valuates to 1 if and only if its two string arguments are equal (have the same length and the same contents), otherwise it values to 0. variables within the strings (e.g., “%mykit%”) are replaced by their values. Note that variables that you want expanded must be put in %’s, otherwise the raw text will be used. Note also that STRING_EQUAL and STRING_COMPARE are similar, but STRING_EQUAL has more intuitive return values. |
or | String STRING_EQUAL_CASE String | As STRING_EQUAL, but the comparison ignores case. That is, "ANOMEN" and "aNoMeN" are considered equal. |
or | String STRING_MATCHES_REGEXP String | As STRING_COMPARE_CASE, but the second string is treated as a regexp. Thus "AR1005" STRING_MATCHES_REGEXP "AR[0-9]+" evaluates to 0 (“no difference”). You may use STRING_COMPARE_REGEXP as a synonym. |
or | String STRING_CONTAINS_REGEXP String | As STRING_MATCHES_REGEXP, but it evaluates to 0 if the first string contains the second regexp. Thus "AR1005" STRING_CONTAINS_REGEXP "[w-z]" evaluates to 1 (“mismatch”). |
or | GAME_IS String | Returns true if the IE game variant
is one of the entries in String, otherwise it returns false. String is a list of
whitespace-separated entries, chosen from (case doesn’t matter)
bg2=soa, tob, iwd2, pst, bg1, totsc, iwd=iwd1, how, tolm=totlm, tutu, tutu_totsc, bgt, ca .
In the list, bg2=soa means that bg2 and soa are synonyms,
and you can use whichever one you like. Please note that the items in the list are
defined to be mutually exclusive (except tob will detect both plain ToB
and ToB with BGT installed, for legacy purposes), so bg2 will detect a
SoA-only game, without either ToB or Tutu installed. As such, a mod which is
useable with any flavor of BG2 and any flavor of Tutu would contain
REQUIRE_PREDICATE GAME_IS ~bg2 tob tutu tutu_totsc~ . |
or | ENGINE_IS String | Functions like GAME_IS
except bg2=soa will detect SoA-based Tutu games in addition to SoA games,
and tob will detect ToB-based Tutu games in addition to ToB and BGT games.
As such, a mod that requires Throne of Bhaal’s engine enhancements could use
REQUIRE_PREDICATE ENGINE_IS ~tob~ . |
or | IDS_OF_SYMBOL (File String) | Will return the number associated with String in File.ids, or -1 if String is not associated in File.ids. |
or | VARIABLE_IS_SET String | Returns true if the variable String is set (there’s a variable called either
String or %String% , regardless of whether it is a string or an
integer). |
or | IS_AN_INT String | Returns true if the variable String is set to an integer (there’s a variable called either
String or %String% with integer value). |
or | MOD_IS_INSTALLED modTp2Name modComponent | Returns true if the modComponent of modTp2Name is installed. The syntax is the same as with REQUIRE_COMPONENT. |
or | STATE_WHICH_SAYS text FROM String | String must be an in-game or in-override .dlg file. Returns:
|
or | STATE_WHICH_SAYS value IN String1 FROM String2 | As above, except that @x references are taken from
the String1 tra file. In particular, one %s in String1 is expanded to the list
of directories; for example (as of Sola v102), solarom/%s/epilogue.tra would match
solarom/american/epilogue.tra solarom/french/epilogue.tra solarom/german/epilogue.tra solarom/italian/epilogue.tra solarom/polski/epilogue.tra solarom/portuguese/epilogue.tra solarom/russian/epilogue.tra |
or | RANDOM ( value value ) | A random-number generator. The first value is the lower bound, the second value is the upper bound. A random integer between the lower bound and the upper bound (inclusive) is returned. Thus RANDOM(3 5) can return 3, 4 or 5. If the lower bound is greater than the upper bound, zero is returned. See also RANDOM_SEED. |
or | BUFFER_LENGTH | Returns the length of the string currently being patched, or 0 if outside of a patch expression. |
or | STRING_LENGTH String | Returns the length of the argument String (after variable evaluation). |
or | FILE_CONTAINS fileName regexp | Evaluates to 1 if the file fileName contains the regular expression regexp and 0 otherwise. Case is ignored. |
or | FILE_CONTAINS_EVALUATED ( fileName varsRegexp ) | First, all WeiDU variables enclosed in %s in varsRegexp and
fileName are
substituted. The expression is 1 if the resulting regexp occurs
in fileName and 0 otherwise. If fileName does not exist or has
size 0, the result is 0. The comparison ignores case. See also
FILE_CONTAINS. Example:
COPY_EXISTING_REGEXP ~.*\.CRE~ ~override~ READ_BYTE 0x273 class READ_ASCII 0x280 deathvar PATCH_IF (class = 14) AND // class 14 = CLERIC_MAGE (NOT FILE_CONTAINS_EVALUATED(~pdialog.2da~ ~%deathvar%~)) THEN BEGIN ADD_CRE_ITEM ~POTN08~ #10 #10 #10 ~IDENTIFIED~ ~INV15~ END BUT_ONLY_IF_IT_CHANGESThe example gives ten healing potions to all cleric-mages who cannot join the party. Notably it excludes Aerie since she has the death variable Aerie and pdialog.2da contains AERIE. |
is | FILE_EXISTS fileName | Evaluates to 1 if the file exists in the filesystem and has non-zero size (or the bif file is referenced inside the chitin.key) and evaluates to 0 otherwise. |
or | FILE_EXISTS_IN_GAME fileName | Evaluates to 1 if the file exists as a game resource and has non-zero size. Evaluates to 0 otherwise. BIFFs and the override directory are searched in the standard manner. Evaluates to 0 for a file that does not exist but has been ALLOW_MISSING’d. |
or | FILE_MD5 fileName md5sum | Evaluates to 1 if the file exists and has the given MD5 checksum. Evaluates to 0 otherwise. Two different files are exceptionally unlikely to have the same MD5 checksum. In any event, the discovered checksum is printed to the log. If the file does not exist, it evaluates to 0. |
or | FILE_SIZE fileName fileSize | Evaluates to 1 if the size of the file fileName is exactly fileSize. Evaluates to 0 otherwise. |
or | %variable% | The value of the variable is used. |
or | EVALUATE_BUFFER variable | User variables inside the given string are evaluated one additional
time. You may prepend EVALUATE_BUFFER when a value is
called for and you would normally use a string. You may also use it for
SET and SPRINT statements. Example:
SPRINT x ~y~ SET y = 5 SPRINT z EVALUATE_BUFFER ~tricky %%x%%~ SET EVALUATE_BUFFER "%x%" += 77 PATCH_PRINT "y is %y% ; z is %z%"This prints out y is 82 ; z is tricky 5. You may also do hackery like FILE_SIZE "myfile" "%%indirection%%". Be very careful with this feature. |
or | $array(index list) | The so-called array construct. See the tutorial. |
or | %WEIDU_ARCH% | The special variable WEIDU_ARCH is set to either "x86" or "mac" at WeiDU startup and can be used to determine the underlying system architecture. |
or | %WEIDU_OS% | The special variable WEIDU_OS is set to either "win32" or "osx" or "unix" at WeiDU startup and can be used to determine the underlying operating system. |
or | %COMPONENT_NUMBER% | The special variable COMPONENT_NUMBER is set to the number of the component being installed. |
or | %INTERACTIVE% | The special variable INTERACTIVE is set to 1 if the install is being done interactively (IE you launched setup-mymod.exe and you’re installing mymod), or is set to 0 if the install is being done non-interactively (IE you launched setup-mymod.exe and now weidu is reinstalling othermod due to the domino uninstall and reinstall). |
or | variable | The value of the variable is used.
In a patch value, you may use either %myvar% or
myvar to get the value of a variable, provided that your variable’s
name is not the same as a WeiDU syntactic keyword or known
constant. Also note that while it is common to “call out”
variables by putting quotation marks around them, this is not necessary
if the variable is unambiguous. Thus the following three patches
are all equivalent:
SET "x" = "%y%" + "%z%" // these all SET "x" = "y" + "z" // do the SET x = y + z // same thing |
or | NAME1 | The offset within an infinity engine resource where the unidentified general name (e.g., "Battle Axe") is stored. |
or | NAME2 | The offset within an infinity engine resource where the identified general name (e.g., "K’logarath +4") is stored. |
or | UNIDENTIFIED_DESC | The offset within an infinity engine resource where the unidentified description (e.g., "The hand axe or throwing axe is also known as a hatchet ...") is stored. |
or | IDENTIFIED_DESC | As above ... ("Clans have gone to war to possess K’log...") |
or | BIO | As above ... NPC Biography |
or | ... | Almost everything in SNDSLOT.IDS or SOUNDOFF.IDS works as well. |
variable | A variable is a textual name that holds a value. Variables are usually set with READ_BYTE, SET or SPRINT. | |
is | string | You may name the variable whatever you like, but stay
away from the special characters used in regexps. When you want to
obtain the value of a variable, enclose it in %s. Example: If a file contains the two binary integers 33 and 77, then after executing: READ_LONG 0 ~myvar~ WRITE_LONG 4 ( ~%myvar%~ + 11 ) The file will contain the two binary integers 33 and 44. |
Note: it’s advised to read these tutorials in the order they were written, as often a tutorial depends on contents explained in the tutorials before it. If the tutorial-writer is kind enough, they’ll state dependancies while introducing the subject.
This tutorial was thoughtfully provided by Japheth.
The purpose of COPY_EXISTING and COPY_EXISTING_REGEXP is to patch a file for patching. It will grab the file out of the BIFFs, or if an override version exists, it will grab it out of the override folder.
COPY_EXISTING_REGEXP, EXTEND_BOTTOM_REGEXP and EXTEND_TOP_REGEXP can be potentially powerful actions if you need to make some changes to a certain set of files all in one shot.
Consider this example for COPY_EXISTING_REGEXP:
In my mod I want to make it so that all 1-handed swords only do D6 damage, rather than the varying damages they do in SOA. To do this with the "normal" COPY_EXISTING, I would have to write out all the one handed swords like this:
COPY_EXISTING ~sw1h01.itm~ ~override/sw1h01.itm~ ~sw1h02.itm~ ~override/sw1h02.itm~ etc...
That could take a lot of time to do. Using COPY_EXISTING_REGEXP I can minimize writing 70+ lines of code into 3 lines of code. Take a look at this:
COPY_EXISTING_REGEXP ~sw1h..[^abc].*itm~ ~override~ WRITE_LONG 0x88 "6" WRITE_LONG 0x8a "1"
I’ll explain how the regexp wildcards in the above example work: I want to avoid copying over the files sw1h54a.itm, sw1h54b.itm and sw1h54c.itm because they are the three components that make up the Equalizer, and because of that, the offsets of 0x88 and 0x8a won’t work because they don’t exist in those items.
To find which offset to WRITE_LONG at I opened up Near Infinity, (found at http://www.idi.ntnu.no/~joh/ni/) and looked up the offset with it. Fortunately for us, the offset is the same in all items, so we can make the changes in one shot. WRITE_LONG 0x88 "6" is telling WeiDU to write the value of 6 at offset 88 hex. Similarly, WRITE_LONG 0x8a "1" is telling WeiDU to write the value 1 at offset 8a hex.
The 6 is the dice size and the 1 is the number of dice.
This tutorial was thoughtfully provided by Japheth.
Let’s take this situation that I was in:
I am making a Hirelings mod, and in order to get non party NPC’s to transition from area to area, I needed to make use of a combination of MakeGlobal(), InMyArea(O:Object*) and MoveGlobalObject(O:Object*,O:Target*).
I originally put the script that moves them from area to area into baldur.bcs (this script is constantly running in the game), but for some strange reason, the cre’s wouldn’t move to smaller areas, only the large "main" areas. (Like The Docks, Temple District, Slums, etc.) So I figured, "Huh, guess I’m going to have to put the script in every area script". (Area scripts are scripts that are assigned to areas that are ran while you are in the area. Pretty obvious right?)
Now, writing the code to EXTEND_TOP this little script into every area script would of taken quite a while, and well, rather than do that, I humbly requested Wes to implement the regexp feature of EXTEND_TOP and EXTEND_BOTTOM, and he obliged. (My penance for that request is writing the docs you are reading right now.)
So now with Wes’ help, I could EXTEND_TOP to every area in 1 fell swoop. If you were to open up NI and expand the BCS tree, you would see a whole crapload of scripts that begin with the prefix AR. These are the area scripts I mentioned before. They go from AR0014 to AR6400. So now, using regexp, here is how you would extend the script to the top of every area script:
EXTEND_TOP_REGEXP ~ar[0-6].*bcs~ ~pathtoscript/patch.bcs~
So there we are, instead of 100+ lines of code, I minimized it to 1 lines. Now, if you are understanding regexp at all, you probably are going “Hey wait, why didn’t you just go EXTEND_TOP_REGEXP "ar.*bcs" "pathtoscript/patch.bcs" ?” Well, infact, I did do that at first, but I forgot to account that there are other “normal” scripts that begin with AR. So, I had to write it so that the regexp had the number after the initial AR so that WeiDU would know only to patch the script to area files.
This tutorial was thoughtfully provided by Japheth, and shamelessly edited by Bigg to use PATCH_IF rather than IF_EVAL.
READ_BYTE, READ_SHORT and READ_LONG can be potentially powerful functions when used in conjunction with PATCH_IF.
Here’s a brief demonstration.
The scenario: I want to make all longswords in the game do 1D6 damage, rather than their normal 1D8. To do that manually would be a pain, so why not use the great feature of READ_BYTE/LONG/SHORT?
Here’s how it’s done.
First of all, we need to copy all the items so they are ready to be patched. Here’s how we do that:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~
All this does is copy all files with a .itm extension to the override folder.
Next, we read the item category offset into a variable. We’ll use the variable name "type".
READ_BYTE "0x1c" "type"
I found the offset by looking at an item file in Near Infinity. The variable "type" can really be anything, it’s just what I’ve chosen.
We should edit the item only if the item is really a longsword:
PATCH_IF ("%type%" = "0x14") BEGIN
So, to call back our variable, you just wrap it with % signs. 0x14 is just the value for longsword. If you look at the Category offset in Near Infinity (which is 0x1c), and scroll down in the box below, you’ll see Longsword (20). The 20 just represents the decimal value. 0x14 is just 20 in hexadecimal.
Now that we’re done reading that into a variable and doing the checks, we can now WRITE_SHORT the values we want to give our new longswords. Namely, 6 for damage and 1 for dice size. (Or, 1D6 for simplicity sake.)
WRITE_SHORT 0x88 6 WRITE_SHORT 0x8a 1
0x88 is the offset for dicesize and 0x8a is the offset for the number of dice. Again, I found these offsets using Near Infinity beforehand.
END BUT_ONLY_IF_IT_CHANGES
Finally, we close the BEGIN that was after PATCH_IF and tell WeiDU not to copy the item if it didn’t change (IE, if it wasn’t a long sword).
So, to loop through the logic again, this is what we’re saying:
If the category is longsword, make the dice size 6 and the dice damage 1. If there are no changes, ignore the file and don’t copy it over.
And that’s it. Now all items with the category Longsword will have 1D6 as their dice damage. Pretty neat eh? Obviously you can do many, many other types of thing with these functions. So go ahead and experiment.
Here’s the full code:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ READ_BYTE "0x1c" "type" PATCH_IF ("%type%" = "0x14") BEGIN WRITE_SHORT 0x88 6 WRITE_SHORT 0x8a 1 END BUT_ONLY_IF_IT_CHANGES
Oh yeah, and if you’re going "But I don’t know when to READ_SHORT or WRITE_LONG. It’s so confusing," don’t worry, cause it kind of is if you’ve never worked with bytes before. Here’s a very brief synopsis.
Use READ/WRITE_SHORT for anything that is two bytes. You can determine this easily by looking at the field you want to change in NI and subtracting the next listed field’s offset from the offset of the field you want to change.
For example, if I wanted to change Bodhi’s Max HP here’s what I would do. First look at her creature file in NI. (She has many, but for the sake of argument, we’ll use bodhi.cre). If you look at the Max HP field, it says it’s offset is 26h. The next field in the creature file is Animation ID, which is 28h. So, subtracting 28h from 26h gives you 2h. The perfect WRITE_SHORT size.
Use READ/WRITE_LONG for anything that is 4 bytes. Use the above example to figure this out as well.
Use READ/WRITE_BYTE for anything that is one byte. Again, use the above example to figure this out.
Use WRITE_ASCII for things like death variable, dialogues, scripts, etc. Note, that if you’re WRITE_ASCIIing to say, the death variable, and the previous death variable entry is longer than the one you’re writing, then you’ll have spillover after WRITE_ASCII. To avoid this, embed a null character at the end of your string. So, if I have
WRITE_ASCII 0x280 ~Guy~
I would want to put a null character right after the y in Guy. You can do this by using a hex editor and inserting 00 after the y.
These are by no means "hard and fast" rules, but generally they ring true. However, there could be times when you need to WRITE_LONG twice if the field is 8 bytes long, etc.
Also, if you tried to run the template, you’d have noticed that it didn’t work
correctly. That’s because there are some corrupt files in the key (that is,
files of size 0). To avoid that, WeiDU gives us the %SOURCE_SIZE%
variable,
which is set to the original file size. We can simply check that the file is
large enough (using either NI or the IESDP, we see that a basic itm file must
have at least 0x72 bytes long, so we can check that before doing anything:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ PATCH_IF ("%SOURCE_SIZE%" > "0x71") BEGIN READ_BYTE "0x1c" "type" PATCH_IF ("%type%" = "0x14") BEGIN WRITE_SHORT 0x88 6 WRITE_SHORT 0x8a 1 END END BUT_ONLY_IF_IT_CHANGES
Again, we instruct WeiDU to not do a thing if the file is not large enough.
And that, as they say, is that.
This tutorial was thoughtfully provided by Japheth, and shamelessly edited by Bigg to use PATCH_IF rather than IF_EVAL.
Using expressions with WeiDU can get quite complex, however the power they give you is more than worth the effort in learning how to use them.
Say, for instance, that you want to give a creature file a shield that doesn’t have one originally. The old way to do this would be to add the item to the creature file using Near Infinity or ShadowKeeper and then copy that creature file over upon installation of your mod. However, now that INSERT/DELETE_BYTES and READ/WRITE_BYTE/LONG/SHORT take expressions, we can do this "on the fly" when installing your mod.
The benefits of doing it this way should be obvious. Now you no longer have to overwrite someone else’s modifications to that creature file. You can simply add your shield and leave their modifications intact. (Unless of course *they* also decided to give the exact same creature a shield. And if that’s the case, the shield you specify won’t be used because of our PATCH_IF statement.)
Anyways, on to the example. I’ll use acolyte1.cre as my example creature.
The first thing we want to do is copy the creature over so it’s ready for patching, and ensure it isn’t corrupt. That’s easy to do using a simple COPY_EXISTING statement:
COPY_EXISTING ~acolyte1.cre~ ~override/acolyte1.cre~ PATCH_IF ("%SOURCE_SIZE%" > "0x2d3") BEGIN
Next, we need to read in the offsets so we know where to insert our bytes and where to update our number of items. I figured out these offsets by simply viewing the creature file in Near Infinity.
The offsets we need to read in are the item slots offset, the items offset and the number of items. This is how we do it:
READ_LONG "0x2bc" "itemsoffset" READ_LONG "0x2b8" "itemslot" READ_LONG "0x2c0" "#items"
"itemsoffset", "itemslot" and "#items" can be anything you choose, but I would name them something that’s easy to remember because we’ll be using them later.
Next, we’re going to read in the shield slot to make sure that it’s empty. We do this by taking the "itemslot" variable and adding 0x04 to it which we know is always going to be the shield slot. (You can verify this by looking at a creature file in NI. If you look at the offset 0x2b8 it will have the offset for item slots. If you add 0x04 to that number, it should equal the offset for the shield slot.)
This is done like so:
READ_SHORT ("%itemslot%" + 0x04) "shield"
Note: To access your variables after they’ve been read, you always have to wrap them in percent signs.
We want to tell WeiDU only to do this if the shield slot is *empty*. This is where PATCH_IF comes in handy. The way a creature file is setup is that if a slot is set to -1, it means it’s empty. WeiDU
doesn’t read unsigned bytes, so if the value is indeed -1, it will come out as 65535 as the value. So, this is the statement we use to tell WeiDU only to execute all of the above code if "%shield%" is -1:
PATCH_IF ("%shield%" = "65535") BEGIN
Now, we have to update the shield slot to reflect that it’s the newest item out of all the items the creature has equipped. This is a bit misleading because in the slots part of the creature file, they count from 0. So, all we need to do is take the value of 0x2c0 (The number of items the creature has.) and use that as our value to write.
To do that, this is what we do:
WRITE_SHORT ("%itemslot%" + 0x04) "%#items%"
Now we have to update the itemslot offset value to reflect the fact that we’ve added a new item to the creature. An item is *always* 0x14 bytes. We already know the offset of the itemslot offset and we’ve already read it’s value, so all we need to do is add 0x14 to it
WRITE_LONG 0x2b8 ("%itemslot%" + 0x14)
The last step before inserting our bytes is to increase the #items by 1 to reflect the fact that we’ve added an item to the creature. We’ve already read in the number of items into "%#items%" so this is dead easy:
WRITE_LONG 0x2c0 ("%#items%" + 1)
Now we can actually insert out bytes. To do this we have to take the number of items that the creature has multiplied by 0x14 and add that to the itemsoffset. (The reason we can’t just insert it right at "%itemsoffset%" is because it would mess up the order of the items in game. If we inserted the shield at the beginning, then the creature would more than likely have some items equipped in some weird places in the game.)
This is done like so:
INSERT_BYTES ("%itemsoffset%" + "%#items%" * 0x14) 0x14
Now we can actually write our information for our item. You don’t need the .itm extension at all, just everything before the .itm. And again, since the item entry is always the first entry in an item field on a creature file, we just need to use the same formula as above because we’ll be writing at the same offset that we inserted our bytes at:
WRITE_ASCII ("%itemsoffset%" + "%#items%" * 0x14) ~shld01~
If you wanted to add a magical shield to the creature file and wanted it to be already identified in the slot, then you would simply have to add 0x10 bytes to the previous expression and WRITE_LONG the value of 1. Like this:
WRITE_LONG ("%itemsoffset%" + ("%#items%" * 0x14) + 0x10) 1
Finally, we close the PATCH_IF statements and add BUT_ONLY_IF_CHANGES.
END END BUT_ONLY_IF_IT_CHANGES
Here’s the full code:
COPY_EXISTING ~acolyte1.cre~ ~override/acolyte1.cre~ PATCH_IF ("%SOURCE_SIZE%" > "0x2d3") BEGIN READ_LONG "0x2bc" "itemsoffset" READ_LONG "0x2b8" "itemslot" READ_LONG "0x2c0" "#items" READ_SHORT ("%itemslot%" + 0x04) "shield" PATCH_IF ("%shield%" = "65535") BEGIN WRITE_SHORT ("%itemslot%" + 0x04) "%#items%" WRITE_LONG 0x2b8 ("%itemslot%" + 0x14) WRITE_LONG 0x2c0 ("%#items%" + 1) INSERT_BYTES ("%itemsoffset%" + "%#items%" * 0x14) 0x14 WRITE_ASCII ("%itemsoffset%" + "%#items%" * 0x14) ~shld01~ END END BUT_ONLY_IF_IT_CHANGES
Note that you can use the PRINT action in a tp2 to debug your code as well. Like this:
COPY_EXISTING ~acolyte1.cre~ ~override/acolyte1.cre~ PATCH_IF ("%SOURCE_SIZE%" > "0x2d3") BEGIN READ_LONG "0x2bc" "itemsoffset" READ_LONG "0x2b8" "itemslot" READ_LONG "0x2c0" "#items" READ_SHORT ("%itemslot%" + 0x04) "shield" PATCH_IF ("%shield%" = "65535") BEGIN WRITE_SHORT ("%itemslot%" + 0x04) "%#items%" WRITE_LONG 0x2b8 ("%itemslot%" + 0x14) WRITE_LONG 0x2c0 ("%#items%" + 1) INSERT_BYTES ("%itemsoffset%" + "%#items%" * 0x14) 0x14 WRITE_ASCII ("%itemsoffset%" + "%#items%" * 0x14) ~shld01~ END END BUT_ONLY_IF_IT_CHANGES PRINT ~The value of the items offset is %itemsoffset%~
I should also mention that the above code for adding an item to a creature file is "portable" in a sense. You would only have to change a couple things. The changes that would have to be made are as follows:
And that’s the end of it. All this can be a bit tricky, but as I mentioned above, it can be quite powerful as well.
This tutorial and this feature were thoughtfully provided by Japheth.
Using ADD_STORE_ITEM is a relatively painless procedure. Consider this example: I want to add a new item to ribald.sto in BGII. There were two ways of doing this before.
Now we can do #2 rather easily thanks to ADD_STORE_ITEM.
First off copy over the store file to set it up for patching as you normally would.
COPY_EXISTING ~ribald.sto~ ~override/ribald.sto~
Next we issue the ADD_STORE_ITEM patch expression with the following arguments:
ADD_STORE_ITEM "myitem" #10 #0 #0 ~IDENTIFIED~ #5
“myitem” is the name of the item file that we will want to appear in the store without it’s .itm extension. #10 is the first extension headers number of charges The two #0s that follow are the second and third extension headers number of charges. IDENTIFED is the flag that we want on the item. Here are the flags that you can use:
IDENTIFIED UNSTEALABLE STOLEN IDENTIFIED&STOLEN IDENTIFIED&UNSTEALABLE
#5 is the number of items that will be in stock.
Pretty easy stuff right?
You can also add an optional + after ADD_STORE_ITEM if you want to overwrite an item that already exists in the store.
So, if we wanted to replace HAMM05.ITM in ribald.sto this is what we’d do:
ADD_STORE_ITEM + ~hamm05~ #10 #0 #0 ~IDENTIFIED~ #5
One final note: remember to copy over your new item because ADD_STORE_ITEM doesn’t do that part for you.
So, a complete set of actions that would patch a store file and copy over the new item would look like this:
COPY_EXISTING ~ribald.sto~ ~override/ribald.sto~ ADD_STORE_ITEM ~myitem~ #10 #0 #0 ~IDENTIFIED~ #5 COPY ~mymod/myitem.itm~ ~override/myitem.itm~
Note that you can also add a final optional string argument if you want the store to have unlimited copies of that item:
ADD_STORE_ITEM ~myitem~ #10 #0 #0 ~IDENTIFIED~ #1 ~UNLIMITED~
That’s about all there is to it.
Note that as of v188 you can specify the position of the item being added:
COPY_EXISTING ~ribald.sto~ ~override/ribald.sto~ ADD_STORE_ITEM ~myitem~ AFTER ~olditem~ #10 #0 #0 ~IDENTIFIED~ #5 ADD_STORE_ITEM ~otheritm~ BEFORE ~olditem~ #10 #0 #0 ~IDENTIFIED~ #5
Valid options for position are:
FIRST will add the item at the top of the stack. This is the default, equivalent to putting nothing there.
LAST will add the item to the bottom of the stack.
BEFORE olditem will add the item exactly before olditem, or to the top of the stack if olditem is missing.
AFTER olditem will add the item exactly after olditem, or to the bottom of the stack if olditem is missing.
As of 192, if there’s a list of items in BEFORE|AFTER olditem (IE ~item1 item2 item3
),
we add the new item BEFORE or AFTER the first item that is present in the list.
As of 211, you can specify a (zero-based) hardcoded position via AT position
.
This tutorial and this feature were thoughtfully provided by Japheth.
In BGII, the game engine was refined so that when any NPC joined your party they would be added to the baldur.gam file, regardless of the fact if they were in there already or not.
Sadly, in BG1 this is not the case. If you were to simply CreateCreature any old NPC and add them to your party, as soon as you moved to another area the game would crash and die.
This is where ADD_GAM_NPC comes in handy for all you BG1 modders out there.
By using ADD_GAM_NPC you can patch your NPC into all the baldur.gam files on an end users install. This includes the default GAM file that’s loaded up when you start a new game as well as all the GAM files in the save and mpsave directories.
Note: However, there really is no foolproof way to back up all the save games, so you may want to warn the end user in a readme to manually back them up themselves beforehand.
Anyways, I digress.
To use ADD_GAM_NPC you have to first copy over the CRE file that you will want to appear in the GAM file. I’m pretty sure we all know how to do this, but for completeness, here it is:
COPY ~mymod/mynpc.cre~ ~override/mynpc.cre~
After the COPY statement, make sure you do all your SAYs and any other type of patching that you’ll want to do for your .cre file. You will always want to use ADD_GAM_NPC last since you want the updates to your cre file to be reflected in the GAM file as well.
Here’s a typical patch statement for most NPC mods:
COPY ~mymod/mynpc.cre~ ~override/mynpc.cre~ SAY NAME1 ~Japh~ SAY NAME2 ~Japh~
So, after doing all that, we would issue ADD_GAM_NPC with the following arguments:
ADD_GAM_NPC "mynpc" "ar2600" #123 #456
“mynpc” is the name of your NPCs cre file that you’re copying over without the .cre extension.
“ar2600” is the area that you want him/her placed in. In this case we’re placing the NPC in Candlekeep.
#123 is the x co-ordinate on the map.
#456 is the y co-ordinate on the map.
That’s all there is to it. Just to recap, here’s the full set of expressions we would use:
COPY ~mymod/mynpc.cre~ ~override/mynpc.cre~ SAY NAME1 ~Japh~ SAY NAME2 ~Japh~ // Do any other patching stuff here ADD_GAM_NPC ~mynpc~ ~ar2600~ #123 #456
Note: Again, I have to stress the importance of warning the end users to backup their save and mpsave directories beforehand. There really is no elegant way to do this via WeiDU (well, maybe there is, but I can’t think of a way) so I’d just warn them.
This tutorial was thoughtfully provided by Idobek.
There have been some queiries about SET_2DA_ENTRY in the WeiDU forum recently. This is a consolidation of my posts on the subject. The command takes this form:
SET_2DA_ENTRY value value value newEntry
The first value is the row, the second is the column and the third is the required column count. The entry on the given column of the given row is set to newEntry, but only rows with at least as many columns as the required column count are considered. The upper-left entry is 0,0. For example, given the following 2DA file:
2DA V1.0 * ROWNAME LOWER MIXED HELP 0 RESERVE * * * 1 BERSERKER 25179 25151 25201 2 WIZARD_SLAYER 25180 25152 25203 3 KENSAI 25181 25153 25204 4 CAVALIER 25182 25154 25206
Then the patch:
SET_2DA_ENTRY 3 1 5 ~SAMURAI~
would result in:
2DA V1.0 * ROWNAME LOWER MIXED HELP 0 RESERVE * * * 1 BERSERKER 25179 25151 25201 2 WIZARD_SLAYER 25180 25152 25203 3 SAMURAI 25181 25153 25204 4 CAVALIER 25182 25154 25206
So the columns and rows you want WeiDU to consider are:
Column0 Column1 Column2 Column3 Column4 ROWNAME LOWER MIXED HELP Row0: 0 RESERVE * * * Row1: 1 BERSERKER 25179 25151 25201 Row2: 2 WIZARD_SLAYER 25180 25152 25203 Row3: 3 KENSAI 25181 25153 25204 Row4: 4 CAVALIER 25182 25154 25206
The total number of columns is five. If, however, you put 1 as your required column count then you tell WeiDU you want to consider all rows with at least 1 column. So your row numbers change:
Row0: 2DA V1.0 Row1: * Row2: ROWNAME LOWER MIXED HELP Row3: 0 RESERVE * * * Row4: 1 BERSERKER 25179 25151 25201 Row5: 2 WIZARD_SLAYER 25180 25152 25203 Row6: 3 KENSAI 25181 25153 25204 Row7: 4 CAVALIER 25182 25154 25206
So the required column count is what you use to tell WeiDU which row you want to use as row0. So using the code
SET_2DA_ENTRY 3 1 1 ~SAMURAI~
would result in:
2DA V1.0 * ROWNAME LOWER MIXED HELP 0 SAMURAI * * * 1 BERSERKER 25179 25151 25201 2 WIZARD_SLAYER 25180 25152 25203 3 KENSAI 25181 25153 25204 4 CAVALIER 25182 25154 25206
Hopefully, that explains the required column count a little better.
Finally, if newEntry is a string that contains any %WeiDU_Variables% they are replaced by their values. That’s the end of this tutorial.
This tutorial was thoughtfully provided by Idobek, and shamelessly edited by Bigg to use PATCH_IF rather than IF_EVAL, and PATCH_IF rather than WHILE.
So what is a WHILE loop? Well very simply put it is a way of applying the same patch to a file multiple times under a different condition each time. Or it is a way of applying different patches depending upon the conditions. This example will show both of these methods. Keeping in theme we are going to modify the damage done by axes. Now, some axes have both a melee and a ranged ability. We are going to want to change both of these abilities but give different damage to ranged and melee.
First, as always, we need to copy the files ready for patching, and throw in the usual sanity check:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ PATCH_IF ("%SOURCE_SIZE%" > "0x71") BEGIN
We only want to modify axes so let’s find out what the item is:
READ_BYTE 0x1c "category"
Once we are done with our patching we will use PATCH_IF to tell WeiDU to only patch axes:
PATCH_IF ("%category%" = 25) BEGIN
Now, we need to find out how many abilities there are and where they are located:
READ_LONG 0x64 "abilitiesoffset" READ_SHORT 0x68 "#abilities"
So far, so good. Here’s where the fun begins. We want to patch examine every ability in the item. We need to use a WHILE loop to patch the abilities one by one. We open the loop like so:
WHILE ("%#abilities%" > 0) BEGIN
For the WHILE loop to progress and close we need to modify the "#abilities" variable. We use the SET command to do this:
SET "#abilities" = ("%#abilities%" - 1) END
These lines are placed at the end of the WHILE loop. WeiDU will process the abilities from last to first and reduce the "#abilities" variable by one each time. Once the "#abilities" variable hits zero the WHILE loop will stop.
OK, now we need to find out what type of ability we are looking at. Within an item an ability is 0x38 bytes, abilities also count from 0 within items so we must take this into account:
READ_BYTE ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38) "abilitytype"
Now we get to the meat of the patch. This patch is conditional on the ability type we have just read. First we will deal with the melee damage patch. We are going to be using a PATCH_IF command to do this
PATCHING ("%abilitytype%" = 1) BEGIN
So we are patching when the ability type is 1 (melee). Time for the actual patch:
WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x16) 6 WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x18) 2
Time to close the PATCH_IF statement.
END
We use exactly the same method for ranged abilities:
PATCH_IF ("%abilitytype%" = 2) BEGIN WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x16) 12 WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x18) 1 END
We can now close the main loop (using the aforementioned method):
SET "#abilities" = ("%#abilities%" - 1) END
We are done with our patches so we can add our usual END and BUT_ONLY.
END END BUT_ONLY_IF_IT_CHANGES
The full code is:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ PATCH_IF ("%SOURCE_SIZE%" > "0x71") BEGIN READ_BYTE 0x1c "category" PATCH_IF ("%category%" = 25) BEGIN READ_LONG 0x64 "abilitiesoffset" READ_SHORT 0x68 "#abilities" WHILE ("%#abilities%" > 0) BEGIN READ_BYTE ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38) "abilitytype" PATCH_IF ("%abilitytype%" = 1) BEGIN WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x16) 6 WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x18) 2 END PATCH_IF ("%abilitytype%" = 2) BEGIN WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x16) 12 WRITE_SHORT ("%abilitiesoffset%" + ("%#abilities%" - 1) * 0x38 + 0x18) 1 END SET "#abilities" = ("%#abilities%" - 1) END END END BUT_ONLY_IF_IT_CHANGES
So what have we done? Let’s list the steps involved:
That’s the end of this tutorial.
This tutorial was thoughtfully provided by CamDawg, and shamelessly edited by Bigg to use PATCH_IF rather than IF_EVAL.
This was a topic that I struggled with while trying to convert some kits with unusual item restrictions. Thanks to WeiDU’s new bitwise operators, this is a process that can done dynamically and non-destructively and affect all items of a particular type, even if added or altered by another mod (assuming yours is installed after others of course ). The basic idea is to construct a function in WeiDU that:
This tutorial is basically an expansion on Japheth’s excellent READ/BYTE/LONG/SHORT tutorial and I suggest you take a look at that before going further.
My initial problem came about when trying to make specific items usable or unusable by a particular class/race/kit. If you look at the item file structure at IESDP, you see that all of the unusability flags of an item are controlled by the individual bits of bytes 0x1E, 0x1F, 0x20 and 0x21 for class/race alignment restrictions and 0x29, 0x2B, 0x2D and 0x2F for the individual kit restrictions. (Near Infinity combines and displays the four bytes of class/race restrictions as a single chunk of data.)
A quick aside about notation before I proceed. Numbers in binary (the strings of bits) are preceded with 0b and hexadecimal numbers (typically the bytes) are preceded with 0x. Eight bits make a byte and bits read right-to-left. So if the second bit is 1 and the rest 0, then the bit would be written as 0b00000010.
If you were trying to alter the usability of a specific class or race, simply using a WRITE_BYTE command on any of these particular bytes would result in changing the usability of an item by all the classes in the particular byte. The new bitwise operators provide an easier solution.
First we need to look at what the new operators BAND and BOR do with bits. They are both ways of combining two bytes, based on different rules. Both BAND and BOR compare the individual bits (bit 0 vs bit 0, bit 1 vs bit 1, etc. all the way through bit 7 vs bit 7) of two bytes. For each individual bit, the following tables are used to determine the value:
0 BAND 0 = 0 0 BAND 1 = 0 1 BAND 0 = 0 1 BAND 1 = 1 0 BOR 0 = 0 0 BOR 1 = 1 1 BOR 0 = 1 1 BOR 1 = 1
So if byte 0x23 is 0b00110101 and byte 0x24 is 0b10010001, then 0x23 BAND 0x24 is 0b00010001 whereas 0x23 BOR 0x24 is 0b10110101.
Back to relating this to unusability in items. For an item to be flagged as unusable, the corresponding bit must be set to 1. A 0 means a particular item is usable by the class/race. Let’s look at the example of changing mage robes usable by bards. This would go into your TP2 file.
COPY_EXISTING_REGEXP GLOB ~.*\.itm~ ~override~ //copies all item files PATCH_IF ("%SOURCE_SIZE%" > "0x71") BEGIN // avoid crashing on empty items READ_BYTE "0x20" "mage" //reads the byte containing the mage usability flag PATCH_IF (("%mage%" BAND "0b00000100") = "0b00000000") BEGIN // if it is usable by mages READ_BYTE "0x1E" "bard" //reads the byte containing bard usability flag READ_SHORT "0x1C" "type" //reads the byte containing item type PATCH_IF ("%type%" = "67") OR ("%type%" = "2") BEGIN // if it is a robe or armor WRITE_BYTE "0x1E" ("%bard%" BAND "0b10111111") // makes usable by bards END END END BUT_ONLY_IF_IT_CHANGES
This is the same basic idea in Japheth’s BYTE tutorial , except now we’re utilizing the bitwise operators in the writing and evaluation commands. The mage usability flag is the bit 2 in byte 0x20 (IESDP). By checking "%mage%" BAND 0b00000100, the values of any of the 7 other bits (0-1,3-7) are set to 0, whereas bit 2 is equal to 0 if and only if it is 0 to begin with. Therefore the statement will only be true if the item is usable by mages (bit 2 = 0). If that is true and the item is a robe or armor, then we write ("%bard%" BAND "0b10111111") into the bard usability byte, 0x1E. By using 1 in bits 0-5 and 7, this ensures that the original value of the flag is preserved, and using 0 in bit 6 ensures that bit 6 is set to 0 regardless of its previous value--making it usable by bards whether it was before or not.
Just as an aside, you could read the entire usability block with a READ_LONG at 0x1E. However, it becomes a pain because then you need to start writing out all 32 bits when doing the bitwise operations.
Let’s try another example. Let’s try making large weapons such as two-handed swords and halberds unusable by short folks--gnomes, halflings, and dwarves.
COPY_EXISTING_REGEXP GLOB ~.*\.itm~ ~override~ //copies all item files PATCH_IF ("%SOURCE_SIZE%" > "0x71") BEGIN // avoid crashing on empty items READ_BYTE "0x21" "race" //reads the byte containing race usability flags we're interested in READ_BYTE "0x31" "prof" //reads the byte containing item type PATCH_IF ("%prof%" = "93") OR // two-handed sword ("%prof%" = "99") BEGIN // or halberd WRITE_BYTE "0x21" ("%race%" BOR "0b00010101") // makes unusable by dwarves, halflings, and gnomes END END BUT_ONLY_IF_IT_CHANGES
In this case, rather than reading the ’type’ byte of the item, I’ve opted for the ’proficiency’ byte. Many two-handed swords in BG2 are classed as long swords for some reason so in this case proficeiency is a better indicator if it is a two-handed sword IMHO.
The unusability flags for dwarves, halflings and gnomes are all in byte 0x2F at bits 0, 2, and 4 respectively (thanks again IESDP). By writing "%race%" BOR "0b00010101" to byte 0x2F we’re preserving the values of bits 1,3, and 5-7 while setting the 0, 2, and 4 bits to 1 (unusable) regardless of their previous values.
Thanks especially to Smoketest for helping me with this, Japheth for the tutorial that inspired this and the help given, and the IESDP team.
While the BCS and BAF script formats are common to all Infinity Engine games, there are three main variants. The variations deal largely with how “object parameters” are expressed. The style used by Baldur’s Gate, Baldur’s Gate 2, and Icewind Dale is considered to be the default. This document will not explain Infinity Engine scripting -- look for a tutorial elsewhere.
BG BAF Object | EA.GENERAL.RACE.CLASS.SPECIFIC.GENDER.ALIGN |
PST BAF Object | EA.FACTION.TEAM.GENERAL.RACE.CLASS.SPECIFIC.GENDER.ALIGN |
IWD2 BAF Object | EA.GENERAL.RACE.SUBRACE.CLASS.SPECIFIC.GENDER.ALIGNMNT |
If SUBRACE.IDS is present in CHITIN.KEY, the scripting style is autodetected as IWD2. Otherwise, if FACTION.IDS is present, the scripting style is autodetected as PST. Otherwise the scripting style is set to BG. The --script-style command-line argument and the SCRIPT_STYLE TP2 flag both override this default. You should not need to use --script-style explicitly if the --game is set correctly.
Also, IWD1 and IWD2 have different rules for parsing certain IDS files, which contain spaces and/or commas in the tokens.
This tutorial was thoughtfully provided by CamDawg, and edited by the Bigg to account for FORCED_SUBCOMPONENT.
SUBCOMPONENTs allow to group together a set of mutually exclusive mod components into a single menu-style selection. The primary purpose of SUBCOMPONENTs is to streamline mod installation and to make life easier for end users. For example, SUBCOMPONENTs are ideal if you wish to provide multiple portrait selections for an NPC mod, have several kits available to one NPC, or for changes that conflict with one another (i.e. raising an XP cap to 10 million or raising it to 20 million). We’ll use the first one (multiple portrait options) as an example.
Without SUBCOMPONENTs, the install dialogue would look something like this:
Install Component [Delainy Portrait 1 by Bob]? [I]nstall, [N]ot Install, or [Q]uit Install Component [Delainy Portrait 2 by Fred]? [I]nstall, [N]ot Install, or [Q]uit Install Component [Delainy Portrait 3 by Wilhelmus]? [I]nstall, [N]ot Install, or [Q]uit
The end user would need to click through the options each time. Worse, conflicting components could be installed. These problems can be limited somewhat by judicious use of predicates, but taking advantage of the SUBCOMPONENT feature yields far superior results:
Install Component [Delainy Portrait]? [N]o, [Q]uit, or choose one: 1] Portrait 1 by Bob 2] Portrait 2 by Fred 3] Portrait 3 by Wilhelmus
Only one of these options can be installed at any time; re-installing and selecting a different SUBCOMPONENT will automatically uninstall the previously installed one. Setting this up is dead simple:
BEGIN ~Portrait 1 by Bob~ /* The string above is displayed in the subcomponent listing, i.e. the list with 1] 2] 3] etc. You can, of course, use TRA references instead for this and the SUBCOMPONENT string below. */ SUBCOMPONENT ~ Delainy Portrait~ /* The string above is displayed as the component listing and must be the same for each SUBCOMPONENT. The tp2 code that follows is only executed if this SUBCOMPONENT is selected. */ COPY ~Delainy/portraits/opt1G.bmp~ ~override/CDDELAIG.bmp~ COPY ~Delainy/portraits/opt1M.bmp~ ~override/CDDELAIM.bmp~ COPY ~Delainy/portraits/opt1S.bmp~ ~override/CDDELAIS.bmp~ BEGIN ~Portrait 2 by Fred~ SUBCOMPONENT ~ Delainy Portrait~ COPY ~Delainy/portraits/opt2G.bmp~ ~override/CDDELAIG.bmp~ COPY ~Delainy/portraits/opt2M.bmp~ ~override/CDDELAIM.bmp~ COPY ~Delainy/portraits/opt2S.bmp~ ~override/CDDELAIS.bmp~ BEGIN ~Portrait 3 by Wilhelmus~ SUBCOMPONENT ~ Delainy Portrait~ COPY ~Delainy/portraits/opt3G.bmp~ ~override/CDDELAIG.bmp~ COPY ~Delainy/portraits/opt3M.bmp~ ~override/CDDELAIM.bmp~ COPY ~Delainy/portraits/opt3S.bmp~ ~override/CDDELAIS.bmp~
Any REQUIRE_FILEs or other module requirements for the whole group should be put with the first subcomponent. If such a requirement fails, none of the subcomponents can be installed. In addition, each individual subcomponent can be guarded by its own predicate. If that predicate fails, that particular subcomponent cannot be installed. Example:
BEGIN ~Imoen: Turnip-Mage~ SUBCOMPONENT ~Imoen Mage Kit~ (FILE_EXISTS_IN_GAME ~turnip.spl~) // This particular subcomponent will only be displayed if TURNIP.SPL // exists. If it does not, you can still install other subcomponents.
One note about SUBCOMPONENTs and mod ordering: WeiDU will display SUBCOMPONENTs in a single grouping no matter if they fall consecutively in the TP2. However, the component number (the one that gets placed in weidu.log and the one you check for in REQUIRE_COMPONENT et al.) is still based on their TP2 order.
As of v192, if all subcomponents fail their requirement, then we don’t ask about the specific group; moreover, if at least one of the subcomponents is defined with FORCED_SUBCOMPONENT rather than SUBCOMPONENT, then the user is forced to install one of them. This could be useful if you want to ask the user if he wants biffing of wav files or not (as you need different actions for either choice).
This tutorial was thoughtfully provided by Japheth.
This command is a little bit more complex. First I’ll list all the valid flags and inventory slots.
Flags:
Inventory Slots:
Now here is the syntax:
COPY_EXISTING ~alynar.cre~ ~override/alynar.cre~ ADD_CRE_ITEM ~itemname~ #charge1 #charge2 #charge3 ~Flags~ ~Inventory Slots~ [EQUIP] [TWOHANDED]
EQUIP and TWOHANDED are optional and are only needed when dealing with weapons. I’ll give some examples now.
Example #1:
If I want to add a Ring of Protection to Sir Alynar, here’s what I’d do:
COPY_EXISTING ~alynar.cre~ ~override/alynar.cre~ ADD_CRE_ITEM ~ring06~ #0 #0 #0 ~IDENTIFIED~ ~RRING LRING~
This will add the item to his right ring slot. If that slot is full, it will be added to the left ring slot. If both are full, we move the current right ring (the first slot in the list) to an an empty inventory slot so that the Ring of Protection is still put in it’s appropriate slot.
Example #2:
If I want to add a +2 Longsword to his second weapon slot, but don’t want to equip it, here’s what I’d do:
COPY_EXISTING ~alynar.cre~ ~override/alynar.cre~ ADD_CRE_ITEM ~sw1h06~ #0 #0 #0 ~IDENTIFIED~ ~WEAPON2~
Again, if there happens to be a weapon in his second weapon slot already, then we simply move that weapon to an empty inventory slot.
Example #3:
Same deal as example #2 except now I want to equip the sword.
COPY_EXISTING ~alynar.cre~ ~override/alynar.cre~ ADD_CRE_ITEM ~sw1h06~ #0 #0 #0 ~IDENTIFIED~ ~WEAPON2~ EQUIP
This will put the sword in his second weapon slot and equip it. And again, if there’s already a weapon in that slot, it will be moved to an empty inventory slot.
Example #4:
If you want to add a two-handed weapon and equip it, here’s what you do:
COPY_EXISTING ~alynar.cre~ ~override/alynar.cre~ ADD_CRE_ITEM ~sw2h01~ #0 #0 #0 ~NONE~ ~WEAPON2~ EQUIP TWOHANDED
This will place the item in the second weapon slot (and again, if there’s already a weapon there it’s moved to inventory), remove anything that’s in the shield slot and put it in an empty inventory slot and finally it will equip the weapon.
If you just want to add a two-handed weapon to a cre but don’t need it equipped, then you don’t have to worry about specifying EQUIP or TWOHANDED.
Finally, it is worth noting that any WeiDU variables inside %s in the itemname will be replaced by their values.
Phew, that’s it.
Note: starting from v189, you can use INV, QITEM, QUIVER, WEAPON and RING as shortcuts respectively to INV1 INV2 ... INV15 INV16, QITEM1 QITEM2 QITEM3, QUIVER1 QUIVER2 QUIVER3, WEAPON1...WEAPON4, RRING LRING.
In IWD2, INV unfolds to INV1..INV24, SHIELD to SHIELD1..4.
In PST, INV unfolds to INV1..20, TATTOO to TATTOO1..3, EARRING to EARING1..2, QITEM to QITEM1..5, QUIVER1..5. PST also supports HAND, but loses SHIELD and HELMET.
This tutorial was thoughtfully provided by Rastor.
A customized installer can make your mod look more unique to your end user, giving a feel that “This isn’t your basic, cookie-cutter mod.” WeiDU actually allows for a great deal of customization of the mod installer program (ie. the setup-mymod.exe file). For the most part, just about every line of text can be changed to suit your needs. This tutorial will show you how.
Understanding the Installer. WeiDU, like most computer programs, draws every line of text that it displays from a series of strings. In a default installer (simply a renamed weidu.exe and the appropriate .tp2), all of the strings that WeiDU displays are programmed into WeiDU’s source code. Fortunately, weidu allows you to overwrite these strings with ones of your own choosing.
prompts.tra. In the WeiDU download package (from http://www.weidu.org), you will find a file in the examples folder called prompts.tra. This is the file that you modify to customize your installer. Go look at it now.
Upon loading, weidu always looks to see if a prompts.tra file exists in the folder that corresponds to the language that the user selects. If so, then weidu automatically loads the file and uses it to display your strings instead of the ones that are normally coded into it. In order to change what the setup-mymod.exe file displays to the user, you need to edit this file and then place it into the relevant language folder. Your translators should then translate this into whatever language and include it in that language folder.
An Example. Let us assume for a minute that you want the question asking if users want to install to read:
You are a big stupid moron if you do not install the [Component Name]. Install? [Y]es you idiot or [N]o, bonehead or [Q]uit
For obvious reasons, you would never want your mod’s installer to actually say that but it will serve as a good working example.
To change the install component text, find the section of prompts.tra that
corresponds with the text that you wish to change. In this case, we want
to change @-1006 and @-1008. If you look at these lines, you will
see a strange
n that does not actually appear when the installer is
running. This is a carriage return (or “new line”) character. It simply
indicates that you want any text following it to be inserted below the text
that precedes the carriage return.
The name of the component in your mod is always inserted immediately following @-1006 without any spaces. This means that you are going to want to include a space or a bracket or some other distinguishing character at the end of your new text, otherwise the installer’s text will look quite strange. The name of the component also is always followed (with no spaces) by either line @-1007 or @-1008 depending on whether the component is already installed or not.
If you want to include more text in your custom line before the [I]nstall or
[N]ot Install stuff, then simply place it before the
n character.
To generate the custom text that I presented above, change the lines in your prompts.tra file to read:
@-1006= "You are a big stupid moron if you do not install the [" @-1008= "]. Install?\n[Y]es you idiot or [N]o, bonehead or [Q]uit"
After making these changes, place prompts.tra into the TRA folder.
If you are familiar with the default text in the installer, then you will
recognize the all of the lines that are listed in prompts.tra. You may
change any of the text, but simply remember the meaning of the carriage
return. The WeiDU installer will not word wrap, meaning that if you insert
a very large line that is not broken up by
n, the user will not be able
to see all of it. You should do some experimentation to determine the
optimal length for each line (it’s roughly 72 characters).
What you cannot change. Although WeiDU does allow for a large degree of customization, there are a few things that you cannot change. You cannot change the input keys used to collect the player’s response. You cannot change the Copying one file and Compiling dialogues (and the similar) text.
Normally I would start off with don’t use this feature, but since this patch can basically only be used to do really obscure things, you’re probably already a master if you’re considering it.
We’ll use INNER_ACTION in the context of a hypothetical mod that causes all monsters to drop special “skins” that can be worn by the PCs, granting the PCs all of the “natural talents” (e.g., fire resistance, seeing through invisibility, animation avatar) of that monster.
First, we will specially prepare a “template” object -- a special blank “skin” item that has slots equipped effects like “fire resistance bouns +0” and “natural armor class 10” and “change creature animation”. Each creature “skin” will be a specialized copy of this template.
Then we’ll consider each monster in turn. We’ll read the fire resistance value from the monster and write that value into a copy of the “skin” item. Then we’ll repeat the process for “armor class”, “cold resistance”, etc. We’ll also copy over any “CRE effects” the creature might have (e.g., “can see invisible”) and make them additional equipped effects of the “skin” item. Finally we’ll add that personalized “skin” item to the monster’s inventory. If the monster is C6BODHI.CRE, we’ll call the skin C6BODHI.ITM.
We’ll give the code in pieces. First, we want to go over every existing creature and make sure that CREATURE.ITM doesn’t already exist in the game. If it does we won’t make a “skin” for that monster. We’ll also include a little size sanity check to make sure that we’re not ready any degenerate 0-byte files.
COPY_EXISTING_REGEXP GLOB ~.*\.cre~ ~override~ PATCH_IF (SOURCE_SIZE > 0x2c8) AND (NOT FILE_CONTAINS_EVALUATED("%SOURCE_RES%.ITM" "ITM")) THEN BEGIN // %SOURCE_RES%.ITM does not exist.
See SOURCE_RES. Using "ITM" as the second argument makes FILE_CONTAINS_EVALUATED behave like FILE_EXISTS_EVALUATED, since any real item contains ITM as part of its header.
Now we’re going to read the animation, armor class, resistances, and creature effects. We’ll also equip this creature with its special skin, even through we haven’t made that skin yet.
READ_SHORT 0x28 anim_id READ_SHORT 0x46 natural_ac READ_ASCII 0x59 resists (11) // read 11 bytes (even 0's) READ_LONG 0x2c4 eff_off READ_LONG 0x2c8 num_eff READ_ASCII eff_off effects (num_eff * 264) ADD_CRE_ITEM ~%SOURCE_RES%~ #0 #0 #0 ~IDENTIFIED~ ~INV10~
Since there are 11 resistances (e.g., fire, cold, magic fire, ...) we are just leaving them in a string as an array of bytes rather than making up individual variables for each one. Now we want to go about creating our special skin item. We’ll use the mysterious INNER_ACTION to do so.
INNER_ACTION BEGIN COPY ~foo/footemp.itm~ ~override/%SOURCE_RES%.ITM~ WRITE_SHORT 0x76 anim_id WRITE_SHORT 0xa6 natural_ac
Using SOURCE_RES as part of a computed COPY destination is normally somewhat dangerous because it is reset every time you enter a COPY -- but we’re OK here.
Our template footemp.itm item has a number of pre-made equipped abilities (e.g., no dispel, bypass resistance, apply 100% of the time to the wearer, last while equipped) that are just lacking concrete values. Here we fill in the animation and armor class. Next we’ll do the resistances:
// copy over all resistances FOR (i=0; i<11 ; i=i+1) BEGIN INNER_PATCH "%resists%" BEGIN READ_BYTE i resist_i END foo_resist_off = 0xd6 + (i * (0x102 - 0xd2)) WRITE_LONG foo_resist_off resist_i END
For each of the 11 resistances we remember its value from the creature by using INNER_PATCH to get it out of our resists “array”. The resistance effects in footemp.itm are stored in the same order as the creature resistances are in a CRE file.
Having handled the standard resistances we now turn to creature effects. We don’t know in advance how many creature effects there will be, so for each one we will make a new effect structure and insert it into our skin item. We’ll get the new effect structure by copying one of the ones that was already there (called dummy_eff in the code below).
Unfortunately, creature effects are not stored in quite the same way as item effects, so we can’t copy them over directly. Instead, from each creature effect we’ll extract the opcode, value and resource fields. We end up with a three step process: (1) insert some bytes into the skin file to hold the new effect, (2) copy over the dummy_eff effect structure as a framework, and (3) fill in the appropriate values from the creature effect. Here it is:
// copy all CRE effects PATCH_IF num_eff > 0 THEN BEGIN READ_ASCII 0x72 dummy_eff (0x30) FOR (i=0; i<num_eff; i=i+1) BEGIN INNER_PATCH "%effects%" BEGIN READ_ASCII ((i * 264) + 8) "cre_eff_opcode" (0x4) READ_ASCII ((i * 264) + 20) "cre_eff_value" (0x8) READ_ASCII ((i * 264) + 40) "cre_eff_res" (0x8) END INSERT_BYTES 0x72 0x30 // step 1 WRITE_EVALUATED_ASCII 0x72 "%dummy_eff%" // step 2 WRITE_EVALUATED_ASCII 0x72 "%cre_eff_opcode%" // step 3 WRITE_EVALUATED_ASCII 0x76 "%cre_eff_value%" // step 3 WRITE_EVALUATED_ASCII 0x86 "%cre_eff_res%" // step 3 END READ_SHORT 0x70 num_global_itm_eff WRITE_SHORT 0x70 (num_global_itm_eff + num_eff) END END
Huzzah. Notice our use of variables to “pass information” between the main copying action and the INNER_ACTION. Yes, I hate WeiDU syntax as well. Why do you ask? Anyway, after that we update the item header to take into account the new global effects.
END BUT_ONLY_IF_IT_CHANGES
Finally we close off our sanity-checks and throw in the ever-popular BUT_ONLY_IF_IT_CHANGES. It would be “easy” (read: annoying and time-consuming but possible) to extend this hackery so that it also stole all of the effects from undroppable items held by creature (e.g., most “undead” immunities are actually stored in RING95.ITM and not as creature effects) by nesting yet-another INNER_ACTION. You could even be selective and avoid copying over effects like “minimum hit points”. Then you could perhaps create an on-the-fly description for the “skin” item using SAY_EVALUATED.
Bonus points if you actually understood this code. Additional bonus points if you find a real use for INNER_ACTION in your mod. As a closing warning, do not try to re-invent the WeiMorph wheel using a TP2 script -- email Japheth instead.
SET_2DA_ENTRY_LATER and SET_2DA_ENTRIES_NOW are an extension of the SET_2DA_ENTRY patch, so you are encouraged to study the SET_2DA_ENTRY tutorial by Idobek before reading this one.
The main difference between the standard SET_2DA_ENTRY and these new variants is that S_2_E changes the file each instance, while the LATER variant stores a list of changes in the memory and then the NOW part flushes them in one go, resulting faster processing than S_2_E. According to Weimer’s calculations, using LATER/NOW gives an edge when 5 or more S_2_E are to be applied to the same file.
We’ll begin with a side-by-side confrontation of the code, for a mod which allows Thieves, Cleric/Thieves and Mage/Thieves to use 3 proficiency points in Two Weapon Fighting:
using S_2_E:
COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY 34 7 1 3 SET_2DA_ENTRY 34 39 1 3 SET_2DA_ENTRY 34 40 1 3
using S_2_E_L/N:
COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 7 3 SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 39 3 SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 40 3 SET_2DA_ENTRIES_NOW ~_#_#_#weapprof~ 1
The things to change are the following:
~_#_#_#weapprof~
) before other parameters.
You can also split S_2_E_L sessions across multiple parts - but it is easy to shoot yourself in the foot:
// this is OK and works COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 7 3 SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 39 3 DO_SOMETHING_ELSE COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 40 3 SET_2DA_ENTRIES_NOW ~_#_#_#weapprof~ 1 // this, however, is a bug: COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 7 3 /* (*) */ DO_SOMETHING_ELSE COPY_EXISTING ~kitlist.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 40 3 SET_2DA_ENTRIES_NOW ~_#_#_#weapprof~ 1 // BUG: the (*) action is processed to kitlist.2da, not to weapprof.2da. // you can use this to scare off newbies if you're mean.
Now, the formalities - which could be useful to avoid hopelessly crashing of your tp2.
SET_2DA_ENTRY_LATER ~string~ row col value
does the following variable associations (%i% is a counter which increases with
each S_2_E_L instruction):
SET string%i%r = row
SET string%i%c = col
SPRINT string%i%v value
SET_2DA_ENTRIES_NOW ~string~ row_count
begins scanning all variables of the form string%i%r, string%i%c and
string%i%v which I spoke about earlier, applies their changes and erases
all those variables (so you can use the same string again).
So it is possible to cheat and/or introduce bugs:
COPY_EXISTING ~weapprof.2da~ ~override~ SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 7 3 SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 39 3 SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 40 3 SET _#_#_#weapprof1v = 5 // BUG: you're doing SET_2DA_ENTRY_LATER ~_#_#_#weapprof~ 34 39 5 (and not 3) SET_2DA_ENTRIES_NOW ~_#_#_#weapprof~ 1
SUGGESTION: prefix your S_2_E_L/N strings (which can accept %variables%,
by the way) with a standard prefix (and don’t use variables starting with that
prefix), to avoid accidental overwrites. I like to use, as a string,
_#_#_#FILE_NAME
where FILE_NAME is the name of the file I’m editing.
READ_2DA_ENTRIES_NOW and READ_2DA_ENTRY_FORMER are a derivation of READ_2DA_ENTRY pretty much the same way that SET_2DA_ENTRY_LATER and SET_2DA_ENTRIES_NOW are a derivation of SET_2DA_ENTRY, so you are advised to be familiar with how READ_2DA_ENTRY, SET_2DA_ENTRY_LATER, and SET_2DA_ENTRIES_NOW work. Ready? Let’s go!
Here’s a comparison of a R_2_E_N/F with R_2_E:
using R_2_E:
COPY_EXISTING ~kitlist.2da~ ~override~ READ_2DA_ENTRY 0 1 9 somevar // somevar will be "RESERVE"
using R_2_E_N/F:
COPY_EXISTING ~kitlist.2da~ ~override~ READ_2DA_ENTRIES_NOW ~_#_#_#read_kitlist~ 9 READ_2DA_ENTRY_FORMER ~_#_#_#read_kitlist~ 0 1 somevar // somevar will be "RESERVE"
If you have read the tutorial about SET_2DA_ENTRY_LATER and SET_2DA_ENTRIES_NOW you should already be familiar with the way to translate this, so I’ll only explain the technicalities.
READ_2DA_ENTRIES_NOW string col_count
will store inside the variable %string_row_col%
the contents of the
cell at row %row% and coloumn %col%, for all rows
with at least col_count
items.
The variable %string%
will be set to the number of large enough
rows (so, in effect, you also have processed COUNT_2DA_ROWS freely).
READ_2DA_ENTRY_FORMER string row col variable
will store inside variable
the contents of %string_row_col%
which should be the content of the cell row col col_count
as defined
by READ_2DA_ENTRY, unless the user edited the variable on his own.
Here, col_count
is defined by READ_2DA_ENTRIES_NOW.
Differently from SET_2DA_ENTRIES_NOW, neither of these two patches will clear variables from former READ_2DA_ENTRIES_NOW, so be extra-careful in choosing the string to be unique.
As you can see, I use _#_#_#read_FILE_NAME
as a string (the read_
part is to proactively avoid possible overlappings with
SET_2DA_ENTRY_LATER, while the WeiDU code also makes sure that these
problems to not occur accidentally).
Macros were originally intended as a way to repeat a certain tp2 truncon with small or no changes without resorting to copy-and-paste.
First, disclaimers: it’s very easy to shoot yourself in the foot, since
there aren’t any limits to the variable scoping of macros. Due to this reason, this tutorial
will skip on most basic stuff - if you’re going to consider this, you should
already be a master of patching.
Also, for brevity, I’ll give only examples involving tp2 actions. You can
do everything you’re about to read here by substituting DEFINE_ACTION_MACRO
and LAUNCH_ACTION_MACRO with DEFINE_PATCH_MACRO and LAUNCH_PATH_MACRO. There
is no change in the treatment of LOCAL_SPRINT/LOCAL_SET.
For starters, you might want to define a macro that prints "Hello World!" to the screen.
// at the tp2 flag level, IE after AUTHOR but before components. DEFINE_ACTION_MACRO ~HW~ BEGIN PRINT ~Hello World!~ END /* ... */ // a component BEGIN ~Print Hello World and do nothing~ LAUNCH_ACTION_MACRO ~HW~
So, the syntax is quite simple:
DEFINE_ACTION_MACRO name-of-the-macro BEGIN local variables list action list END
will bind the action listed here with name-of-the-macro. We’ll deal with local variables later. After that, LAUNCH_ACTION_MACRO name-of-the-macro will work like if you copy-pasted there the action list defined in DEFINE_ACTION_MACRO.
Well, suppose you wanted to say ’Hello Jon!’ or ’Hello Jack!’ or any other name, rather than only ’Hello World!’. Do you have to define a macro for each of these? Of course not, since all variables defined before the LAUNCH_ACTION_MACRO instruction is called are not cleared, and so can be read (or written) during execution of the macro.
// before calling, have the string ~HW_Name~ contain the name of the person // to greet. DEFINE_ACTION_MACRO ~HW~ BEGIN PRINT ~Hello %HW_Name%!~ END /* ... */ OUTER_SPRINT ~HW_Name~ ~Jon~ LAUNCH_ACTION_MACRO ~HW~ OUTER_SPRINT ~HW_Name~ ~Jack~ LAUNCH_ACTION_MACRO ~HW~
Variables inside macros that are changed keep their new value once the macro has finished being computed. This is bad and good at the same time. For example, you might want to calculate a factorial:
// set ~%factorial_index%~ to the factorial you'd like to compute, // ~%factorial_result%~ to 1, // keep a backup copy of ~%factorial_index%~ // ~%factorial_result%~ will yield the resulting factorial. DEFINE_ACTION_MACRO ~factorial~ BEGIN ACTION_IF ~%factorial_index%~ != 1 THEN BEGIN OUTER_SET ~factorial_result~ = ~%factorial_result%~ * ~%factorial_index%~ OUTER_SET ~factorial_index~ = ~%factorial_index%~ - 1 LAUNCH_ACTION_MACRO ~factorial~ END END /* ... */ OUTER_SET ~factorial_index_old~ = 5 OUTER_SET ~factorial_index~ = 5 OUTER_SET ~factorial_result~ = 1 LAUNCH_ACTION_MACRO ~factorial~ PRINT ~%factorial_index_old%! = %factorial_result%~
TThe fact that variables are kept through the entire tp2 execution (formally,
that they are global) is good, since you have a way of ’remembering’ the result;
however, the bad part is that it may happen that variables are changed even if you
didn’t want them to.
Note: risks grow exponentially as you use macros by somebody else.
By the way, another (rarely) useful feature of macros used here is that they can call themselves (formally, they are recursive).
To solve the problem with variables being changed inside a macro, I provided a rudimentary form of scoping: the aforementioned local variables list part.
Note: if ’global’ and ’local’ variables mean nothing to you (for example, if you don’t have experience in programming in real life languages) you might have problems in understanding the logic beneath the rest of this tutorial. You’ll do fine with the tutorials about standard macros.
Basically, at the beginning of the macro declaration, you may use
LOCAL_SET name = value LOCAL_SPRINT variable string
to declare a variable as local, and assign it a new value. Use this for all variables that don’t need to be changed by the macro. After computing the macro, these variables are set to the value they had before the macro was processed. They work for both actions and patches. For example, here is a corrected factorial macro (also converted to patch):
// set ~%tb#factorial_index%~ to the factorial you'd like to compute, // ~%tb#factorial_result%~ to 1, // ~%ftb#actorial_result%~ will yield the resulting factorial. DEFINE_PATCH_MACRO ~tb#factorial~ BEGIN LOCAL_SET ~tb#factorial_index~ = ~%factorial_index%~ PATCH_IF ~%tb#factorial_index%~ != 1 THEN BEGIN SET ~tb#factorial_result~ = ~%tb#factorial_result%~ * ~%tb#factorial_index%~ SET ~tb#factorial_index~ = ~%tb#factorial_index%~ - 1 LAUNCH_PATCH_MACRO ~tb#factorial~ END END /* ... */ SET ~tb#factorial_index~ = 5 SET ~tb#factorial_result~ = 1 LAUNCH_PATCH_MACRO ~tb#factorial~ PATCH_PRINT ~%tb#factorial_index%! = %tb#factorial_result%~
For your convenience, WeiDU includes some standard macros for you to use. See Macros Listing for a listing of all macros and how to invoke them in detail. Here I’ll describe the standard syntax.
In our example, we’ll use the tb#factorial macro (which is the same as the last example in the preceeding tutorial).
Description (from Macros Listing):
tb#factorial
: computes the factorial of a number.
This is a PATCH macro.
tb#factorial_index
to the factorial you’d like to compute.
tb#factorial_result
to 1;
tb#factorial_result
.
Read the instructions on the variables to set:
SET ~tb#factorial_index~ = 5 SET ~tb#factorial_result~ = 1 LAUNCH_PATCH_MACRO ~tb#factorial~ PATCH_PRINT ~%tb#factorial_index%! = %tb#factorial_result%~
(since you’ve read the warnings from the previous tutorial, standard macros make use of LOCAL_SET/SPRINT to avoid overwriting variables that shouldn’t change).
Another example:
tb#fix_file_size
: overwrites all files matching a certain regexp with a
certain standard file, if they are under a certain size.
This is an ACTION macro.
tb#fix_file_size_min
to the maximum allowed size
(files whose size is exactly this value are NOT overwritten).
tb#fix_file_size_target
to the standard file used to replace file(s) that match the regexp.
tb#fix_file_size_category
to a descriptive name.
tb#fix_file_size_regexp
to the regexp for the file(s) to be checked for a match.
So, to overwrite all item files that are smaller than 0x72 with sw1h01.itm (for example the corrupted iplot*.itm files)
OUTER_SET tb#fix_file_size_min = 0x72 OUTER_SPRINT tb#fix_file_size_target "SW1H01.ITM" OUTER_SPRINT tb#fix_file_size_category "items" OUTER_SPRINT tb#fix_file_size_regexp "^.*\.itm$" LAUNCH_ACTION_MACRO ~tb#fix_file_size~
The GROUP flag came about in discussing component management with an abnormally large number of independent components, e.g. BG2 Tweaks. Previously WeiDU lacked a satisfactory way to organize such a modthe closest feature would be the top level ASK_EVERY_COMPONENT, which only allows for a single all-or-none approach.
The solution is a new component flag, GROUP, with the following syntax (and example):
Code:
GROUP string Code: BEGIN ~100% Learn Spells~ GROUP ~Convenience Tweaks~ // component code BEGIN ~Identify All Items~ GROUP ~Convenience Tweaks~ // component code BEGIN ~Give Edwin his BG2 Stats~ GROUP ~NPC Tweaks~ // component code BEGIN ~ Give Jaheira her BG2 Stats~ GROUP ~NPC Tweaks~ // component code
Upon installing the mod, the player is now presented with meta-options on each group at a high level:
Code:
Would you like to display the category [Convenience Tweaks]? [Y]es/[N]o Would you like to display the category [NPC Tweaks]? [Y]es/[N]o
Selecting [N]o
on any group suppresses those options from being
displayed, leading to a simpler and more controlled installer experience for
the player. In the provided example, selecting [N]o
to Convenience
Tweaks and [Y]es
to NPC Tweaks would result in WeiDU starting
installation by asking to install the Give Edwin his BG2 Stats component.
GROUP operates independently of SUBCOMPONENT, meaning you can use both to organize the mod as needed. A few other items of note:
BEGIN ~1~ GROUP ~1~ BEGIN ~2~ GROUP ~2~ BEGIN ~123~ GROUP ~1~ GROUP ~2~ GROUP ~3~the 3 group will not be displayed. To workaround, add a dummy component with only that GROUP:
BEGIN ~1~ GROUP ~1~ BEGIN ~2~ GROUP ~2~ BEGIN ~123~ GROUP ~1~ GROUP ~2~ GROUP ~3~ BEGIN ~Dummy component because it's too hard to fix this in WeiDU~ DEPRECATED ~Dummy component~ GROUP ~3~
The array construct (for lack of a better term) can be used to create with ease arrays at the install time. The construct is formed as follows:
$variable(modifier list)
It is equivalent to saying
~variable_%modifier1%_%modifier2%_etc.~
When it is on the right handed side of a SPRINT or SET statement, the contents of a variable with the resulting name, if existing, is used in place of the variable name. This does not happen on the left handed side (this is good for you).
Example dummy use:
OUTER_SPRINT ~var_1_sbar~ ~300~ OUTER_SPRINT ~var_1~ ~bar~ OUTER_SPRINT $asd(doo) ~f00~ // equivalent to OUTER_SPRINT ~asd_doo~ ~f00~ OUTER_SET $s(x) = 4 // equivalent to OUTER_SET ~s_x~ = 4 OUTER_SET one = 1 OUTER_SPRINT two ~sbar~ OUTER_SET ~x~ = $var("%one%" "%two%") // equivalent to OUTER_SET ~x~ = ~var_%one%_%two%~ // which is again equivalent to OUTER_SET ~x~ = ~var_1_sbar~ (which is 300) OUTER_SPRINT ~y~ $var("%one%") // equivalent to OUTER_SPRINT ~y~ (the contents of variable ~var_%one%~/~var_1~ ) OUTER_SPRINT ~z~ $asd(doo) // equivalent to OUTER_SPRINT ~z~ (the contents of variable ~asd_doo~ ) OUTER_SPRINT ~t~ $miss(ing) // equivalent to OUTER_SPRINT ~t~ (the contents of variable ~miss_ing~ // there's no variable "miss_ing", so we default to the "miss_ing" text instead ) OUTER_SET ~f~ = $s(x) // equivalent to OUTER_SET ~f~ = ~s_%x%~ OUTER_SPRINT $var(x) ~p~ // equivalent to OUTER_SPRINT ~var_x~ ~p~ OUTER_SPRINT $var(x) ~y~ // equivalent to OUTER_SPRINT ~var_x~ ~y~ // notice that the ~var_x~ variable is not checked, so this doesn't become OUTER_SPRINT ~p~ ~y~ PRINT ~"%x% %y% %z% %t% %f% %var_x%" should yeld "300 bar f00 miss_ing 4 y"~
We expect somebody to actually write a mod which uses this form more seriously this construct and use their code as a more useful example.
The MODDER flag allows for more debugging info to be reported. It is suggested to leave MODDER enabled while you’re developing your mod and then disable it before you ship it to the players. However, given different coding styles, it is possible that something gets interpreted as a critical bug for some users and as expected behavior for others. For this reason, it’s possible to fine-tune the reporting level for some of the bugs that are found. This is done by writing additional modifiers after your MODDER statement, for example:
MODDER setup_tra fail area_variables none missing_extern warn
Each option can be set to either NONE, WARN or FAIL with the above syntax, with the obvious meaning. If none are chosen, WARN is assumed. WARN may generate false positives. In this case, "INSTALLED WITH WARNINGS" is not displayed.
Here the list of feedback options that you can configure, along with use cases when you should deactivate them. In all other cases, using FAIL or WARN is a case of personal preference (for example, use WARN early in the development process and FAIL when you’re near the beta status).
*Global*("varname","scope")
statement is six characters long, but not GLOBAL, MYAREA, LOCALS, or a valid area name.As you can see, all of these are catching either really stupid coding practices (yes, biffing ARE files is really stupid) or are avoidable with minor code restructuring, but with some people these kind of things are enough to start flame wars, so here’s a way to choose what things I should let through. *shrugs*
This tutorial has been brought to you by plainab.
The goal of this tutorial is to explain the use of READLN within the tp2 structure. READLN can be used as an action or as a patch. My experience has been in using it as an action command. I will describe the usage of READLN and then give you some examples of its use. To put it bluntly READLN stands for "read line" and it’s use causes WeiDU to pause and read any information entered by the user on the interface screen and to then store that information within the given variable.
Syntax:
ACTION_READLN ~variable~ PATCH_READLN ~variable~
Lets build an example from scratch:
First we need to have something we want to change. For this example we will use the stack amount for item files. We could pick a high number and simply use:
COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ WRITE_SHORT 0x38 9999 //stack value BUT_ONLY_IF_IT_CHANGES
Or we can let the end user choose how many items they want in their stacks. To do that we first need to ask the user a question.
PRINT ~How many of one item do you want to be able to be stacked?~
If we leave it at this, weidu will display the question, but keep running. We have to make weidu pause and that is where READLN comes in. So we now have:
PRINT ~How many of one item do you want to be able to be stacked?~ ACTION_READLN ~new_stack~
We can then take this entered value and use it on our copy block.
So the usage in this example would be:
BEGIN ~Modify the stack value of all items~ PRINT ~How many of one item do you want to be able to be stacked?~ ACTION_READLN ~new_stack~ COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ WRITE_SHORT 0x38 %new_stack% //stack value BUT_ONLY_IF_IT_CHANGES
You could leave it at this and it will work, but I will tell you now that the end user can input any character from their keyboard. READLN doesn’t care what is entered, but weidu may have problems if it is expecting an integer rather than text. To get around that issue we need to add a loop that will re-ask the question when the user enters the incorrect type of information.
BEGIN ~Modify the stack value of all items~ PRINT ~How many of one item do you want to be able to be stacked?~ ACTION_READLN ~new_stack~ OUTER_WHILE NOT(IS_AN_INT %new_stack%) BEGIN PRINT ~How many of one item do you want to be able to be stacked?~ ACTION_READLN ~new_stack~ END COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ WRITE_SHORT 0x38 %new_stack% //stack value BUT_ONLY_IF_IT_CHANGES
The loop we added looks to see if the entered value is an integer and if the entered value is not an integer then it will re-ask the question. This loop will go indefinitly until the user inputs the correct type of information. Should your user not understand why they keep getting asked the question you may wish to add new text to your questions to help them to understand what it is you are looking for. As in this example:
BEGIN ~Modify the stack value of all items~ PRINT ~How many of one item do you want to be able to be stacked? Please enter your answer as an integer.~ ACTION_READLN ~new_stack~ OUTER_WHILE NOT(IS_AN_INT %new_stack%) BEGIN PRINT ~Your answer was not an integer. Please enter your answer as an integer. How many of one item do you want to be able to be stacked?~ ACTION_READLN ~new_stack~ END COPY_EXISTING_REGEXP ~.*\.itm~ ~override~ WRITE_SHORT 0x38 %new_stack% //stack value BUT_ONLY_IF_IT_CHANGES
That is the basics on asking for information from the user and then applying that information later on in the install process. PATCH_READLN works just like ACTION_READLN only it’s location of use within the tp2 is different.
Another use of READLN is to create options that the user can choose from.
Top level options are taken care of by the use of SUBCOMPONENT and/or GROUP, but since
WeiDU cannot nest subcomponents, you can use READLN to offer additional choices beneath
the first set of subcomponents. Here’s an example of a finished product where READLN is
used to offer additional choices to a subcomponent. The relevant parts to this discussion
were taken from iiProjectileRetrievalMod. I helped to create this portion of the mod.
It works. I have however since learned of better methods that would eliminate some of the
READLN usage from this example.
BEGIN ~Easy TUTU~ SUBCOMPONENT ~Projectile Retrieval Mod~ INCLUDE ~iiprojectiler/E22.tph~ // Easy TUTU version of mod BEGIN ~All other Infinity Engine Games~ SUBCOMPONENT ~Projectile Retrieval Mod~ PRINT ~Please tell me how you'd like to modify your files. 1. I want to modify ALL thrown/projectiles at once. 2. I want to modify each thrown/projectile one at a time. PLEASE ENTER 1 OR 2 ~ ACTION_READLN how OUTER_WHILE NOT(IS_AN_INT %how%) || (%how% > 2) || (%how% < 1) BEGIN PRINT ~Please tell me how you'd like to modify your files. 1. I want to modify ALL thrown/projectiles at once. 2. I want to modify each thrown/projectile one at a time. PLEASE ENTER 1 OR 2 ~ ACTION_READLN how END ACTION_IF (%how% = 1) THEN BEGIN PRINT ~Please choose how you'd like ALL your projectiles retrieved. 1. Plain projectile only -> Plain projectile only 2. All projectiles -> Plain projectile only 3. Each projectile -> Each projectile 4. Magical projectiles -> Reduced Magical/Plain projectiles PLEASE ENTER 1, 2, 3, OR 4 ~ ACTION_READLN choice OUTER_WHILE NOT(IS_AN_INT %choice%) OR (%choice% > 4) OR (%choice% < 1) BEGIN PRINT ~Please choose how you'd like ALL your projectiles retrieved. 1. Plain projectile only -> Plain projectile only 2. All projectiles -> Plain projectile only 3. Each projectile -> Each projectile 4. Magical projectiles -> Reduced Magical/Plain projectiles PLEASE ENTER 1, 2, 3, OR 4 ~ ACTION_READLN choice END
The only drawback to using READLN when offering additional choices to the user is that there is no record within the weidu.log file as to which choices were made. Should a problem arise during game play it may be difficult to determine which choice is causing the problem, especially if the player has forgotten which choices they made. This could be solved by creating your own log file, but that is another tutorial...
Macros can have unintended side effects because they change the global variable environment. You have to be extremely careful when interacting with the “outside”.
That’s why functions were introduced in v209. Any changes to the variables are only visible inside the function. You can also provide default arguments and return values.
As with macros these functions come in an ACTION and a PATCH variant. So the corresponding commands are: DEFINE_ACTION_FUNCTION / DEFINE_PATCH_FUNCTION (both actions) and LAUNCH_ACTION_FUNCTION (action) / LAUNCH_PATCH_FUNCTION (patch).
Before you can use a function you have to DEFINE it:
DEFINE_PATCH_FUNCTION count_spells INT_VAR level = 1 STR_VAR type = "WI" RET num BEGIN SET num = 0 PATCH_IF (~%SOURCE_EXT%~ STRING_EQUAL_CASE ~CRE~) THEN BEGIN // do some magic here that actually counts the number of spells and sets // the variable num to that number. END END
After that you can LAUNCH it:
COPY_EXISTING ~my.cre~ ~override~ // count number of level 1 wizard spells (default arguments of the function) LAUNCH_PATCH_FUNCTION count_spells RET lvl1_wizard = num END // and now the level 5 priest spells SET level = 5 SPRINT type = "PR" LAUNCH_PATCH_FUNCTION count_spells RET lvl5_priest = num END PATCH_PRINT ~This CRE has %lvl1_wizard% level 1 wizard spells and %lvl5_priest% level 5 priest spells~
Notice that arguments are passed indirectly to the function. It inherits the variable environment from the caller. So when you launch it, all variables that are visible at that point are also visible from within the function.
You can provide default values for variables with INT_VAR (integers) and STR_VAR (strings). If the variable was not defined in the caller’s environment the default value is used.
Any variables you change inside the function will revert back to their old value
after the function finishes.
That’s why you have to specify what values should be return values.
You can do this in the RET part of the function definition (RET num
).
When you launch the function you can use these variables to import the value
into your current variable environment (RET lvl1_wizard = num
).
Separate the variables with a space character if you have multiple return values.
If you don’t need certain features simply omit them:
DEFINE_ACTION_FUNCTION xyz INT_VAR x = 1 y = 2 z = 3 BEGIN PRINT ~x: %x%, y: %y%, z: %z%~ END LAUNCH_ACTION_FUNCTION xyz END
Starting from V210, you can also specify a variable that will be local to the function environment, without being passed back to the calling environment, by adding either STR_VAR or INT_VAR to the LAUNCH bit:
DEFINE_ACTION_FUNCTION xyz INT_VAR x = 1 y = 2 z = 3 BEGIN PRINT ~x: %x%, y: %y%, z: %z%~ END LAUNCH_ACTION_FUNCTION xyz INT_VAR x = 0 END
These macros are automatically shipped with WeiDU.
tb#factorial
: computes the factorial of a number.
This is a PATCH macro.
tb#factorial_index
to the factorial you’d like to compute.
tb#factorial_result
to 1;
tb#factorial_result
.
tb#fix_file_size
: overwrites all files matching a certain regexp with a
certain standard file, if they are under a certain size.
This is an ACTION macro.
tb#fix_file_size_min
to the maximum allowed size
(files whose size is exactly this value are NOT overwritten).
tb#fix_file_size_target
to the standard file used to replace file(s) that match the regexp.
tb#fix_file_size_category
to a descriptive name.
tb#fix_file_size_regexp
to the regexp for the file(s) to be checked for a match.
The following macros were provided by Gort. Note that, differently from the standalone versions, you need
to set all the needed values - they aren’t initialized automatically. This is untrue in the FUNCTION version
(they are initialized to either 0 or the empty string, except for probability1 which is initialized to 100).
DELETE_SPELL_EFFECT
: deletes all extended effects with specified opcode from a spell.
This is a PATCH macro and function.
DELETE_ITEM_EFFECT
: deletes all extended effects with specified opcode from an item.
This is a PATCH macro and function.
DELETE_ITEM_EQEFFECT
: deletes all equipping effects with specified opcode from an item.
This is a PATCH macro and function.
DELETE_CRE_EFFECT
: deletes all effects with specified opcode from a creature. (Warning: doesn’t check EFF version)
This is a PATCH macro and function.
ITEM_EFFECT_TO_SPELL
: copies all extended effects from the current item to the first extended header of a spell.
This is a PATCH macro and function.
ADD_SPELL_EFFECT
: adds an extended effect to a spell. All variables except probability1 are 0 by default.
This is a PATCH macro and function.
ADD_ITEM_EFFECT
: adds an extended effect to an item.
This is a PATCH macro and function.
ADD_ITEM_EQEFFECT
: adds an equipping effect to an item.
This is a PATCH macro and function.
ADD_CRE_EFFECT
: adds an effect to a creature. All variables except probability1 are 0 by default. (Warning: doesn’t check EFF version)
This is a PATCH macro and function.
DELETE_CRE_ITEM
: deletes all matching items from a creature. Regexp alowed.
This is a PATCH macro and function.
DELETE_STORE_ITEM
: deletes all matching items from a store. Regexp alowed.
This is a PATCH macro and function.
DELETE_AREA_ITEM
: deletes all matching items from an area. Regexp alowed.
This is a PATCH macro and function.
REPLACE_STORE_ITEM
: replaces all matching items in a store with another item. Regexp alowed.
This is a PATCH macro and function.
REPLACE_AREA_ITEM
: replaces all matching items in an area with another item. Regexp alowed.
This is a PATCH macro and function.
ADD_AREA_ITEM
: adds an item to a container of an area.
This is a PATCH macro and function.
ADD_CRE_ITEM_FLAGS
: adds flags to all matching items posessed by a creature. Regexp allowed.
This is a PATCH macro and function.
REMOVE_CRE_ITEM_FLAGS
: removes flags from all matching items posessed by a creature. Regexp allowed.
This is a PATCH macro and function.
SET_CRE_ITEM_FLAGS
: set flags to all matching items posessed by a creature. Regexp allowed.
This is a PATCH macro and function.
READ_SOUNDSET
: reads soundset of a creature and stores it as an array of string references This is an ACTION macro.
WRITE_SOUNDSET
: writes soundset to all matching creatures. Regexp allowed.
This is an ACTION macro.
ADD_AREA_REGION_TRIGGER
: adds an area region to the current are file.
This is a PATCH macro and function. Unlike other macros you do not need to set every variable. You are only required to set those values that you need to write.
The vertex points are seperated by their XY values
You may have as many pairs as you need. Just copy this set and increment the #.
Example: For the X values: ab_RT_Vx_X_0, ab_RT_Vx_X_1, ab_RT_Vx_X_2
For the Y values: ab_RT_Vx_Y_0, ab_RT_Vx_Y_1, ab_RT_Vx_Y_2
NOTE: The original code released in WeiDU v211 has been overhauled and re-released in v212. Therefore, the user defined variable names have been changed. If you used the 211 version you will need to update your mod to use the variables in this version. Sorry for the inconvenience...
If you rename WeiDU.EXE to something of the form Setup-MyMod.exe, it will behave as if the following arguments were present:
To distribute your mod, rename WeiDU.EXE to Setup-MyMod.EXE (or whatever), put Setup-MyMod.TP2 file in the same directory and go! Typically mods are distributed a ZIP files or self-extracting archives that put the EXE, TP2 and module data files in the main BGII directory.
If your mod adds new resources to the game (e.g., via COPY) you should be careful to pick a unique name so that you resource will not conflict with one created by another mod. For example (and I speak from personal experience here), naming a creature BOO2.CRE is just asking for trouble. One common approach here is to pick a special prefix that uses a character not found in normal game resources. For example, I might use the prefix W# and call a sword W#SWORD.ITM and a new spell W#SPELL.SPL. However, it it still possible to run into conflicts, so I recommend that you check out the Community Mod Filename Prefix Reservations project at http://forums.blackwyrmlair.net/index.php?showtopic=113 and register your own unique prefix.
The WeiDU.exe (or Setup-MyMod.exe) process will terminate with a “return value” (or “exit code” or “status code” or “error level”) based on the success or failure of certain operations. Depending on your operating system you can use this value to guide shell scripts or batch files that include WeiDU. In general, WeiDU returns 0 on success and non-zero on failure. Here is a concrete list of return values.
A regular expression or regexp is "somewhat" like a DOS wildcard but not quite. The big difference is that if you would say * in DOS you say .* in regexp-land. Here’s a definition:
The syntax for regular expressions is the same as in Gnu Emacs. The special characters are:
$^.*+?[]\
The following constructs are recognized:
. matches any character except newline * (postfix) matches the previous expression zero, one or several times + (postfix) matches the previous expression one or several times ? (postfix) matches the previous expression once or not at all [..] character set; ranges are denoted with -, as in [a-z]; an initial ^, as in [^0-9], complements the set ^ matches at beginning of line $ matches at end of line \| (infix) alternative between two expressions \(..\) grouping and naming of the enclosed expression \1 the text matched by the first \(...\) expression (\2 for the second expression, etc) \b matches word boundaries \ quotes special characters.
So spe.* matches "sper01.itm" and "sper.eff" and "special".
Hopefully this is understandable to most people.
The following keywords can be used in values and evaluate to the given integer (usually an offset in a CRE file).
// KEYWORD VALUE AREA_CITY 0252 AREA_DAY 0260 AREA_DUNGEON 0256 AREA_FOREST 0248 AREA_NIGHT 0264 ATTACK1 0220 ATTACK2 0224 ATTACK3 0228 ATTACK4 0232 BATTLE_CRY1 0200 BATTLE_CRY2 0204 BATTLE_CRY3 0208 BATTLE_CRY4 0212 BATTLE_CRY5 0216 BIO 0x1cc BORED 0196 COMPLIMENT1 0352 COMPLIMENT2 0356 COMPLIMENT3 0360 CRITICAL_HIT 0424 CRITICAL_MISS 0428 DAMAGE 0236 DESC 0x54 DIALOGUE_DEFAULT 0412 DIALOGUE_HOSTILE 0408 DYING 0240 EXISTANCE1 0444 EXISTANCE2 0448 EXISTANCE3 0452 EXISTANCE4 0456 EXISTANCE5 0460 HAPPY 0172 HURT 0244 IDENTIFIED_DESC 0x54 INITIAL_MEETING 0164 INSULT 0340 INTERACTION1 0320 INTERACTION2 0324 INTERACTION3 0328 INTERACTION4 0332 INTERACTION5 0336 INVENTORY_FULL 0436 LEADER 0188 MISCELLANEOUS 0384 MORALE 0168 NAME1 8 NAME2 12 PICKED_POCKET 0440 REACT_TO_DIE_GENERAL 0376 REACT_TO_DIE_SPECIFIC 0380 RESPONSE_TO_COMPLIMENT2 0388 RESPONSE_TO_COMPLIMENT3 0392 RESPONSE_TO_INSULT1 0396 RESPONSE_TO_INSULT2 0400 RESPONSE_TO_INSULT3 0404 SELECT_ACTION1 0292 SELECT_ACTION2 0296 SELECT_ACTION3 0300 SELECT_ACTION4 0304 SELECT_ACTION5 0308 SELECT_ACTION6 0312 SELECT_ACTION7 0316 SELECT_COMMON1 0268 SELECT_COMMON2 0272 SELECT_COMMON3 0276 SELECT_COMMON4 0280 SELECT_COMMON5 0284 SELECT_COMMON6 0288 SELECT_RARE1 0416 SELECT_RARE2 0420 SPECIAL1 0364 SPECIAL2 0368 SPECIAL3 0372 TARGET_IMMUNE 0432 TIRED 0192 UNHAPPY_ANNOYED 0176 UNHAPPY_BREAKING 0184 UNHAPPY_SERIOUS 0180 UNIDENTIFIED_DESC 0080 HIDDEN_IN_SHADOWS 0444 SPELL_DISRUPTED 0448 SET_A_TRAP 0452 STORE_NAME 12 SCRIPT_OVERRIDE 0x248 SCRIPT_CLASS 0x250 SCRIPT_RACE 0x258 SCRIPT_GENERAL 0x260 SCRIPT_DEFAULT 0x268 DEATHVAR 0x280 DIALOG 0x2cc AREA_SCRIPT 0x94
This section briefly explains some common file formats. The definitive reference is http://iesdp.gibberlings3.net/.
The source code to WeiDU is available under the GNU General Public License, as detailed in the file COPYING. If for some reason you are unable to obtain a copy of the GPL, merely announce that fact on some public forum and your mailbox will be full of copies of it for life. It’s a great way to meet new people.
Since this is the world of Windows, I distribute a pre-compiled binary. WeiDU is written in OCaml, a functional programming language that includes automatic memory management, higher-order functions and efficient native code generation. If these terms mean nothing to you, you probably won’t be able to modify the source code.
However, if you do want to modify the source and then recompile WeiDU, it’s quite easy. Make sure that you have OCaml 3.10 (or newer), sed (or Perl 5.6), make and either gcc or (optionally) cl (the Microsoft Visual C Compiler).
Edit the Configuration file and pick your configuration. If you’re not x86/windows/cygwin or x86/linux, you’ll have to do some tweaking. Then just type make clean and then make. Presto, you’ve recompiled it.
WeiDU uses the GPL’d zlib and xdiff libraries, as well as portions from the BSD’d Elkhound program.
Weimer’s thank you: I would like to thank the fine folks at the Infinity Engine File Format Hacking Project, without which this would not have been possible: http://iesdp.gibberlings3.net/.
In addition, I make heavy use of Near Infinity for general IE mods: http://www.idi.ntnu.no/~joh/ni/.
The Infinity Engine Editor Pro was very good at changing ITMs and SPLs, but DLTCEP is now a better choice: http://www.gibberlings3.net/tools/dltcep.php.
Kudos to my main out-of-house developer:
Special thanks to my main bug-finders *cough* I mean beta-testers:
Special thanks to Greg Henry for being the first person to mention WeiDU to me in a face-to-face conversation. I was quite impressed. Jason Compton is the first person to mention WeiDU to me in a telephone conversation. Scott McPeak is the first person to discuss my modding hobby with me in a face-to-face conversation (and he was quite well-informed).
The Bigg would additionally like to thank the following contributors:
and additionally
WeiDU was originally made by Westley Weimer. I (Valerio Bigiani, AKA The Bigg) decided to mantain the ’WeiDU’ software according to the GPL terms, under which WeiDU is distributed, starting from version 186 (which was a closed beta, while the first public release was the 187 one).
What this means is that all praise and credits should go to Weimer, but all bug reports and feature requests should go to Valerio Bigiani (AKA ’The Bigg’).
Since WeiDU was released under GPL, with an added notice that you are also allowed to distribute unmodified binaries, I assume the same holds true for mods distributed with this version of WeiDU. However, if you feel like being nice add a line in your readme stating that the current WeiDU may be downloaded from <insert link>. Current authorized mirrors include (but may not be limited to) all those from IEGMC, the Dragon’s Hoard Download Centre, and the SHS Download Centre.
Currently undocumented:
|
|
See the file README-WeiDU-Changes.txt for a description of how WeiDU has changed over time.
This document was translated from LATEX by HEVEA.