tes3lint (http://www.gamesas.com/bgsforums/index.php?showtopic=930873) is now in Beta.

tes3lint is a command-line tool for investigating potential problems in TES3 plugins. It is written in Perl and runs natively on Windows or Linux.

Highlights from the http://code.google.com/p/mlox/wiki/Tes3lint:

Features
* Find out if the plugin defines autocalc'ed spells.
* Find out if the plugin has an implicit dependency on Tribunal/Bloodmoon functions because it uses them but does not list the expansion as a Master.
* List Evil GMSTs.
* Find deprecated usage of Leveled List scripting functions on standard Bethesda Lists.
* Find duplicate (i.e.: "dirty/unclean") records that are unchanged from one of the masters.
* Shows duplicate and modified INFO dialogs.
* Find CELLs that trigger the FOGBUG.
* Find any CELLs where fog density in DATA is different from AMBI subrecords.
* Find problematic usage of GetSoundPlaying for detecting events.
* Find dirty junk cells that are apparently injected by some CS bug.
* Find scripts that are unaware of MenuMode.
* Show if plugin is missing Author/Description/Version from plugin header.
* Find duplicate/modified INFOs.
* Finds potentially buggy case where INFO dialogs have had their IDs changed.
* Find all records in the plugin that override records in its Masters.
* Find problematic scripts attached to Bethesda doors.

Here is a http://code.google.com/p/mlox/wiki/Tes3lintSample to get an idea of some of the things it will tell you about a plugin.

The wiki page explains how to http://code.google.com/p/mlox/wiki/Tes3lint#Get_tes3lint and http://code.google.com/p/mlox/wiki/Tes3lint#Installing_and_Running_tes3lint tes3lint.

Congratulations on this release!

* Swiveller *

Congratulations indeed.

Is the new name the hint of a coming tes4lint?

Ha ha Not really. It's just that I wanted to emphasize that this tool was not also usable for TES4 plugins. I'm sure the issues with TES4 plugins are quite different.

Once I get the next 3 or 4 Morrowind tools I have in mind out the door, maybe I'll move on to Oblivion and see what I'd like to do there.

Have you considered flagging up invalid headers? Where the TES3.DATA doesn't match file size of the file defined in the corresponding TES3.MAST?

That would be pretty easily done. I've got a little library of over 2000 plugins installed, so I'll go do a test and see if that comes up with anything.

There are a couple good plugins to do tests on, that I thought I might mention:
MultiMark.esp contains a hexedited GMST that can choke a strict record parser.
morrowland.rar/From Hell To Heaven.esm is almost 70MB, so it's a good stress tester.

Edit: This was mildly interesting: I just tried a check for the number of records in the TES3.HEDR vs the actual count, and quite a few plugins differ. I conclude that particular field is pretty worthless

I've checked in a new flag for this (MASTERS - lists masters out of sync with installed masters), available from svn:
http://mlox.googlecode.com/svn/trunk/util/tes3lint (Right click/Save)

It'll be in the next release.

Another idea: When loading a plugin with the Enchanted Editor sometimes I get this warning:Seem to get this with mods edited by TESAME and those cleaned with TESPCD.

I have that check written in the code now, but I haven't enabled it yet. It ended up not seeming very interesting. It doesn't seem to affect either Morrowind or the CS, they just read all the records in the file anyway. But I could enable it, if you like.

Well, maybe it's just me but I always fix those errors when I come across them :shrug:

Okey dokey, next release will check the actual number of records vs. those recorded in the header.

It is a really strange record that, as it isn't used by Morrowind at all. (That I can make out :shrug:)


(Not saying that it is a bad thing to check for or anything... )

I look forward to the release date (I'm not knowledgeable enough about all this to provide any beta-testing feedback ). Keep up the great work, john.moonsugar! This utility will be a blessing. :tops:

The version of tes3lint in svn now has NUMRECS flag enabled:
http://mlox.googlecode.com/svn/trunk/util/tes3lint

example:

tes3lint -f numrecs "Zyndaar's Bows.esp"Zyndaar's Bows.esp: NUMRECS	[NUMRECS]: number of records in TES3.HEDR (20) != Actual count (8)

John, have you seen abot's / LizTail's recent posts in the MGE thread? They talk about how to determine the end of a script (since the method MGE used previously didn't work on some mods). Might be worthwile to check whether tes3lint uses one of the two suggested working methods (record length check / check for "END" command) or the one the doesn't work on all mods (null-termination).

Alas no, I don't follow MGE, as I haven't gotten it working on Linux yet

I'll go see if I can find what you're talking about. I'm not sure I understand what the issue is to begin with.

One version of MGE assumes that SCPT.SCTX was null termianted. It isn't in all cases.

Sorry, I should've been more clear. I was thinking this:

tes3lint parses scripts in the esp. Scripts are terminated, and after termination other code may follow (for example if the script has been edited), which never gets executed. There are two potential problems with this:

1. If a utility that parses the scripts misses the end of a script, it then it treats the following records as part of the script code. (This is what the MGE beta apparently did.)

2. If the utility continues to read script code after the "END" command, then it may report problems of this code which are effectively false hits (because the respective code will never get executed).

Or just check posts #187 and #188 in the current MGE thread.

Ah, I just read that. Thanks.

tes3lint doesn't care if the scripts are null-terminated or not.

However, abot's comment on script text termination at the first "END" is interesting, I didn't know that people did that, although I think abot mentioned it before, it didn't register in my brain.

I can probably do an okay job of recognizing END, but I'll have to see. I'm using regexps, not a real parser.

Edit: maybe it's time to reverse-engineer the compiled script! Has anyone done that yet?

D'oh, I did make a mistake. It was 36 plugins/65 scripts had non-comment code after the end. I think

Wow, fast work
Not that I think appending comments/previous versions of code after the end is a good practice, still it is something existing and working in some mods, like using Basic like = operator instead of C like ==, or <> instead of != ... MWEdit complains, TESCS happily compiles.
As the number of mods writing something after the first final end seems to be really low, all in all I don't think it is something critical to detect for MGE, but it is indeed good for this specialized tool :thumbsup:

It's possible some of them were mistakes too. It would be easy to overlook a little "end" in the middle of your script if you put it in and forgot, or were using someone else's script as a template. I'll take a closer look and see if I find any that are suspicious. Maybe I should flag these.

Yes, maybe different persons working on scripts and keeping previous version when in doubt.
I just tried tes3lint-0.51 with one I stumbled upon when fixing some scripts (LGNPC_PaxRedoran_v1_12.esp, jo_BolvynCorpseScript script), it does not seem to be detected at the moment,
I am using this line
C:\Perl\bin\perl.exe tes3lint -r "G:\_games\_rpg\Morrowind\Data Files\LGNPC_PaxRedoran_v1_12.esp" > tes3lint.log
am I missing some parameter or is this feature not included in current uploaded version already?

Are you saying it does scan past the End? Or that it does not flag that there is code past the End?
The former is implemented in 0.51 so it should not scan past the End, the latter I was just musing about, haven't actually done anything yet.