In Neeras voice: "half-drow !" :-)

Perhaps because she sees that the PC has the potential to re-acquire spiders bane and is a dangerous foe on top

(while the common folk are not) ? She is certainly not a "good" character, but that does not mean suicidal. I hate that

in the original game the only thing you can do is kill her. Without any context or explanation why she is there in the

graveyard and does what she does. Spiders are of course typically drow, but breeding them under a human city ?

That is interesting and uncommon, Pai'Na is not your typical trash mob ? A missed story and the change fits in my

opinion. There are enough obvious dangling plot hooks left (rune, mind flayers) which cannot be "solved" by a small

modification.

I would agree this is still a bit weak an argument as the original game does not give you any real clue. You probably

really have to check the cre file and then have to retcon something.

Surely a matter of taste. I think this thread is a nice read and thank you for it.

Small comment to Pai'Na. If this web page  contains the information from her original creature file then the modification by unfinished business makes more sense. According to this she is a

Because of that I actually find the change quite nice. I never noticed the logical problem with revealing plot points. On the other hand, she basically lives next doors and might know a bit about the neighbourhood ?

I'm sorry, I could not  resist.

I made another test, absolute minimal packages with stratagems v31. Already this gives an "nOpcode out of bounds", the lucky number this time is 408. Using the same bases packages with stratagems v32.6 on top complains at least about 452,469,470,490 and 491. The same base packages without either version of  stratagems do not show any tobex complaints in the console. So it appears that stratagems finally is an EE mod :-)

That it already happens with stratagems v31 would indicate that it really makes no sense to bother DavidW with it, I think that is what Jarno Mikkola implied ? Or should I open a thread over at G3 just in case ? Any recommendations ?

What exactly constitutes an EE mod ? There are several mods in the list which support EE in principle or have components for EE, e.g. klatu or stratagems. Probably the only way to figure that out if I wanted to would be to test what happens if mods are left out ?

Well, there are several mods that are modernized and as a result of a bad moding practice, have updated their content to include EE-only features in a non exclusive manner that by doing

I tested a setup which would be the bare minimum of mods to me, see attached weidu.log. Even with these I get the tobex console messages. So it appears there is nothing I can do about it. At some point I should actually start playing the game, maybe I discover what it is that does not work then :-)

Thank you so far !

Hello ... and everyone !

nOpcode ... 400 ...

This has to do with an EE mod being installed on a non-EE game. As you can see from these links, the non-EE stops at 351, while the EE one is 471.

It won't crash the game... but what ever the actions that were supposed to happen, won't.

I even have a 490 which is not in the EE table :-)

What exactly constitutes an EE mod ? There are several mods in the list which support EE in principle or have components for EE, e.g. klatu or stratagems. Probably the only way to figure that out if I wanted to would be to test what happens if mods are left out ?

In any case thank you !

Hello Tom and everyone !

It is very nice to hear that you are happy with the result now. I have one question you (or someone else) might be able to answer. I made a test with a bigger mod selection. Installation went fine (weidu.log attached). I started the game the in windowed mode because this particular windows 10 version and I have different ideas how screen resolution should work. Anyway, I can see the tobex console when starting the game. I see a lot of

DETOUR_CDerivedStats::DETOUR_GetStat(): nOpcode out of bounds (490; expected 1-400)

Instead of 490 varous other numbers above 400, e.g. 491, 470, appear. There was a second message about nOpcode out of bounds which scrolled away.

Do you see the same messages ? Should I worry about ? Googling the error turns up some very old posts only, I am not sure they still apply.

Best Regards

Hello Leonardo Watson !

I can confirm that the workaround in the BWP Installpack 18.2.5 is effective, I managed to install the minimally modded game which failed previously. I also tested the beginning of the tutorials in BG1 and BG2 and there were no problems. I attach the weidu.log. Thank you very much for your help !

I have an additional question: I would like to eventually switch the weapon proficiency system to IWD with the Tweaks Anthology component 2163. You explicitely warn against components 2161 and 2162 in connection with Jarls BGT Tweak Pack. I will not install  Jarls BGT Tweak Pack. Are you aware of any other potential incompatibilites ?

Best Regards

ckoe said:

Did you notice ? stratagems v32 works (officially) only with a 64 bit weidu, and 2*32=64 ... OK, probably unrelated :-)

Well, it was an attempt to make a joke. Did not work.

Leonardo Watson doesn't have a 64-bit OS capable computer, which is odd as those are the main consumer sold PCs the last ~12 years, or so(the first consumer-grade PC goes all the way to 2005).

Or does not have a 64 bit OS ? Anyway, what I actually meant to offer was to run a test with a 32 bit weidu if that would make sense on my computer just to see if the bitness of the weidu would makes any difference. I might try anyway when testing the workaround, although the comment by new(Mike) is probably an indicator that it is hopeless to try 32bit weidu.

EDIT: Changed the first quote, originally had the wrong part of the post there. I did not figure fully out how the quoting system of this forum software works ...

We were talking about a different things, you are talking about the SPACE character that's in STATS.00002.IDS -file... I am talking about the additional lines at the end of the file, as each iteration has different amout of them. ...000.2da to ...001.2da has a difference of 20 empty lines. The ...001.2da to ..002.2da has a difference of -18, and at least one number was moved to a later line. But it still has 2 empty lines at the end. ... and the last .ids has 5 empty lines( so additional +3).

Ah, I see. Sorry then !

I just added a few lines to the Installpack to fix the stats.ids issue. I don't know which mod causes the issue because it seems to be a mod that I don't have in my installation. All the other empty lines will be trimmed correctly by the Installpack no matter whether they are leading or tracking ones.

I will try it in the next few days and report back. Thank you !

If you need any further tests to find out if it is actually stratagems v32 (which is still my hypothesis) or related to a particular weidu version or bitness, as suggested by Jarno Mikkola if I understood his argument correctly,  please let me know. I assume it would be a good thing to understand and to fix the underlying issue (if any and if possible) instead of working around it.

Did you notice ? stratagems v32 works (officially) only with a 64 bit weidu, and 2*32=64 ... OK, probably unrelated :-) Or does it eventually still work with a 32 bit weidu for testing ?

Hmm, is a little more complicated...
As even the tobex adds a space to the end of the file... aka it might not be SCS that does this. But the weidu -function that this originates from.

Sorry, but which file version are  you refering to ? I am asking because:

There are never (white)spaces (I can find) in the empty lines at the end of the file in any of the revisions. All I can see with a hex editor are "0D0A" (aka CR LF) sequences in these lines at the end of the file. Always. This is also true for the first empty line in STATS.00000.IDS and STATS.00001.IDS, it just is "0D0A".  Are you sure that  tobex does add a (white)space to any empty line on top of a "0D0A" ?

However, looking at the bytes at the start of the file I see:

1. STATS.00000.IDS: 0D 0A 31
2. STATS.00001.IDS: 0D 0A 31
3. STATS.00002.IDS: 20 0D 0A  <- This appears to be the offending whitespace (#20) CR (#0D) LF (#0A) sequence

I am not sure to which changelog entry STATS.00002.IDS belongs. And you are right, the "200D0A" sequence could be from some particular weidu call nothing else uses.

... if necessary manually, saving stats.IDS for later inspection during every pause if you think that would be helpful ? 

No need to do that... you can just --change-log the file and get all the files from their backups printed out to the folder that it makes.

OK, I tried that. There was a "irregularity" in the sense that the change-log.bat after I added the filename did not work. I the just executed the Weidu command from the .bat by hand on the command line. This gives the files attached which were in the change-log folder. The STATS.00002.IDS appears to be the first one with the space ?

Let me know if I can do anything else.

EDIT: To clarify, I used a plain "stats.IDS" as the filename argument to the weidu command. So not override\stats.IDS or something.

Yet, the only difference in the two WeiDU.logs new(Mike) posted, i.e. in post number 122 and number 115, appears to be stratagems v32.7. My stats.IDS.bak with v32.6 has this white space also, see e.g. post 87, I did not explicitely check the most recent one. While I have been a bit undisciplined between tests, i.e. changing mod and component selections, I had no problems with v31.

If you would like to have something tested I can offer to try it. Maybe inserting a pause after each mod/ installation step (or even mod component) and, if necessary manually, saving stats.IDS for later inspection during every pause if you think that would be helpful ?  Or, that would be really easy now I think about it, I could grab the previous v32 minor versions including release candidates from github and give each of them a try ?

You might want to consider that I have been seeing this exact problem with stratagems v32.6. This is nothing which happened in the version bump to v32.7.

Please check the stats.IDS.bak which I uploaded last week or so as proof of that.

EDIT: I attach now again the weidu.log and the (zipped) stats.IDS.bak from the near minimal install which was the most recent one for easier reference. The stats.IDS in override is 0 Byte long. This is with v32.6, if it is a stratagems change then it is reasonable to assume that the change in the changes tagged v32.6 or before

Hello Jarno Mikkola,

to answer your question to new(Mike) from my experience, it does _Not_ start generalized biffing because it fails before finishing close2.bat, "crashing" to a shell prompt. So, there is no debug log for generalized biffing I could post I believe. You might want to take a look at the screenshots I posted which show what happened in close2.bat in my install.

Please take also a moment to review the weidu.logs I posted and the posts themselves. I would like to stress again that (I admit only mostly) identical mod and component selections appear to work, i.e. generalized biffing is executed and the install finishes with an appropriate success message, without stratagems v32 or including v31 instead.

This (yes that it is the same problem is a jump to a conclusion) problem when v32.6 or v32.7 is included was reported independently by two persons.

I admit that it might be the OS, individual.bat or something else. But v32 makes a problem visible. And yes, if the Big World Project does not finish regularly I consider it broken at the moment. I believe the word "probably" in the second part of the sentence you quote makes it clear that the second part about the future is an extrapolation.

If you have any concrete suggestions to either diagnose this further or to work around I will gladly help with that.

Thank you very much !

Regarding your problems with the close2.bat I noticed both of you have installed Stratagems v32.7 or v32.7. Maybe it conflicts in any sort with the batch routine even if I cannot imagine why.

Yes, exactly this is the point :-)

Stratagems v32.6 and v32.7 appear to break the Big World Project at the moment. That is why I did this bare bones mod install in my latest post including the weidu.log: to proof as far as possible that it is really the addition of stratagems v32.6 either itself or with one of the required minimal components (discounting BGT Tweaks) which leads to this breakage.

In previous post I gave information about an attempt with stratagems v31 with the same mods and largely the same individual.bat. With stratagems v31 everything works.

You mentioned being unable to test it before. If you have anything I can test for you let me know.

But at the moment the conclusion is that the Big World Project stopped working with the currently developed version of stratagems and will probably not work with any versions developed from it in the future if this is not fixed.

@tomkaz: DivideBy0 is most probably completely right, and please only invest time into my suggestion if and only if nothing else comes to mind any more.

As  I wrote, I would not have brought it up if I did not find what tomkaz reports pretty much confusing.

I am still not sure what the problem with bg1ub was, but corruption during extraction  is a possibility ?

When we were still buying machines without ECC (20 years ago) or on our non-workstation machines and laptops without ECC I have seen some pretty weird stuff happening with faulty RAM, which happens once every few years, i.e. very rarely but it happens.  It might be a bit which only flips once in a while based on some factor, or it might be that the OS runs the binaries at some other position in memory so that the faulty region is not accessed by chance.

And both tests are non-destructive and save once you have the necessary tools.

Hello everyone !

In some sense it is good to see that Mike (?) has the same problem. So it is not my computer or some other external factor.

I tried now a very minimal install (BGT Tweak could be removed) with and without the fix Leonardo Watson suggested I should try. Unfortunately it still fails in either case, assuming I made the right changes to close2.bat. I attache the WeiDU.log and the close2.bat of that failure for reference. This looks like it will take some time to figure out and fix :-) Again, I can offer to do testing.

@tomkaz: What you report from your computer problems is very confusing. I assume you did some diagnostics when it first broke down. However, apparently there is still something wrong. While there might be a permission problem I wonder if the hard disk (S.M.A.R.T) and the RAM (memtest86 comes to mind) have been checked ? Especially bad RAM could lead to silent and seemingly random data corruption and other issues. Both can be checked with the right tools basically for free but require some more involved steps, eventually some type of SMART test is already integrated in the computers firmware diagnostics if it has any ?

Sorry that I cannot offer more help than this stab in the dark. If it was not so very confusing I would not bring this up at all and eventually send you on a wild goose chase.

Best Regards

Hello Leonardo Watson !

Yes, I will try that on Monday. And by all means take your time ! I post this often  to not lose the overview about what I am doing myselves. You are spending your free time helping me and I am very grateful for that !

Hello Leonardo Watson,

with stratagems v31 (and iwdification thrown in on top) the install works, no problem at all.

I will try a version with stratagems 32.6 and only ToBEx, Baldurs Gate Trilogy, and BG2 Fixpack on Monday. I believe those are the absolute minimum mods. This should really proof that it is stratagems 32.6 and not some other mod which messes stats.ids up and then the error only shows up with stratagems on top.

While tomkaz is reading individual.bat: There is a section

findstr /I /M "SANDRAHRTF.TP2~.*#.*#0./" %ELOG%>nul
if errorlevel 1 (
        Call ".\%BWIPI%\BGTTweak.bat
)

which appears to trigger for me and then installs components 1800 to 1808 from BGTTweak. At least that is what I suspect, because commenting it out removes those from the install (I did that earlier, but when changing stratagems v31 components I grabbed an older version) and nowhere else in the individual.bat is BGTTweak called with these components. The guide says not to install them. In fact, after installation the Installpack informs that 1800 and 1806 were not installed because the actual NPC mods are included which is nice.

I attach the weidu.log of the successfull install with stratagems v31. I will have to fiddle with the component selection a bit more.

Best Regards

Hello Leonardo Watson,

so the install attempt with the slowed down trim_ids.bat, which I thought I would not be able to do today, finished with a failure at line 1437 as before. The stats.ids is empty and stats.ids.bak is the same as ever. On the other hand I am not sure where you wanted to insert a pause. So, if I can test anything for you please let me know.

I will probably revert to stratagems 31 then. Too bad, I really liked the after-install customization which I could test briefly by starting one of the earlier failed installs.

Best Regards

Hello Leonardo Watson,

after testing I can confirm that the problem is stratagems 32.6 . I wanted to have it because of its additional customizability after install.

The install fails with just adding that, the tweaks component is set to 2160 and Jarls BGT Tweaks was not included. The weidu.log together with the stats.ids is attached. The install would work without stratagems.

I will add timeout statements in trim_IDS.bat next to see whether that changes everything. I plan to add these to :trimLeft, :trimRight and :ProcessLine just before the final "goto :eof" of each of these blocks.

If I should test something different please let me know.

Best Regards

Hello tomkaz,

what is your problem with v27 ? I have it in my install and as far as I can see it does not make a problem for me.

It would be possible to roll back to v27rc1 by checking out the right commit from the git repository, so eventually it is no big deal for the author to help you ?

Best Regards

Hello Leonardo Watson !

Thank you for looking into this issue. But please wait a moment with any work. There are a few things which I discovered in the meantime.

The current state is that I have nearly all mods included and did not have that problem again. The mods missing are essentially stratagems 32.6 (which I assume you cannot test) and Jarls BGT Tweak pack (which I will probably remove in the end, but for testing it goes back in). Also I had an old copy of BG1 UB, version 14 instead of 16.

Looking at your install guide I see that there is a potential problem with components 2160 to 2164 from Tweaks Anthology with Jarls BGT Tweak pack. My original individual.bat had component 2163 instead of 2160. The warning is about 2161 and 2162, but in hindsight this does not imply 2163 is save to use. If you check the stats.ids I posted there are seveal lines mapping the same number to different weapon proficiencies. I am not sure if that is ok, but I do not like it especially missile and melee weapons mapping to the same number. So I would like to make sure, if possible, that it is neither a stratagems problem nor a problem with the combination of mods and option 2163.

The install is on an external USB3 rotating hard disk (otherwise I would destroy the internal SSD pretty fast I fear). Assuming I can reproduce the problem in the new setup, everything else would be a bit embarassing, I can add a timeout command in the batch file and test.

This will take a few more days. I will report back with the results ! Thank you very much again.

Edit: changed the broken formatting.

Edit2: According to the weidu.logs BG1 UB was at version 16 all the time. Will have to check the logs very carefully again.

Hello !

I started to build up the installation basically one by one mod according to the Big World Project guide now. I run a full installation after adding only one or a few mods. With minimal mods, i.e. only bgt and section 1 of the guide, the installation is successfull. I expect to be able to pin down why it failed previously with this procedure, although it will take some time.

Best Regards

Hello everybody,

I have the results with the echo statements now.

The first screenshot (trim_error.png) shows the command window with the error and the additional output from the echo statements on the left side and most of the modified trim_IDS.bat (so that one can see the location of the echos) on the right side. So the script fails directly after apparently moving "override\stats.IDS" to "override\stats.IDS.bak" when trying to execute :trimright (!, trimleft is finished at that point).

The second screenshot (stats.png) shows the "override\stats.IDS.bak" file, the "override\stats.IDS" has zero bytes size (windows hides extensions unfortunately). The .bak appears to have an empty first line, apart from that it looks fine to me. Would it be ok to upload the actual file ?

At this point I really need some input from people understanding the file formats and eventually someone can shed a light on the question why there is an empty first line (which I assume is the problem). Also, how to proceed here ? I could try to introduce a pause statement in close2.bat, manipulate the files and the unpause ?

Best Regards

EDIT: Looking more closely the first line is not empty but contains a single white space. I simply attach the file now in zipped form (because the forum software does not let me upload it without) together with the current weidu.log.