195 réponses à ce sujet

[Baaleos]

My Compliments to the chef.

I ran the new toolset mod, and it pointed out a script error where I had a prototype for a function, that mismatched the declaration.

eg- One returned an int, and the actual function was a void.

[Karvon]

For those needing this missing file, here's a link to a copy plus a readme for proper installation.

Karvon

[ShadowM]

Wow nice, I tested it out. Seem the outside PRC compiler a little faster. Great job. You all seem to bring out new stuff all the time. I just trying to keep up with it all. Thanks again

[SkywingvL]

ShadowM wrote...

Wow nice, I tested it out. Seem the outside PRC compiler a little faster. Great job. You all seem to bring out new stuff all the time. I just trying to keep up with it all. Thanks again

I would recommend using this compiler instead of the original PRC compiler, as this compiler includes a number of bugfixes and additional compiler warning/errors that are not present in the (unmaintained) PRC build.

Modifié par SkywingvL, 20 juin 2011 - 02:10 .

[FunkySwerve]

Just installed, very nice. Now, of course, I've grown to prefer command-line...so it's off to Skywing's other post.

Funky

[Baaleos]

Hi guys,
Not sure whats going on, but

My module builds fine in both compilers. No errors.

However, when I build with the new extended toolset compiler, and go into game, a 'Traveller Charm' item I have, which acts as a multi-location recall stone, ceases to work.

But only some of the scripts.

When I build it again, using the default toolset, the scripts and the item, functions again.

Cant post the script contents, as I am at work.
I was just wondering if anyone noticed anything querky like that?

[Baaleos]

Does anyone know

The 16k Limit, is this a limit of the .mod filetype, or limitation imposed by toolset or nwnserver.exe etc


Is it possible to mod this?
It just seems that if we have already improved the compiler, why stop there.

[leo_x]

Baaleos wrote...

Hi guys,
Not sure whats going on, but

My module builds fine in both compilers. No errors.

However, when I build with the new extended toolset compiler, and go into game, a 'Traveller Charm' item I have, which acts as a multi-location recall stone, ceases to work.

But only some of the scripts.

When I build it again, using the default toolset, the scripts and the item, functions again.

Cant post the script contents, as I am at work.
I was just wondering if anyone noticed anything querky like that?

I noticed something similar too.  One of my scripts seemed to get corrupted and have the entire contents of the module in it, so the size of the module would essentially double and some scripts would cease to function.

Compiling scripts with the compiler was not even necessary to recreate it, it seems to happen on module save every time for me.  Almost like it's saving the .BackupMod in side the module proper or something.

Modifié par pope_leo, 20 juin 2011 - 01:17 .

[Thayan]

New release problems, aye?

Sadly, I have to report a similiar situation. The problem was due to my own errors - but the compiler was compiling all my scripts without errors. So then, the change I was testing to a starting conditional script was not occurring. Even when I put in a simple debug line in the SC script saying "this script is firing" and did a full build on the module, the debug info didn't get sent when I tried testing it.

I eventually ended up going back to the default toolset .exe and compiler and found using that, there were a couple scripts that I had recently modified which had logic errors (like a parenthesis in the wrong spot, and so on). Once I fixed those and did a full build using the default toolset compiler, and then the old PrC compiler, my scripts worked as intended and the debug line fired as expected too.

Modifié par Thayan, 20 juin 2011 - 03:39 .

[SkywingvL]

Could you send me a copy of the scripts in question in their problematic state and all of their include dependencies?

Also, could you give building the scripts with the standalone compiler from the ASC distribution a shot? This will help us identify whether there is a problem with the toolset integration part or the compiler core.

Modifié par SkywingvL, 20 juin 2011 - 04:55 .

[Lazarus Magni]

Wow that's messed up, my virus soft ware show viruses in the file from the link virusman posted.

Modifié par Lazarus Magni, 21 juin 2011 - 07:06 .

[virusman]

Since it injects a DLL into nwtoolset process to modify its behavior, it's no wonder some paranoid AVs may give a false positive.

Modifié par virusman, 21 juin 2011 - 01:28 .

[Baaleos]

Virus Software picks it up...
The author is called Virusman.....

HELLO!!!
WAKE UP AND SMELL THE COFFEE.....


Kidding.
Nothing from Virusman has ever set off my virus scanners. (YET).

[FunkySwerve]

AHA! It's as I always suspected. You play a long game, vman, building our trust with years of useful plugins, just so you can release upon an unsuspecting public a virus to turn us all into santa-hat-wearing-will o' wisps - doubtless part of your plans for word domination. Well, played, sir, well played.

Funky

[Baaleos]

You know...
The amount of people that use nwnx, imagine the DDOS attack he could accomplish if he used nwnx to Zombie all the Servers using it. Lol

CIA - Tango Down!!

Virusman - The man behind LulzSec!!

(Any libelous statements above are meant in jest - Disclaimer)

[Lazarus Magni]

Well I am glad to hear it is not malicious, and just to be clear I was not accusing anyone of anything, just reporting what happened. I certainly am one of those who benifits from all the hard work from the nwnx team, and appreciate it very much! I got to say though, I was wondering why I woke up this morning with a santa hat on? Or was it a dunce cap?

Modifié par Lazarus Magni, 21 juin 2011 - 08:26 .

[ShadowM]

Just a thought on people who are having scripts not working after compile, I had similar problems with the other prc compiler that I found was fixed when I took out the optimize script option. I do not know if this one has that in by default. Oh and if there another update version for NWN1 is it at the other link or is there another vault entry for just nwn1 version?

Modifié par ShadowM, 24 juin 2011 - 03:15 .

[SkywingvL]

The Advanced Script Compiler distribution supports NWN1 as well.

If you have a script that isn't working after compiling with the ASC or virusman's plugin, but compiles with the standard BioWare compiler (or silently fails to complete with the ASC or virusman's plugin), can you drop me a copy of the script and all its dependencies?

Unfortunately, if there is a problem here, so far there hasn't been sufficient actionable information to track it down.

[leo_x]

SkywingvL wrote...

The Advanced Script Compiler distribution supports NWN1 as well.

If you have a script that isn't working after compiling with the ASC or virusman's plugin, but compiles with the standard BioWare compiler (or silently fails to complete with the ASC or virusman's plugin), can you drop me a copy of the script and all its dependencies?

Unfortunately, if there is a problem here, so far there hasn't been sufficient actionable information to track it down.

sorry for the delay, for my part, the issue was with my own setup, not your plugin / compiler.

thanks!

[ShadowM]

Here one that does not work when I run it through the toolset compiler. I clearing the actions because I want this conversation to start even in combat. When I compile it with the basic toolset or prc compiler it works, but when I run it through the update compiler toolset it does not fire in game.

object oTarget = GetPCSpeaker();
void main()
{

AssignCommand(oTarget,ClearAllActions(TRUE));
AssignCommand(oTarget,ActionStartConversation(oTarget,"player_guide", TRUE, FALSE));
}

[SkywingvL]

With this script snippet, I observe identical code generation to the BioWare compiler without optimizations when using the standalone ASC build, and identical code generation to the PRC build with optimizations when using the standalone ASC build.

Can you verify that if you build the script locally using the ASC's NWNScriptCompiler.exe that it functions as expected?

[SkywingvL]

After a bit more analysis, I can confirm that this is a pre-existing code generation problem with the optimizer. The same script is incorrectly optimized by the PRC build as well.

I'll get back to you on a fix for the optimizer.

[ShadowM]

Thank you

[SkywingvL]

Please try this compiler build:

http://valera-ext.ny...CompilerDll.ndl

The optimizer bug (longstanding issue as far as I can tell) may cause incorrect code generation in cases where there is a global that is written to only during its initialization expression, and for which the initilization expression involves a function call, action service handler call, or intrinsic call.

I'll release an updated ASC build later to correct the code generation issue in the standalone compiler.

[ShadowM]

Getting No compiler errors, but getting stack underflows.