[texworks] Texworks quit working (on windows)

[K. Frank]

Hello Stefan!

On Mon, Oct 3, 2011 at 3:34 PM, Stefan Löffler <st.loeffler at gmail.com> wrote:
> Hi,
>
> this is getting stranger and stranger.

Indeed, it does seem strange to me.

> To follow up on the dll issue:
> There once was a nice program called "Dependency Walker" at
> http://www.dependencywalker.com/ but this links seems dead now.
> It seems that http://support.microsoft.com/kb/256872/en refers to an
> older version, though, so you might want to give it a try, load
> TeXworks.exe in it (the real one) and see if any dlls look out of place
> (i.e., loaded from somewhere else than C:\windows or the directory Tw
> itself resides in).

Okay, I finally bit the bullet and got Dependency Walker.  I downloaded
both the older version from your microsoft link, and the version for which
Philip found the Way-Back link.

The basic story is as follows:

The "real" texworks.exe (both the "stand-alone" and the one that came
bundled with TeX Live) do show a possible missing dependency:

The error / warning log window shows:

   Warning: At least one delay-load dependency module was not found.
   Warning: At least one module has an unresolved import due to a
   missing export function in a delay-load dependent module.

and the Module List window shows:

   GPSVC.DLL
      Error opening file.  The system cannot find the file specified (2).
   IESHIMS.DLL
      Error opening file.  The system cannot find the file specified (2).

These (for whatever reasons) smell like red herrings to me.  (But what
do I know...)

I then ran TeXworks.exe inside Dependency Walker (Profile / Start Profiling...).
It generated a whole slew of messages, but to my eye, they all looked okay.

It started with:

   Starting profile on 10/3/2011 at 4:44:37 PM
   ...
   Started "TEXWORKS.EXE" (process 0x1164) at address 0x00400000.
Successfully hooked module.
   ...
   Entrypoint reached. All implicit modules have been loaded.
   ...
   Exited "TEXWORKS.EXE" (process 0x1164) with code 1 (0x1).

There was no mention of GPSVC.DLL or IESHIMS.DLL in the profiling
output.  The profiling did pick up one other missing dependency:

   LoadLibraryW("C:\Windows\system32\wintab32.dll") returned NULL.
Error: The specified module could not be found (126).

but again, I think this is a false positive, so to speak.

I will note that neither the static nor dynamic dependency analysis
showed any kind of dependency on a python dll.

> On a different front, there are a few things I'd like to rule out in
> connection with startup. Therefore, I've stripped the Tw program from
> all the graphical user interface and retained only the startup code. The
> resulting program is a console app, and you can download it from the
> following link (for some time, anyway):
> http://www.filehosting.org/file/details/272564/startup-test.zip

filehosting.org wanted an email address that I didn't feel like providing.
Perhaps you could just attach the test program to a posting here, or
email it to me directly?

> Please unzip it and run it. If all goes well, it should print the
> numbers 1-9.
>
> On 2011-10-03 18:41, K. Frank wrote:
>> To follow up on the python angle:  Do the other parts of TeX Live rely on
>> python?  Or is just TeXworks python dependent?
>
> Not that I know of. At least not in the way Tw does. But as Tw supplies
> its own python dll, changes to other pythons on your system shouldn't be
> a problem (at first glance, anyway). Besides, dll errors normally cause
> some error messages, don't they?

I am assuming that texworks.exe has some sort of python library linked
in statically.  Why?  Because several posters have indicated that
texworks.exe does use python and my ...\TeXworks-w32-0.5-r932\Lib
directory (and the analogous ...\texworks'lib directory in the TeX Live
installation) contains a bunch (eighteen) .pyc files.  So it looks like
TeXworks uses python.  But neither the TeXworks nor the TeX Live
installation comes with any python dll's, and Dependency Walker
shows no dependencies on any python dll's.

So would anyone know if python is, in fact, linked in statically?

> Anyway, we'll see how the startup test goes...

If you could kindly email me the startup test or post it here, I would
be happy to give it a try.

> HTH
> Stefan

Thanks very much for your help with this.

This is becoming quite a mystery to me.  I'm almost more interested
in solving it just to resolve the mystery than I am in using TeX!

Best.


K. Frank