[Bug 14981] New: Catia's file reading is unreliable

[wine-bugs at winehq.org]

http://bugs.winehq.org/show_bug.cgi?id=14981

           Summary: Catia's file reading is unreliable
           Product: Wine
           Version: CVS/GIT
          Platform: PC-x86-64
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: kernel32
        AssignedTo: wine-bugs at winehq.org
        ReportedBy: mah at jump-ing.de


Created an attachment (id=15593)
 --> (http://bugs.winehq.org/attachment.cgi?id=15593)
log/backtrace of Catia's internal facilities

In my quest to get DassaultSystemes' Catia V5 to an useable state, I currently
face unreliable file operations. While I can open and view a single CAD model
(CATPart) once or even twice, additional attempts as well as an attempt to open
a multi-model scene (CATProject) fails and quits Catia with a Visual C++
Runtime Exception.

The interesting thing is, for startup, Catia loads some thousand configuration
files (small text files) and obviously, none of these operations fails. But
opening an additional two or three files after startup, failure is almost sure.

To debug the thing I've set up Catia similar to how it's decripbed in the AppDB
http://appdb.winehq.org/objectManager.php?sClass=version&iId=12939 . Wine is
compiled unoptimised with CFLAGS="-g". To launch the app my current favourite
is:

  env WINEPREFIX="/home/mah/.wine" WINEDEBUG=+seh,+tid winedbg
"C:\Programme\Dassault Systemes\B16\intel_a\code\bin\catstart.exe" -run
"CNEXT.exe" -env CATIA.V5R16.B16 -direnv "C:\windows\profiles\All
Users\Anwendungsdaten\DassaultSystemes\CATEnv"

This eventually stops in the debugger, giving a backtrace like this (it's
short, so inline):

=>1 0xf7fc642e (0x0033f270)
  2 0x00000800 (0x0033f3b0)
  3 0x7b84ab74 ReadFile+0x191(hFile=0x4c, buffer=0x33f448, bytesToRead=2048,
bytesRead=0x33fcc0, overlapped=(nil)) [/home/mah/wine/dlls/kernel32/file.c:451]
in kernel32 (0x0033f420)
  4 0x0040367b in catstart (+0x367b) (0x0033fc58)

file.c, line 451 reads:

    status = NtReadFile(hFile, hEvent, NULL, cvalue, io_status, buffer,
             bytesToRead, poffset, NULL);

To my knowledge, the second line of the backtrace shouldn't read 0x00000800 but
something in NtReadFile.

Catia has it's own backtracing facilities built in, showing up in a
Wine/Windows terminal window. While I can't decipher this, I've attached the
last few hundred lines of such a log.

Last not least I've seen Catia halts in the debugger always at the exact same
place, if the steps done to reproduce the bug are identical.


-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.