[Bug 51531] New: DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

Bug ID: 51531
           Summary: DrMingw problem
           Product: Wine
           Version: 6.13
          Hardware: x86-64
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

Although I installed DrMingw, wine doesn't call him if the exception occurs.

d: && cd \drmingw-0.9.3-win64\bin && drmingw -i -v
Dr.Mingw has been installed as default application debugger

wine64 version=wine-6.13-173-g04d52eb83fa

/mnt/WD-Elements-25A1/src/wine-git/install64/system.reg
...
[Software\Microsoft\Windows NT\CurrentVersion\AeDebug] 1623820558
#time=1d7626eb12154b6
"Auto"="1"
"Debugger"=""d:\drmingw-0.9.3-win64\bin\drmingw.exe" -p %ld -e %ld -v"
...
[Software\Wow6432Node\Microsoft\Windows NT\CurrentVersion\AeDebug]
1623820558
#time=1d7626eb121692e
"Auto"="1"
"Debugger"=""d:\drmingw-0.9.3-win64\bin\drmingw.exe" -p %ld -e %ld -v"
...

/asoft-src/ispp-64/linux/cross-compile/llvm-mingw/ispp64-llvm-mingw-start.sh
0110:fixme:process:CmdBatNotification 1
0110:fixme:process:CmdBatNotification 0
[root@vlada-kuci tmp 00:41:04]$ wine: Unhandled division by zero at address
000000014000F806 (thread 0120), starting debugger...
0120:err:seh:start_debugger Couldn't start debugger L"" (2)
Read the Wine Developers Guide on how to set up winedbg or another debugger
wine: Unhandled page fault on write access to 0000000000000000 at address
00007F96D844E9BB (thread 0070), starting debugger...
0070:err:seh:start_debugger Couldn't start debugger L"" (2)
Read the Wine Developers Guide on how to set up winedbg or another debugger

Vladimir Koković, DP senior(70), Serbia, Belgrade, 28.July 2021

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

Bernhard Übelacker  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #1 from Bernhard Übelacker  ---
Hello, I tried to reproduce this issue with a 64-wine build
from git at wine-6.14.

I took drmingw from here:
https://github.com/jrfonseca/drmingw/releases/download/0.9.3/drmingw-0.9.3-w...

And in a clean wineprefix unpacked it and installed it like you did,
but just in the drive_c.
This gave me an interactive messagebox with the same message you wrote
for the installation.

Then, because I don't know the content of your script
ispp64-llvm-mingw-start.sh, , I created a small test case:
  echo "int main() { return 0/0; }" > divzero.c
  x86_64-w64-mingw32-gcc -g -O0 -Wno-div-by-zero divzero.c -o divzero.exe

And finally tested it by:
  wine divzero.exe

This brought up successfully the interactive window of drmingw, showing
the exception information, source and backtrace with source file and lines.

I compared also my system.reg file and it showed the same entries
like yours (except the drive letter).

The error message from your crash indicates, that the crashing process
was just able to retrieve an empty debugger command line L"".
So maybe you can reduce your test case, because currently the
ispp64-llvm-mingw-start.sh script hides the most part of it.

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

--- Comment #2 from Vladimir Koković  ---
File
/asoft-src/ispp-64/linux/cross-compile/llvm-mingw/ispp64-llvm-mingw-start.sh
contents:
---
#!/bin/bash

set -e

export WINEPREFIX="/mnt/WD-Elements-25A1/ispp-sql-mingw-64/windows-10"
export FREETYPE_PROPERTIES="truetype:interpreter-version=35"

filebat="vk-ispp-llvm-mingw.bat"

cat << EOF > "/tmp/$filebat"
rem
Z:\asoft-src\ispp-64\linux\cross-compile\llvm-mingw/ispp64-llvm-mingw-start.bat
PATH=.;%PATH%;z:\asoft-src\ispp-64\linux\cross-compile\vk-llvm-mingw\x86_64-w64-mingw32\bin;d:\drmingw-0.9.2-win64\bin;Z:\asoft-src\ispp-64\linux\cross-compile\llvm-mingw/pdcurses-cross-PDCursesMod-4.2.0-wincon\dist\Debug\llvm-mingw-Linux;Z:\asoft-src\ispp-64\linux\cross-compile\llvm-mingw\libodbc-llvm-mingw\dist\Debug\llvm-mingw-Linux
set LSTDIR=d:\ispp-64\ispp-64\liste
set XENVIRONMENT=d:\ispp-64\ispp-64\XCurses.def
z:
cd
Z:\asoft-src\ispp-64\linux\cross-compile\llvm-mingw\ispp-llvm-mingw\dist\Debug\llvm-mingw-Linux
start /MAX ispp-llvm-mingw.exe
"-xDRIVER={PostgreSQL};DATABASE=ispp-pprostor-beograd;SERVER=localhost;PORT=5432;UID=vlada;PWD=vladakk;MaxVarcharSize=65536;MaxLongVarcharSize=65536;ByteaAsLongVarBinary=1"
EOF

wineconsole z:\tmp\$filebat

exit 0
---

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

--- Comment #3 from Bernhard Übelacker  ---
Hello, you added the path to your system.reg in the first message.
> From that I assume you installed drmingw with such an environment?

WINEPREFIX=/mnt/WD-Elements-25A1/src/wine-git/install64

But ispp64-llvm-mingw-start.sh changes that to:
  WINEPREFIX="/mnt/WD-Elements-25A1/ispp-sql-mingw-64/windows-10"

Therefore, for the wine processes started afterwards, I guess the AeDebug
setting from the system.reg below this wineprefix matters.
Is DrMingw installed to this wineprefix too?

Another thing could be, what filesystem is the mount /mnt/WD-Elements-25A1?

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

--- Comment #4 from Vladimir Koković  ---
(In reply to Bernhard Übelacker from comment #3)
> Hello, you added the path to your system.reg in the first message.
> From that I assume you installed drmingw with such an environment?
>   WINEPREFIX=/mnt/WD-Elements-25A1/src/wine-git/install64
>
> But ispp64-llvm-mingw-start.sh changes that to:
>   WINEPREFIX="/mnt/WD-Elements-25A1/ispp-sql-mingw-64/windows-10"
>
> Therefore, for the wine processes started afterwards, I guess the AeDebug
> setting from the system.reg below this wineprefix matters.
> Is DrMingw installed to this wineprefix too?
>
> Another thing could be, what filesystem is the mount /mnt/WD-Elements-25A1?

It's my fault, I apologize to everyone,
but if I put export wineprefix=/mnt/wd-elements-25a1/src/wine-git/install64,
Wine don't do start program !

blkid
/dev/sdd1: UUID="423b17a7-7cc7-48eb-bbbb-3646b13d57b1" BLOCK_SIZE="4096"
TYPE="ext4" PARTLABEL="Elements"
PARTUUID="10f9b972-3188-4b46-a2d4-a6d647f313f4"

mount
/dev/sdd1 on /mnt/WD-Elements-25A1 type ext4 (rw,noatime)

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

--- Comment #5 from Bernhard Übelacker  ---
The mounts filesystem is ext4, so I think that is no problem.
I was wondering if this might be some network filesystem ...

Depending on what is installed in the install64 prefix, it
might be faster to rename and reinstall it?

If I understand you right, from your side this bug can be closed?

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

--- Comment #6 from Vladimir Koković  ---
(In reply to Bernhard Übelacker from comment #5)
> The mounts filesystem is ext4, so I think that is no problem.
> I was wondering if this might be some network filesystem ...
>
> Depending on what is installed in the install64 prefix, it
> might be faster to rename and reinstall it?
>
> If I understand you right, from your side this bug can be closed?

Yes.

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

Bernhard Übelacker  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |INVALID

--- Comment #7 from Bernhard Übelacker  ---
Maybe side effect of mixing wineprefixes.
Closing as confirmed by bug creator.

[Bug 51531] DrMingw problem

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=51531

Gijs Vermeulen  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #8 from Gijs Vermeulen  ---
Closing.