[Bug 38220] New: Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

Bug ID: 38220
           Summary: Chemcraft crashes with wine-1.7.38
           Product: Wine
           Version: unspecified
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

Created attachment 51020
  --> https://bugs.winehq.org/attachment.cgi?id=51020
Chemcraft crash backtrace

After upgrading wine from 1.7.33 to 1.7.38 Chemcraft_b405
(http://chemcraftprog.com/files/Chemcraft_b405_win.exe) doesn't start anymore.

I try bisecting following http://wiki.winehq.org/RegressionTesting.
Git says that f16cf541068831074a6891688d947f0ce32abdd9 comit is the first bad
commit. I try ot revert this commit but it's failed because of some commits
with this code was made.

Also I reproduce this crash with the current git wine
(ed92458754ce21b4d4719168515866fc1bb340ad).

I attach backtrace which was made by winedebugger.

P.S. Old versions of Chemcraft (b382 and below) works with wine-1.7.38.

[Bug 38220] Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

Sebastian Lackner  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |download, regression
                URL|                            |http://chemcraftprog.com/fi
                   |                            |les/Chemcraft_b405_win.exe
                 CC|                            |[e-mail],
                   |                            |[e-mail]
            Version|unspecified                 |1.7.38
    Regression SHA1|                            |f16cf541068831074a6891688d9
                   |                            |47f0ce32abdd9

--- Comment #1 from Sebastian Lackner  ---
Adding the author of the specific commit.

[Bug 38220] Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #2 from Bruno Jesus  ---
Thanks for the bug report, I can reproduce the error myself. I don't know why
the commit caused the issue since the program uses GetTempPathA instead of
GetTempPathW and the buffer passed from A to W is always good. Maybe related to
the EXECryptor used by the application, I'll have to test when I have more
time.

[Bug 38220] Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|-unknown                    |kernel32

[Bug 38220] Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

--- Comment #3 from Bruno Jesus  ---
This simple change makes it work for me, changing the variable with its value.
Any hints?

diff --git a/dlls/kernel32/path.c b/dlls/kernel32/path.c
index 7fe229f..8f590da 100644
--- a/dlls/kernel32/path.c
+++ b/dlls/kernel32/path.c
@@ -623,7 +623,7 @@ DWORD WINAPI GetTempPathW( DWORD count, LPWSTR path )
     {
         lstrcpynW(path, tmp_path, count);
         /* the remaining buffer must be zeroed */
-        memset(path + ret, 0, (count - ret) * sizeof(WCHAR));
+        memset(path + ret, 0, (260 - ret) * sizeof(WCHAR));
         ret--; /* return length without 0 */
     }
     else if (count)

[Bug 38220] Chemcraft crashes with wine-1.7.38

https://bugs.winehq.org/show_bug.cgi?id=38220

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|[e-mail]        |[e-mail]

--- Comment #4 from Bruno Jesus  ---
Ok, I just found the issue. My first assumption was wrong, it doesn't call
GetTempPathA, just GetTempPathW.

The problem is that the count value passed is 65534 but the buffer does not
seem to really have that size so the memset to clear the memory makes the
application crash. I'll add more tests for such invalid sizes and check what
native does.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Chemcraft crashes with      |Chemcraft crashes
                   |wine-1.7.38                 |

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

Sergey  changed:

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

--- Comment #5 from Sergey  ---
I also can confirm this bug and I made regression test.
Something wrong with my attempt, it said
~~~~
The first bad commit could be any of:
2f804307bda0f62ac23c9a3387e185150868fede
7f618695c624f532300ecff8735f94fec3f3a86c
8ed9c2a11bf8a61909ef0940d0dee3611a6c003e
db80aba5ad526b8691dace1c4c20b5cbc85389a4
1fcc75299f3e640ec60ef3caade9d14fa2e1299a
We cannot bisect more!

~~~~

I am looking around and have one doubt
~~~~
            if (!gl_info->supported[ARB_DRAW_INSTANCED] &&
!gl_info->supported[ARB_INSTANCED_ARRAYS])
            {
                FIXME("Instanced drawing not supported.\n");
            }
            else
            {

~~~~
should it be || instead of && ?
"else" will be executed if (1 && 0) that leads to crash.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

--- Comment #6 from Bruno Jesus  ---
Hi, Sergey. I believe you posted to the wrong bug. This bug is well-known and
in process of getting fixed already.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

--- Comment #7 from Sergey  ---
(In reply to Bruno Jesus from comment #6)
> Hi, Sergey. I believe you posted to the wrong bug. This bug is well-known
> and in process of getting fixed already.

I just think it is related because I found it here.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |24fc876e616db6f2249811c21c2
                   |                            |c399c57c7df2f
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #8 from Bruno Jesus  ---
Fixed by
http://source.winehq.org/git/wine.git/?a=commit;h=24fc876e616db6f2249811c21c...

Thanks for the bug report.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

Alexandre Julliard  changed:

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

--- Comment #9 from Alexandre Julliard  ---
Closing bugs fixed in 1.7.39.

[Bug 38220] Chemcraft crashes

https://bugs.winehq.org/show_bug.cgi?id=38220

--- Comment #10 from Sebastian Lackner  ---
The application DipTrace was also affected by this, leaving this note here so
that its also documented on the upstream bugtracker:

https://bugs.wine-staging.com/show_bug.cgi?id=173