[Bug 17413] New: riched32/editor tests fail with +heap enabled

http://bugs.winehq.org/show_bug.cgi?id=17413

Summary: riched32/editor tests fail with +heap enabled
           Product: Wine
           Version: 1.1.15
          Platform: PC
               URL: http://test.winehq.org/data/6a1537c4b5037be5253c33021fda
                    3a02430533f7/wine_xp_ae-ub-810-heap/riched32:editor.html
        OS/Version: Linux
            Status: NEW
          Keywords: download, source, testcase
          Severity: enhancement
          Priority: P2
         Component: richedit
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Looks like some massive heap corruption:
austin@austin-desktop:~/wine-git/dlls/riched32/tests$ wineserver -k ; make
testclean ; rm -rf ~/.wine ; WINEDEBUG=+heap make editor.ok 2> /dev/null | grep
"Test failed:" | wc -l
rm -f editor.ok
186

editor.c:101: Test failed: WM_SETTEXT round trip: strcmp = -1
editor.c:101: Test failed: EM_GETLINECOUNT returned 12, expected 1
...
editor.c:319: Test failed: streamed text length is 13, expecting 12
editor.c:320: Test failed: streamed text different, got TestSomeText
...
editor.c:395: Test failed: 0: 1!=9
editor.c:406: Test failed: 0: expected_bytes_written=10
editor.c:406: Test failed: 1: expected_bytes_written=2
editor.c:395: Test failed: 2: 2!=3
...
editor.c:457: Test failed: Incorrect line count of 45
editor.c:461: Test failed: Length of line at offset 0 is 1, expected 9
editor.c:461: Test failed: Length of line at offset 5 is 1, expected 9

[Bug 17413] riched32/editor tests fail with +heap enabled

http://bugs.winehq.org/show_bug.cgi?id=17413

Dylan Smith  changed:

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

--- Comment #1 from Dylan Smith   2009-02-24 11:52:09 ---
Seems like it is fixed by commit b9b9835f9368c83e15451e04c8e0781012ac4a2f which
I submitted to fix Bug 17502.

The patch initializes an uninitialized value.  Turning on the heap channel for
debugging initializes allocated bytes that aren't initialized, so it caused the
tests to fail.

[Bug 17413] riched32/editor tests fail with +heap enabled

http://bugs.winehq.org/show_bug.cgi?id=17413

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Austin English   2009-02-24 13:52:19 ---
(In reply to comment #1)
> Seems like it is fixed by commit b9b9835f9368c83e15451e04c8e0781012ac4a2f which
> I submitted to fix Bug 17502.
>
> The patch initializes an uninitialized value.  Turning on the heap channel for
> debugging initializes allocated bytes that aren't initialized, so it caused the
> tests to fail.

Yep:

http://test.winehq.org/data/86f79e3a558ca1fb5d596e9f62403ca89178d3cf/wine_xp...

Now we just need to fix those other 9 failures ;-). Some of them
(mshtml/msi/winmm) are crashing, and may be a bit harder. Crypt32's got a few
that don't like debugging traces...

[Bug 17413] riched32/editor tests fail with +heap enabled

http://bugs.winehq.org/show_bug.cgi?id=17413

Alexandre Julliard  changed:

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

--- Comment #3 from Alexandre Julliard   2009-02-27 16:31:44 ---
Closing bugs fixed in 1.1.16.