[Bug 20334] New: SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

Summary: SysReAllocStringLen should return null terminated
                    strings
           Product: Wine
           Version: 1.1.31
          Platform: PC
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: oleaut32
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Created an attachment (id=24050)
 --> (http://bugs.winehq.org/attachment.cgi?id=24050)
wine log +seh,+tid,+relay

The call like SysReAllocStringLen (&buffer, NULL, EXPECTED_LEN) should return
null terminated strings.

I found that problem, then playing with Texas Instruments Code Composer Studio
v3.3 under wine.

Test and logs are attached.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #1 from A. Kochetkov   2009-10-12 08:56:48 ---
Created an attachment (id=24051)
 --> (http://bugs.winehq.org/attachment.cgi?id=24051)
The patch

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #2 from A. Kochetkov   2009-10-12 08:59:24 ---
Created an attachment (id=24052)
 --> (http://bugs.winehq.org/attachment.cgi?id=24052)
The test programm

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #3 from Nikolay Sivov   2009-10-12 09:07:40 ---
Hi, thanks for report.

Could you provide a test diff for Wine test suite (oleaut32/tests/vartype.c)?
It's too complicated now, all you need to check is does it add null terminator
and how it depends on previous string length.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #4 from A. Kochetkov   2009-10-12 11:34:21 ---
Created an attachment (id=24053)
 --> (http://bugs.winehq.org/attachment.cgi?id=24053)
test diff

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #5 from Austin English   2009-10-12 13:05:51 ---
Excellent work. Please submit the test/patch to [e-mail]

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #6 from Nikolay Sivov   2009-10-12 13:09:24 ---
(In reply to comment #4)
> Created an attachment (id=24053)

--> (http://bugs.winehq.org/attachment.cgi?id=24053) [details]
> test diff

(In reply to comment #5)
> Excellent work. Please submit the test/patch to [e-mail]

A small comment here. Could you reduce the buffer size? Or allocating 8K really
means something?

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #7 from Juan Lang   2009-10-12 17:08:10 ---
(In reply to comment #4)
> Created an attachment (id=24053)

--> (http://bugs.winehq.org/attachment.cgi?id=24053) [details]
> test diff

One further small comment:  any tests that currently fail on wine need to be
preceded with todo_wine.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

A. Kochetkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #24051|0                           |1
        is obsolete|                            |
  Attachment #24053|0                           |1
        is obsolete|                            |

--- Comment #8 from A. Kochetkov   2009-10-13 03:54:28 ---
Created an attachment (id=24070)
 --> (http://bugs.winehq.org/attachment.cgi?id=24070)
New version of patch

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #9 from A. Kochetkov   2009-10-13 03:55:09 ---
(In reply to comment #6)
> A small comment here. Could you reduce the buffer size? Or allocating 8K really
> means something?

It doen't means something for wine, but under XP with 95, 98, Me compatability
mode, new allocated string is outside of previous if CHUNK_SIZE is less than
16384. Range test fails, so I cannot assert is null terminator always present.
But, with 16394, range test pass, new string allocated on the old string with
trash, and new string is null terminated.

(In reply to comment #7)
> One further small comment:  any tests that currently fail on wine need to be
> preceded with todo_wine.

I provide bug fix and test. Test with bug fix doesn't fail. Should I add
todo_wine?

Please review another version of patch.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #10 from Nikolay Sivov   2009-10-13 04:25:01 ---
(In reply to comment #9)
> (In reply to comment #6)
> > A small comment here. Could you reduce the buffer size? Or allocating 8K really
> > means something?
>
> It doen't means something for wine, but under XP with 95, 98, Me compatability
> mode, new allocated string is outside of previous if CHUNK_SIZE is less than
> 16384. Range test fails, so I cannot assert is null terminator always present.
> But, with 16394, range test pass, new string allocated on the old string with
> trash, and new string is null terminated.

Didn't know that, if it will fail on native you should use required 8K size of
course.
> (In reply to comment #7)
> > One further small comment:  any tests that currently fail on wine need to be
> > preceded with todo_wine.
>
> I provide bug fix and test. Test with bug fix doesn't fail. Should I add
> todo_wine?

Of course not, patch shouldn't bring new failures including success inside
todo_wine.
> Please review another version of patch.

Send it to wine-patches for review.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

--- Comment #11 from A. Kochetkov   2009-10-13 05:20:00 ---
The patch was sended to [e-mail].
Nikolay Sivov, Austin English, Juan Lang thanks a lot for your help.

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

A. Kochetkov  changed:

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

--- Comment #12 from A. Kochetkov   2009-10-14 10:13:55 ---
Fixed, in commits:
http://source.winehq.org/git/wine.git/?a=commit;h=ae4de9d6956a18b2e8bff0cfd2...
http://source.winehq.org/git/wine.git/?a=commit;h=b6da072551cf6eb88d1500df51...

[Bug 20334] SysReAllocStringLen should return null terminated strings

http://bugs.winehq.org/show_bug.cgi?id=20334

Alexandre Julliard  changed:

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

--- Comment #13 from Alexandre Julliard   2009-10-23 13:19:35 ---
Closing bugs fixed in 1.1.32.