[Bug 14641] New: OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Summary: OLE implementation improvement
           Product: Wine
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P2
         Component: ole32
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Created an attachment (id=15053)
 --> (http://bugs.winehq.org/attachment.cgi?id=15053)
add default inprocess handler support

Out-of-process OLE component with a in-process handler are not managed. For
exemple, with Microsoft Office Powerpoint 2003, it is impossible to insert a
graph or an equation. An "server not found" error occurs.

I'm currently trying to implement this. I have good results for instance.
Object can be inserted and the corresponding server is launched. But it seems I
still have problem with datas transfert because the inserted object is blank
and no change are saved.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #1 from François DORIN   2008-07-26 08:34:49 ---
Created an attachment (id=15054)
 --> (http://bugs.winehq.org/attachment.cgi?id=15054)
needed file for default inprocess_handler implementation

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #2 from François DORIN   2008-07-26 08:35:36 ---
Created an attachment (id=15055)
 --> (http://bugs.winehq.org/attachment.cgi?id=15055)
needed file for default inprocess_handler implementation

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #3 from François DORIN   2008-07-26 08:36:08 ---
Created an attachment (id=15056)
 --> (http://bugs.winehq.org/attachment.cgi?id=15056)
needed file for default inprocess_handler implementation

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #4 from François DORIN   2008-07-26 08:36:27 ---
Created an attachment (id=15057)
 --> (http://bugs.winehq.org/attachment.cgi?id=15057)
needed file for default inprocess_handler implementation

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #5 from François DORIN   2008-07-26 08:38:02 ---
Created an attachment (id=15058)
 --> (http://bugs.winehq.org/attachment.cgi?id=15058)
fix a wrong return value for DefaultHandler_SetExtent

This patch was also submitted to [e-mail]

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #6 from François DORIN   2008-07-26 08:40:00 ---
Created an attachment (id=15059)
 --> (http://bugs.winehq.org/attachment.cgi?id=15059)
modify the beheavior of a non implemented function

Modify the behavior of GetDestCtx to return MSHCTX_NOSHAREDMEM instead of
MSHCTX_DIFFERENTMACHINE to allow a handle marshalling.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #7 from François DORIN   2008-07-26 08:45:08 ---
Created an attachment (id=15060)
 --> (http://bugs.winehq.org/attachment.cgi?id=15060)
fix a bug when marshalling a null pointer string

Fix a crash when NdrConformantString receive a null pointer. I don't know if
these functions can deal with null pointer on Windows. I didn't found any
documentation on this point. So, I don't know if this behavior is correct or
not.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #8 from François DORIN   2008-07-26 08:46:04 ---
Created an attachment (id=15061)
 --> (http://bugs.winehq.org/attachment.cgi?id=15061)
fix a wrong offset managment

This patch was also submited to [e-mail]

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #9 from François DORIN   2008-07-26 08:51:05 ---
Created an attachment (id=15062)
 --> (http://bugs.winehq.org/attachment.cgi?id=15062)
fix an exception when dealing with a unknown TYMED value

A unknown tymed value raises an exception and stops the program.

I don't know if this behavior is wanted or not, but it seems strange for me.
So, I just do nothing with a FIXME instead of raising an exception.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #10 from Vitaliy Margolen   2008-07-26 11:34:43 ---
Please send all patches to wine-patches mailing list. Patches are not being
picked up from bugzilla.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #11 from François DORIN   2008-07-26 12:42:18 ---
(In reply to comment #10)
> Please send all patches to wine-patches mailing list. Patches are not being
> picked up from bugzilla.

I already sent some of them. But others are tricks to try having a working
default inprocess handler. And while this doesn't work, I'm not sure it's a
good idea to send these patches to the wine-patches mailing list, because some
of the modification can introduce bugs and/or regressions.

Moreover, these bugs can be due to a wrong implementation of what I implemented
and correct the bad effects, but not correct the source of the problem.

Do I have to send them anyway ?

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Dan Kegel  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
           Severity|enhancement                 |normal

--- Comment #12 from Dan Kegel   2008-07-27 03:36:03 ---
Hi again Francois!  
Please write conformance tests for everything - you can
submit those (with todo_wine surrounding the tests that
fail) even if your implementation isn't ready to submit.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

François DORIN  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #15060|0                           |1
        is obsolete|                            |

--- Comment #13 from François DORIN   2008-07-28 10:56:18 ---
(From update of attachment 15060)
After some tests on windows, an exception is raised when a null pointer is
passed in pszMssage and in pMemory. So, the patch is useless. I mark it as
obsolete.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #14 from François DORIN   2008-07-28 11:01:49 ---
Hi Dan,

As you can see with my previous post, I'm currently making some test under
windows. After that, I will try to write some conformant tests.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #15 from Dan Kegel   2008-07-28 13:27:29 ---
Excellent.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #16 from François DORIN   2008-07-29 05:29:30 ---
(From update of attachment 15061)
The bug is in fact a widl one.

For more information :
http://www.winehq.org/pipermail/wine-devel/2008-June/066587.html

Thanks to Robert Shearman for these links.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

François DORIN  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #15062|0                           |1
        is obsolete|                            |

--- Comment #17 from François DORIN   2008-07-31 12:44:03 ---
(From update of attachment 15062)
This patch seems to be useless now. I can't reproduce the error for which I
wrote it. I mark it as obsolete.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #18 from François DORIN   2008-07-31 12:47:17 ---
I'm looking for writting a conformance test for NdrConformantString, but in the
rpcrt4/tests/ndr_marshal.c exceptions seems not to be take into account.

Is it possible to check a function raises the right exception when a bad
parameter is given, for exemple ? If yes, how to do this ?

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Rob Shearman  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1

--- Comment #19 from Rob Shearman   2008-09-11 01:33:03 ---
It has recently become possible to write tests for functions raising
non-continuable exceptions, but yours will be the first to use the new
functionality.

What exception is being raised by NdrConformantStringMarshall?

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

--- Comment #20 from Austin English   2009-03-30 12:51:50 ---
Is this still an issue in current (1.1.18 or newer) wine?

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |ABANDONED

--- Comment #21 from Austin English   2009-10-01 20:48:30 ---
Abandoned. If this is still an issue in current (1.1.30 or newer) wine, and you
can provide the needed information, feel free to reopen.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Jeff Zaroyko  changed:

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

--- Comment #22 from Jeff Zaroyko   2009-10-02 21:35:40 ---
closing abandoned.

[Bug 14641] OLE implementation improvement

http://bugs.winehq.org/show_bug.cgi?id=14641

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Platform|All                         |Other
         OS/Version|All                         |other

--- Comment #23 from Austin English  2012-02-23 15:03:02 CST ---
Removing deprecated 'All' Platform/OS.