comctl32.dll: Running siap and modules under wine, first steps

[Maximiliano Curia]

Hi,

I'm trying to use a program called "siap"[1] under wine. It's the required 
program to pay taxes that the Argentinian government uses.
So, it's quite important for the free software community in Argentina to be 
able to use it in a free software environment. However, I haven't heard 
of any real success on it, so far.

I've been trying to make it run using winetools[2], and differents versions of 
wine. The furthest I've gone, has been with version 20050310 and all the 
"plugins" that winetools provides, installed.

Only then, could I install the application, run it, switch between the 
different modules (although it worked only sometimes), but I can't really use 
it.

There are a couple of important widgets that seem to be incompatible with 
wine. But as I'm not an expert neither in Windows nor in wine, the one I could 
really track down is a TreeView widget (from comctl32.dll) (from the main 
window of the program), that is not fully implemented in wine, and by using a 
native comctl32.dll (I tried using the win98, winme, win2k libraries) the 
widget would work but some of the other stuff, like switching 
to one of the modules wouldn't.

I made a screenshot to make this clearer, the wine version[3] and the w2k 
version[4].

So, what should I do next?

Is there some documentation of how to implement a certain Windows widget in 
wine?

Is anybody working in this dll, or has worked on this widget?

I'd appreciate any comments.

Thanks,

[1] http://www.afip.gov.ar/programas/siap_main.asp (spanish only)
[2] http://www.von-thadden.de/Joachim/WineTools/
[3] http://gnuservers.com.ar/~maxy/wine/siap-wine.png
[4] http://gnuservers.com.ar/~maxy/wine/siap-w2k.png

-- 
Saludos,
 /\/\ /\ >< `/

[Robert Shearman]

Maximiliano Curia wrote:
> ...
>
> There are a couple of important widgets that seem to be incompatible with 
> wine. But as I'm not an expert neither in Windows nor in wine, the one I could 
> really track down is a TreeView widget (from comctl32.dll) (from the main 
> window of the program), that is not fully implemented in wine, and by using a 
> native comctl32.dll (I tried using the win98, winme, win2k libraries) the 
> widget would work but some of the other stuff, like switching 
> to one of the modules wouldn't.
>
> I made a screenshot to make this clearer, the wine version[3] and the w2k 
> version[4].
>
> So, what should I do next?

Start the program with WINEDEBUG=+treeview wine program.exe &> 
treeview.log and send the treeview log to the list (compressed if it is 
over 50kB).
> Is there some documentation of how to implement a certain Windows widget in 
> wine?

Not really. My guess is that either some message is failing because it 
is hitting some border case.
> Is anybody working in this dll, or has worked on this widget?

I have worked on the control, but I don't really have time to look at 
the specific problem, other than giving a 5 minute look over any log 
files you send.

-- 
Rob Shearman

[Maximiliano Curia]

On Saturday 17 September 2005 20:45, Robert Shearman wrote:
> Start the program with WINEDEBUG=+treeview wine program.exe &>
> treeview.log and send the treeview log to the list (compressed if it is
> over 50kB).

It's about 31 Kb, it gets to the line 272 to open the main window and showing 
the treeview, as shown in the screenshot. The rest is just the program 
closing.

I don't know if this could actually help but, I've noticed, after looking 
fiercely at the log, that the tree is being set to expand, but the Paint 
seems to only be calling DrawItem to the root (twice), but looking at the 
code I can't tell wheter the TREEVIEW_Expand is not setting the siblings to 
visible or if TREEVIEW_Paint is just failling.

-- 
Saludos,
 /\/\ /\ >< `/



No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.2/105 - Release Date: 19/09/05

[Dimi Paun]

On Sat, 2005-09-17 at 17:46 -0300, Maximiliano Curia wrote:
> Is anybody working in this dll, or has worked on this widget?

Nobody currently works on the treeview unfortunately, and it
needs love. Lots of it. If you want to take a stab at it,
you can find it in dlls/comctl32/treeview.c

-- 
Dimi Paun <dimi(a)lattica.com>
Lattica, Inc.

[Alex Villací­s Lasso]

Maximiliano Curia wrote:
> Hi,
>
> I'm trying to use a program called "siap"[1] under wine. It's the required 
> program to pay taxes that the Argentinian government uses.
> So, it's quite important for the free software community in Argentina to be 
> able to use it in a free software environment. However, I haven't heard 
> of any real success on it, so far.
>
> I've been trying to make it run using winetools[2], and differents versions of 
> wine. The furthest I've gone, has been with version 20050310 and all the 
> "plugins" that winetools provides, installed.

I found some free time to investigate the SIAP program issue. As I can 
remember,
the SIAP package is a software program to prepare and submit taxes for the
government of Argentina. This software is written in MS Visual Basic 5.

When an attempt is made to run the program without any overrides with the
current CVS (2005/09/23), the following messages appear (with 
WINEDEBUG=+ole):

...
trace:ole:ITypeLib2_fnRelease (0x7cb9f198)->(1)
trace:ole:WINE_StringFromCLSID 
0x2311b940->{00028C01-0000-0000-0000-000000000046}
trace:ole:LoadTypeLib (L"C:\\ARCHIVOS DE 
PROGRAMA\\S.I.AP\\AFIP\\DBGRID32.OCX",0x7fa5f658)
trace:ole:LoadTypeLibEx (L"C:\\ARCHIVOS DE 
PROGRAMA\\S.I.AP\\AFIP\\DBGRID32.OCX",0,0x7fa5f658)
trace:ole:LoadTypeLibEx File L"C:\\ARCHIVOS DE 
PROGRAMA\\S.I.AP\\AFIP\\DBGRID32.OCX" index 1
trace:ole:TLB_ReadTypeLib cache hit
trace:ole:ITypeLib2_fnAddRef (0x7cb9f198)->ref was 1
trace:ole:LoadTypeLibEx  returns 00000000
trace:ole:LoadRegTypeLib (IID: {00028c01-0000-0000-0000-000000000046}) 
load SUCCESS (0x7cb9f198)
trace:ole:ITypeLib2_fnGetTypeInfoOfGuid (0x7cb9f198)
    guid:    {00028c02-0000-0000-0000-000000000046})
trace:ole:ITypeLib2_fnGetTypeInfoOfGuid -- found (0x7cbb3bd8, 
L"IMsDgridCtrl")
trace:ole:ITypeLib2_fnAddRef (0x7cb9f198)->ref was 2
trace:ole:ITypeInfo_fnAddRef (0x7cbb3bd8)->ref is 3
trace:ole:ITypeLib2_fnRelease (0x7cb9f198)->(2)
trace:ole:ITypeInfo_fnGetTypeAttr (0x7cbb3bd8)
trace:ole:ITypeInfo_fnReleaseTypeAttr (0x7cbb3bd8)->(0x7fdb8e58)
trace:ole:ITypeInfo_fnGetFuncDesc (0x7cbb3bd8) index 0
trace:ole:ITypeInfo_fnReleaseFuncDesc (0x7cbb3bd8)->(0x7cbb3d48)
trace:ole:ITypeInfo_fnGetFuncDesc (0x7cbb3bd8) index 1
...
trace:ole:ITypeInfo_fnReleaseFuncDesc (0x7cbb3bd8)->(0x7cbb4eb8)
trace:ole:ITypeInfo_fnGetFuncDesc (0x7cbb3bd8) index 12
fixme:ole:ITypeInfo_fnGetRefTypeInfo Can't find pRefType for ref 19
trace:ole:ITypeInfo_fnGetRefTypeInfo (0x7cbb3bd8) hreftype 0x0019 loaded 
FAILURE (0x7ca52dc8)
wine: Unhandled exception (thread 0009), starting debugger...
WineDbg starting on pid 0x8
Unhandled exception: page fault on read access to 0x00000010 in 32-bit 
code (0x230f4f61).
In 32 bit mode.
Register dump:
 CS:0073 SS:007b DS:007b ES:007b FS:1007 GS:0033
 EIP:230f4f61 ESP:7fa5f5bc EBP:7fa5f624 EFLAGS:00010202(   - 00      - -RI1)
 EAX:00000004 EBX:00000020 ECX:7fa5f5d0 EDX:7ca52dc8
 ESI:7ca52ddc EDI:7ca52b70
Stack dump:
0x7fa5f5bc:  7ca52dc8 7fa5f5d0 7ca52df8 7cbb3bd8
0x7fa5f5cc:  7ca52dc8 7fa5f624 230f4ae8 7cbb3bd8
0x7fa5f5dc:  7ca52ddc 7cbb5048 7ca52b70 7ca505a0
0x7fa5f5ec:  7f6bb480 7f6857a0 7f686444 00000000
0x7fa5f5fc:  7fdb8e58 0000000c 00000032 00000011
0x7fa5f60c:  7ca52b88 7ca52b70 7cbb5028 7fa5f660
0200: sel=1007 base=7fe9a000 limit=00001fff 32-bit rw-
Backtrace:
=>1 0x230f4f61 in dbgrid32.ocx (+0x34f61) (0x7fa5f624)
  2 0x230e9c2f in dbgrid32.ocx (+0x29c2f) (0x7fa5f66c)
  3 0x230e7573 in dbgrid32.ocx (+0x27573) (0x7fa5f6b4)
  4 0x230c3bee in dbgrid32.ocx (+0x3bee) (0x7fa5f6e8)
  5 0x230cd0a9 in dbgrid32.ocx (+0xd0a9) (0x7fa5f754)
err:dbghelp:pe_load_dbg_file -Unable to peruse .DBG file MSVBVM50.dbg 
("MSVBVM50.dbg")
  6 0x0f045840 in msvbvm50 (+0x45840) (0x7fa5f7a4)
  7 0x0f046096 in msvbvm50 (+0x46096) (0x7fa5f7b4)
  8 0x0f0293dd in msvbvm50 (+0x293dd) (0x7fa5f7f4)
  9 0x0f028f7b in msvbvm50 (+0x28f7b) (0x7fa5f824)
  10 0x0f028e6e in msvbvm50 (+0x28e6e) (0x7fa5f858)
  11 0x0f028ed5 in msvbvm50 (+0x28ed5) (0x7fa5f890)
  12 0x0f029feb in msvbvm50 (+0x29feb) (0x7fa5f8c8)
  13 0x0f029b4e in msvbvm50 (+0x29b4e) (0x7fa5f928)
  14 0x0f029a57 in msvbvm50 (+0x29a57) (0x7fa5f94c)
  15 0x0f0299e9 in msvbvm50 (+0x299e9) (0x7fa5f96c)
  16 0x0f023e77 in msvbvm50 (+0x23e77) (0x7fa5f9a4)
  17 0x0f027329 in msvbvm50 (+0x27329) (0x7f1e5420)
  18 0x00000000 (0x0f10e1f0)
0x230f4f61: call    *0xc(%eax)
fixme:winedbg:be_i386_is_func_call Unsupported yet call insn (0x10) at 
0x230f4f61
Modules:
...

This result suggests that builtin oleaut32 is unable to cope with the 
typelib
embedded in DBGRID32.OCX. When *both* native oleaut32 and ole32 are 
used, the
program starts up normally.

The following test was made: with MS Visual Basic 6 SP6, I made a sample 
program
that includes a Data control, a DBGrid control from DBGRID32.OCX, and a 
sample
MDB file to query into the DBGrid. This sample program crashes in the 
exact same
way as the SIAP package when no overrides are used, and also runs 
correctly when
native ole32 and oleaut32 are used. After some trial and error, I came 
up with
the following patch:

--- wine-20050830-cvs/dlls/oleaut32/typelib.c   2005-09-21 
10:39:22.000000000 -0500
+++ wine-20050830-cvs-patch/dlls/oleaut32/typelib.c     2005-09-24 
20:34:32.000000000 -0500
@@ -5207,9 +5207,11 @@
     ITypeInfoImpl *This = (ITypeInfoImpl *)iface;
     HRESULT result = E_FAIL;

-    if (hRefType == -1 &&
+    if ((hRefType == -1 &&
        (((ITypeInfoImpl*) This)->TypeAttr.typekind   == TKIND_DISPATCH) &&
-       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  TYPEFLAG_FDUAL))
+       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  TYPEFLAG_FDUAL)) ||
+       ((((ITypeInfoImpl*) This)->TypeAttr.typekind   == TKIND_DISPATCH) &&
+             (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
TYPEFLAG_FDISPATCHABLE)))
     {
          /* when we meet a DUAL dispinterface, we must create the interface
          * version of it.

This patch executes the "create interface" code when typekind == 
TKIND_DISPATCH
and wTypeFlags has TYPEFLAG_FDISPATCHABLE, regardless of the value of 
hRefType,
in addition to the previous condition. With this patch, the native oleaut32
requirement is removed. The typelib test still passes - not that it 
actually
tests the change in code (more on this later). One more VB6 application 
I wrote
also works correctly before as well as after the change. However, I 
don't know
enough about typelibs to evaluate whether the previous code was wrong or 
whether
the change is the "right" thing to do. I must note that the test program 
still
requires native ole32 - when an attempt to run it is made otherwise, the
CLASS_E_NOAGGREGATION error is triggered from OleCreateDefaultHandler() in
defaulthandler.c (what is the purpose of this error anyway?).

BTW, is there a way to extract an IDL definition from a binary OCX file? 
This
could be a basis to build a more complete test for typelib - there is a
difference between native and wine that DBGRID32.OCX depends upon
(please ignore, I found out about oleview.exe in Visual Studio, but I am 
still curious
about whether the bug "corrected" above is in fact exposed by the IDL of 
the
problematic control).

UPDATE: The OleCreateDefaultHandler routine seems to be in error in 
returning
the CLASS_E_NOAGGREGATION error in the DBGRID32.OCX case. If I read the 
comment
correctly (and *please* tell me so if I am not), the aggregation case
(pUnkOuter != NULL) is forbidden when riid is other than IID_IUnknown, and
therefore the test is reversed from what it should be:

--- wine-20050830-cvs/dlls/ole32/defaulthandler.c    2005-09-23 
10:51:36.000000000 -0500
+++ wine-20050830-cvs-patch/dlls/ole32/defaulthandler.c    2005-09-25 
19:19:31.000000000 -0500
@@ -1422,7 +1422,7 @@
    * This is necessary because it's the only time the non-delegating
    * IUnknown pointer can be returned to the outside.
    */
-  if (pUnkOuter && IsEqualIID(&IID_IUnknown, riid))
+  if (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid))
     return CLASS_E_NOAGGREGATION;
 
   /*

This patch allows the DBGrid test to move to the next error: the
DataCache_GetAdvise function crashes after the above patch is applied, 
because
this->sinkInterface is NULL at the time of call, but it does not make 
any check
for it:

--- wine-20050830-cvs/dlls/ole32/datacache.c    2005-07-27 
10:49:38.000000000 -0500
+++ wine-20050830-cvs-patch/dlls/ole32/datacache.c    2005-09-25 
19:45:48.000000000 -0500
@@ -1390,9 +1390,11 @@
 
   if (ppAdvSink!=NULL)
   {
-    IAdviseSink_QueryInterface(this->sinkInterface,
+    if (this->sinkInterface != NULL)
+        IAdviseSink_QueryInterface(this->sinkInterface,
                    &IID_IAdviseSink,
                    (void**)ppAdvSink);
+    else *ppAdvSink = NULL;
   }
 
   return S_OK;

With the above two patches, the DBGrid test sheds all dependencies on native
oleaut32 *and* ole32. In addition, the SIAP package is freed from the 
dependence
on native ole32. I will try to investigate later the remaining dependency on
native oleaut32.

Mr. Maximiliano Curia, could you please indicate where to get sample 
modules for SIAP? The default installer
doesn't seem to come with any modules inside. Of course, this might be 
yet another Wine bug, so please confirm.

Alex Villacís Lasso

[Maximiliano Curia]

On Monday 26 September 2005 15:00, Alex Villací­s Lasso wrote:
> --- wine-20050830-cvs/dlls/oleaut32/typelib.c   2005-09-21
> 10:39:22.000000000 -0500
> +++ wine-20050830-cvs-patch/dlls/oleaut32/typelib.c     2005-09-24
> 20:34:32.000000000 -0500
> This patch executes the "create interface" code when typekind ==
> TKIND_DISPATCH
> and wTypeFlags has TYPEFLAG_FDISPATCHABLE, regardless of the value of
> hRefType,
> in addition to the previous condition.

I backported this patch to the version 20050310.
If you are willing to let the test application available I could check the 
successfullness.
> --- wine-20050830-cvs/dlls/ole32/defaulthandler.c    2005-09-23
> 10:51:36.000000000 -0500
> +++ wine-20050830-cvs-patch/dlls/ole32/defaulthandler.c    2005-09-25
> 19:19:31.000000000 -0500
> @@ -1422,7 +1422,7 @@
>     * This is necessary because it's the only time the non-delegating
>     * IUnknown pointer can be returned to the outside.
>     */
> -  if (pUnkOuter && IsEqualIID(&IID_IUnknown, riid))
> +  if (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid))
>      return CLASS_E_NOAGGREGATION;
>
> /*

IsEqualIID was introduced after the version 20050310 that I'm using, and it 
seem that the:
  if ( (pUnkOuter!=NULL) &&
       (memcmp(&IID_IUnknown, riid, sizeof(IID_IUnknown)) != 0) )
was replaced wrongly replaced.

The (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid)), should have the same 
efect as the previous code.
> Mr. Maximiliano Curia, could you please indicate where to get sample
> modules for SIAP? The default installer
> doesn't seem to come with any modules inside. Of course, this might be
> yet another Wine bug, so please confirm.

Yes, of course, they are in:
http://www.afip.gov.ar/seccion_rg.asp#programas

The most used are:
ftp://www.afip.gov.ar/iva40/iva_v4r1.exe (Taxes)
ftp://www.afip.gov.ar/sijyp25/sijyp.25r0.exe (Pensions)

Although, there are different programs for different activities (and I use 12 
of these programs), these two are used by most of them.

-- 
Saludos,
 /\/\ /\ >< `/

FIXED: Running siap and modules under wine, first steps

[Alex Villací­s Lasso]

Maximiliano Curia wrote:
> On Monday 26 September 2005 15:00, Alex Villací­s Lasso wrote:
> > --- wine-20050830-cvs/dlls/oleaut32/typelib.c   2005-09-21
> > 10:39:22.000000000 -0500
> > +++ wine-20050830-cvs-patch/dlls/oleaut32/typelib.c     2005-09-24
> > 20:34:32.000000000 -0500
> > This patch executes the "create interface" code when typekind ==
> > TKIND_DISPATCH
> > and wTypeFlags has TYPEFLAG_FDISPATCHABLE, regardless of the value of
> > hRefType,
> > in addition to the previous condition.
>
> I backported this patch to the version 20050310.
> If you are willing to let the test application available I could check the 
> successfullness.
> > --- wine-20050830-cvs/dlls/ole32/defaulthandler.c    2005-09-23
> > 10:51:36.000000000 -0500
> > +++ wine-20050830-cvs-patch/dlls/ole32/defaulthandler.c    2005-09-25
> > 19:19:31.000000000 -0500
> > @@ -1422,7 +1422,7 @@
> >     * This is necessary because it's the only time the non-delegating
> >     * IUnknown pointer can be returned to the outside.
> >     */
> > -  if (pUnkOuter && IsEqualIID(&IID_IUnknown, riid))
> > +  if (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid))
> >      return CLASS_E_NOAGGREGATION;
> >
> > /*
>
> IsEqualIID was introduced after the version 20050310 that I'm using, and it 
> seem that the:
>   if ( (pUnkOuter!=NULL) &&
>        (memcmp(&IID_IUnknown, riid, sizeof(IID_IUnknown)) != 0) )
> was replaced wrongly replaced.
>
> The (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid)), should have the same 
> efect as the previous code.
> > Mr. Maximiliano Curia, could you please indicate where to get sample
> > modules for SIAP? The default installer
> > doesn't seem to come with any modules inside. Of course, this might be
> > yet another Wine bug, so please confirm.
>
> Yes, of course, they are in:
> http://www.afip.gov.ar/seccion_rg.asp#programas
>
> The most used are:
> ftp://www.afip.gov.ar/iva40/iva_v4r1.exe (Taxes)
> ftp://www.afip.gov.ar/sijyp25/sijyp.25r0.exe (Pensions)
>
> Although, there are different programs for different activities (and I use 12 
> of these programs), these two are used by most of them.

As of 2006-06-09, the CVS snapshot of Wine seems to run SIAP correctly. 
No crashes at all. Could you please be so kind as to test with the *most 
recent CVS snapshot* and confirm?


Alex Villacís Lasso

[Robert Shearman]

Alex Villací­s Lasso wrote:
> --- wine-20050830-cvs/dlls/oleaut32/typelib.c   2005-09-21 
> 10:39:22.000000000 -0500
> +++ wine-20050830-cvs-patch/dlls/oleaut32/typelib.c     2005-09-24 
> 20:34:32.000000000 -0500
> @@ -5207,9 +5207,11 @@
>     ITypeInfoImpl *This = (ITypeInfoImpl *)iface;
>     HRESULT result = E_FAIL;
>
> -    if (hRefType == -1 &&
> +    if ((hRefType == -1 &&
>        (((ITypeInfoImpl*) This)->TypeAttr.typekind   == 
> TKIND_DISPATCH) &&
> -       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  TYPEFLAG_FDUAL))
> +       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
> TYPEFLAG_FDUAL)) ||
> +       ((((ITypeInfoImpl*) This)->TypeAttr.typekind   == 
> TKIND_DISPATCH) &&
> +             (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
> TYPEFLAG_FDISPATCHABLE)))
>     {
>          /* when we meet a DUAL dispinterface, we must create the 
> interface
>          * version of it.
>
> This patch executes the "create interface" code when typekind == 
> TKIND_DISPATCH
> and wTypeFlags has TYPEFLAG_FDISPATCHABLE, regardless of the value of 
> hRefType,
> in addition to the previous condition. With this patch, the native 
> oleaut32
> requirement is removed. The typelib test still passes - not that it 
> actually
> tests the change in code (more on this later).

This patch looks pretty good, but I can't see the difference between 
TKIND_DISPATCH and TYPEFLAG_FDISPATCHABLE on MSDN. They look like they 
mean the same thing. Have you seen a typelib with one and not the other? 
Also, you can remove all of the casts of This while you're changing that 
if statement because they are unnecessary.
> UPDATE: The OleCreateDefaultHandler routine seems to be in error in 
> returning
> the CLASS_E_NOAGGREGATION error in the DBGRID32.OCX case. If I read 
> the comment
> correctly (and *please* tell me so if I am not), the aggregation case
> (pUnkOuter != NULL) is forbidden when riid is other than IID_IUnknown, 
> and
> therefore the test is reversed from what it should be:
>
> --- wine-20050830-cvs/dlls/ole32/defaulthandler.c    2005-09-23 
> 10:51:36.000000000 -0500
> +++ wine-20050830-cvs-patch/dlls/ole32/defaulthandler.c    2005-09-25 
> 19:19:31.000000000 -0500
> @@ -1422,7 +1422,7 @@
>    * This is necessary because it's the only time the non-delegating
>    * IUnknown pointer can be returned to the outside.
>    */
> -  if (pUnkOuter && IsEqualIID(&IID_IUnknown, riid))
> +  if (pUnkOuter && !IsEqualIID(&IID_IUnknown, riid))
>     return CLASS_E_NOAGGREGATION;
>
> /*

This patch is obviously correct.
> This patch allows the DBGrid test to move to the next error: the
> DataCache_GetAdvise function crashes after the above patch is applied, 
> because
> this->sinkInterface is NULL at the time of call, but it does not make 
> any check
> for it:
>
> --- wine-20050830-cvs/dlls/ole32/datacache.c    2005-07-27 
> 10:49:38.000000000 -0500
> +++ wine-20050830-cvs-patch/dlls/ole32/datacache.c    2005-09-25 
> 19:45:48.000000000 -0500
> @@ -1390,9 +1390,11 @@
>
> if (ppAdvSink!=NULL)
>   {
> -    IAdviseSink_QueryInterface(this->sinkInterface,
> +    if (this->sinkInterface != NULL)
> +        IAdviseSink_QueryInterface(this->sinkInterface,
>                    &IID_IAdviseSink,
>                    (void**)ppAdvSink);
> +    else *ppAdvSink = NULL;
>   }
>
> return S_OK;
>
> With the above two patches, the DBGrid test sheds all dependencies on 
> native
> oleaut32 *and* ole32. In addition, the SIAP package is freed from the 
> dependence
> on native ole32. I will try to investigate later the remaining 
> dependency on
> native oleaut32.

This patch is also correct.

Please send all of these patches to wine-patches.

-- 
Rob Shearman

[Alex Villací­s Lasso]

Robert Shearman wrote:
> Alex Villací­s Lasso wrote:
> > --- wine-20050830-cvs/dlls/oleaut32/typelib.c   2005-09-21 
> > 10:39:22.000000000 -0500
> > +++ wine-20050830-cvs-patch/dlls/oleaut32/typelib.c     2005-09-24 
> > 20:34:32.000000000 -0500
> > @@ -5207,9 +5207,11 @@
> >     ITypeInfoImpl *This = (ITypeInfoImpl *)iface;
> >     HRESULT result = E_FAIL;
> >
> > -    if (hRefType == -1 &&
> > +    if ((hRefType == -1 &&
> >        (((ITypeInfoImpl*) This)->TypeAttr.typekind   == 
> > TKIND_DISPATCH) &&
> > -       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
> > TYPEFLAG_FDUAL))
> > +       (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
> > TYPEFLAG_FDUAL)) ||
> > +       ((((ITypeInfoImpl*) This)->TypeAttr.typekind   == 
> > TKIND_DISPATCH) &&
> > +             (((ITypeInfoImpl*) This)->TypeAttr.wTypeFlags &  
> > TYPEFLAG_FDISPATCHABLE)))
> >     {
> >          /* when we meet a DUAL dispinterface, we must create the 
> > interface
> >          * version of it.
> >
> > This patch executes the "create interface" code when typekind == 
> > TKIND_DISPATCH
> > and wTypeFlags has TYPEFLAG_FDISPATCHABLE, regardless of the value of 
> > hRefType,
> > in addition to the previous condition. With this patch, the native 
> > oleaut32
> > requirement is removed. The typelib test still passes - not that it 
> > actually
> > tests the change in code (more on this later).
>
> This patch looks pretty good, but I can't see the difference between 
> TKIND_DISPATCH and TYPEFLAG_FDISPATCHABLE on MSDN. They look like they 
> mean the same thing. Have you seen a typelib with one and not the 
> other? Also, you can remove all of the casts of This while you're 
> changing that if statement because they are unnecessary.

I need to read the Automation guide throughly. Back when I was trying to 
make this app work, the telltale sign was the "Can't find pRefType for 
ref XXX" right before the crash. Since I was at my home machine, and I 
don't have an Internet connection at home, I could not check whether 
TKIND_DISPATCH and TYPEFLAG_FDISPATCHABLE mean the same thing or should 
be specified together. I can send the (reverse-engineered with OLEVIEW) 
typelib for DBGRID32.OCX, if that might show something.

I already sent the other two patches to wine-patches. I hope they get 
commited.

Alex Villacís Lasso