[Bug 17697] New: SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Summary: SNetTerm won't start
           Product: Wine
           Version: 1.1.16
          Platform: Other
               URL: http://www.securenetterm.com
        OS/Version: other
            Status: NEW
          Keywords: download
          Severity: normal
          Priority: P2
         Component: -unknown
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

http://chrisboucek.blogspot.com/2009/03/installing-snetterm-on-debian-etch.h...
complained about not being able to run snetterm, so I tried it.
Sure enough, it complains

Cannot find import; DLL may be missing, corrupt, or wrong version\nFile
"TAPI32.dll", function "lineGetDevCapsW" (error 126)

but looking at +relay, it seems possible this is a diversion, as the
log looks rather like it's checking software licenses right before the
error.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #1 from Andrew Nguyen   2009-03-23 01:38:12 ---
I've submitted a series of tapi32 patches that should allow the application to
start after overriding atl.dll with a native copy:

http://www.winehq.org/pipermail/wine-patches/2009-March/070918.html
http://www.winehq.org/pipermail/wine-patches/2009-March/070919.html
http://www.winehq.org/pipermail/wine-patches/2009-March/070920.html
http://www.winehq.org/pipermail/wine-patches/2009-March/070921.html

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Louis Lenders  changed:

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

--- Comment #2 from Louis Lenders   2009-11-06 09:01:26 ---
(In reply to comment #1)
> I've submitted a series of tapi32 patches that should allow the application to
> start after overriding atl.dll with a native copy:
>
> http://www.winehq.org/pipermail/wine-patches/2009-March/070918.html
> http://www.winehq.org/pipermail/wine-patches/2009-March/070919.html
> http://www.winehq.org/pipermail/wine-patches/2009-March/070920.html
> http://www.winehq.org/pipermail/wine-patches/2009-March/070921.html

Hi Andrew, any news on this one? It looks like three of your four patches went
into git. What happened with the 4th patch?

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #3 from Kenneth Robinette   2010-01-14 13:41:19 ---
Created an attachment (id=25736)
 --> (http://bugs.winehq.org/attachment.cgi?id=25736)
wine SNetTerm dump

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Kenneth Robinette  changed:

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

--- Comment #4 from Kenneth Robinette   2010-01-14 13:43:35 ---
As the author of NetTerm, SecureNetTerm and SNetTerm, I would like to add
additional information on this problem in order to get it install.

First, I have changed all three products to do a dynamic load of the tapi32.dll
to remove loading problems related to tapi.  I have also changed our logic to
use our Installshield installation routine to not install custom fonts.  It
appears that loading any fonts cause major grief to wine users.

In spite of these changes, we have not been able to get SNetTerm to load.  I am
currenty testing on Red Hat Enterprise Linux Server release 5.4 (Tikanga) with
the following wine rpm's installed:

wine-core-1.0.1-1.el5.i386.rpm
wine-desktop-1.0.1-1.el5.i386.rpm
wine-devel-1.0.1-1.el5.i386.rpm
wine-ldap-1.0.1-1.el5.i386.rpm
wine-nas-1.0.1-1.el5.i386.rpm
wine-tools-1.0.1-1.el5.i386.rpm

After tracing the problem (log attached), I have determined that SNetTerm is
aborting in the MFC CreateDlgIndirect() function.  I have debugged about as far
as I can without assistance from some wine experts.

Please fill free to contact me at anytime.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #5 from Austin English   2010-01-14 15:34:26 ---
> wine: Call from 0x7ee0d7d0 to unimplemented function atl.dll.AtlModuleRegisterWndClassInfoA, aborting

That function is unimplemented in wine. You can either use
AtlModuleRegisterWndClassInfoW, which is implemented, or wait for someone to
stub/implement it. A stub would only prevent the crash, of course, but not
provide the functionality you need.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #6 from Kenneth Robinette   2010-01-14 16:06:00 ---
(In reply to comment #5)
> > wine: Call from 0x7ee0d7d0 to unimplemented function atl.dll.AtlModuleRegisterWndClassInfoA, aborting
>
> That function is unimplemented in wine. You can either use
> AtlModuleRegisterWndClassInfoW, which is implemented, or wait for someone to
> stub/implement it. A stub would only prevent the crash, of course, but not
> provide the functionality you need.

I saw that call in the trace, but I cannot figure out where the call to that
funtion is coming from.  I searched all our source files and it is not
referenced anywhere.  When I did an internet search, I found that funtion is
located in atl.dll.  I tried adding atl.dll to the .wine windows/system32
directory, but it did not help.  Any ideas?

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #7 from Nikolay Sivov   2010-01-14 16:09:56 ---
(In reply to comment #6)
> (In reply to comment #5)
>
> I saw that call in the trace, but I cannot figure out where the call to that
> funtion is coming from.  I searched all our source files and it is not
> referenced anywhere.

Most likely it's coming from mfc runtime module.
> When I did an internet search, I found that funtion is
> located in atl.dll.  I tried adding atl.dll to the .wine windows/system32
> directory, but it did not help.  Any ideas?

You should set it to native using winecfg.

P.S. This function is trivial to implement on top of W version. You could
checkout Wine git tree and look at dlls/atl path.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #8 from Kenneth Robinette   2010-01-14 18:00:39 ---
(In reply to comment #7)
> (In reply to comment #6)
> > (In reply to comment #5)
> >
> > I saw that call in the trace, but I cannot figure out where the call to that
> > funtion is coming from.  I searched all our source files and it is not
> > referenced anywhere.
>
> Most likely it's coming from mfc runtime module.
> > When I did an internet search, I found that funtion is
> > located in atl.dll.  I tried adding atl.dll to the .wine windows/system32
> > directory, but it did not help.  Any ideas?
>
> You should set it to native using winecfg.
>
> P.S. This function is trivial to implement on top of W version. You could
> checkout Wine git tree and look at dlls/atl path.

Thanks, adding atl.dll to native using winecfg worked.  SNetTerm is now
starting so I can continue to debug the problems.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #9 from Kenneth Robinette   2010-01-16 11:20:04 ---
(In reply to comment #8)
> (In reply to comment #7)
> > (In reply to comment #6)
> > > (In reply to comment #5)
> > >
> > > I saw that call in the trace, but I cannot figure out where the call to that
> > > funtion is coming from.  I searched all our source files and it is not
> > > referenced anywhere.
> >
> > Most likely it's coming from mfc runtime module.
> > > When I did an internet search, I found that funtion is
> > > located in atl.dll.  I tried adding atl.dll to the .wine windows/system32
> > > directory, but it did not help.  Any ideas?
> >
> > You should set it to native using winecfg.
> >
> > P.S. This function is trivial to implement on top of W version. You could
> > checkout Wine git tree and look at dlls/atl path.
>
> Thanks, adding atl.dll to native using winecfg worked.  SNetTerm is now
> starting so I can continue to debug the problems.

Another strange problem.  When I exit SNetTerm, it aborts with the attached
trace.  It looks like the problem is in svc_udp_recv??  Any ideas how to
proceed?

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #10 from Kenneth Robinette   2010-01-16 11:21:01 ---
Created an attachment (id=25753)
 --> (http://bugs.winehq.org/attachment.cgi?id=25753)
Program abort trace

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #11 from Nikolay Sivov   2010-01-16 11:28:48 ---
(In reply to comment #10)
> Created an attachment (id=25753)

--> (http://bugs.winehq.org/attachment.cgi?id=25753) [details]
> Program abort trace

Please open separate report for that. Main rule - one problem per report.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #12 from Kenneth Robinette   2010-01-20 14:43:01 ---
(In reply to comment #11)
> (In reply to comment #10)
> > Created an attachment (id=25753)

--> (http://bugs.winehq.org/attachment.cgi?id=25753) [details] [details]
> > Program abort trace
>
> Please open separate report for that. Main rule - one problem per report.

Confirmed that initial problem was due to:

wine: Call from 0x7ee0d7d0 to unimplemented function
atl.dll.AtlModuleRegisterWndClassInfoA, aborting

Current workaround is to use real Microsoft atl.dll.

This bug is present in wine 1.0.1-1 and 1.1 36

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Nikolay Sivov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|-unknown                    |atl

--- Comment #13 from Nikolay Sivov   2010-01-20 14:47:10 ---
(In reply to comment #12)
> Confirmed that initial problem was due to:
>
> wine: Call from 0x7ee0d7d0 to unimplemented function
> atl.dll.AtlModuleRegisterWndClassInfoA, aborting

Ok. W version is implemented, so it's a trivial task.
> Current workaround is to use real Microsoft atl.dll.

Changing component.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #14 from Kenneth Robinette   2010-01-26 16:47:07 ---
(In reply to comment #13)
> (In reply to comment #12)
> > Confirmed that initial problem was due to:
> >
> > wine: Call from 0x7ee0d7d0 to unimplemented function
> > atl.dll.AtlModuleRegisterWndClassInfoA, aborting
>
> Ok. W version is implemented, so it's a trivial task.
> > Current workaround is to use real Microsoft atl.dll.
>
> Changing component.

I download the current 1.1.37 source and compiled.  Once I got everything
compiled, I installed and checked to make sure SNetTerm was still aborting due
to the atl.dll.AtlModuleRegisterWndClassInfoA funtion.  It was.

I then added the atl.dll.AtlModuleRegisterWndClassInfoA() function to the
dlls/atl/atl_main.c module and it compiled clean.  I then transferred the
resulting atl.dll.so to /usr/local/lib/wine replacing the current version.  I
removed the atl.dll override and ran SNetTerm

I still get the following:

wine: Call from 0x7ee0d7d0 to unimplemented function
atl.dll.AtlModuleRegisterWndClassInfoA, aborting

What else do I need to do test to see if my patch works?

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #15 from Dan Kegel   2010-01-26 17:28:42 ---
Don't transfer anything to /usr.  Just run the 'wine' in the top
level of the source tree.

When in doubt, add debug print statements :-)

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #16 from Vitaliy Margolen   2010-01-26 21:34:50 ---
(In reply to comment #14)
> I then added the atl.dll.AtlModuleRegisterWndClassInfoA() function to the
> dlls/atl/atl_main.c module and it compiled clean.

You need to add it to the .spec file as well.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #17 from Kenneth Robinette   2010-01-27 11:12:20 ---
(In reply to comment #16)
> (In reply to comment #14)
> > I then added the atl.dll.AtlModuleRegisterWndClassInfoA() function to the
> > dlls/atl/atl_main.c module and it compiled clean.
>
> You need to add it to the .spec file as well.

I believe I got AtlModuleRegisterWndClassInfoA implemented correctly, however
now it looks like AtlAxDialogBoxA() is what is really causing SNetTerm to
start.  Both AtlAxDialogBoxA() and AtlAxDialogBoxW() are not implemented in any
version of Wine.

I guess I will just have to use the real Windows XP atl.dll and be done with
it.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #18 from Kenneth Robinette   2010-02-02 11:23:37 ---
(In reply to comment #17)
> (In reply to comment #16)
> > (In reply to comment #14)
> > > I then added the atl.dll.AtlModuleRegisterWndClassInfoA() function to the
> > > dlls/atl/atl_main.c module and it compiled clean.
> >
> > You need to add it to the .spec file as well.
>
> I believe I got AtlModuleRegisterWndClassInfoA implemented correctly, however
> now it looks like AtlAxDialogBoxA() is what is really causing SNetTerm to
> start.  Both AtlAxDialogBoxA() and AtlAxDialogBoxW() are not implemented in any
> version of Wine.
> I guess I will just have to use the real Windows XP atl.dll and be done with
> it.

I added the logic to do a Microsoft minidump when SNetTerm aborts to aid in
tracking down the problem.  If I use the default Wine atl.dll, SNetTerm will
abort on startup (prior to showing the window).  I have tracked it down to an
ActiveX control being initialiazed in the view of SNetTerm.  The following is
the stack trace when the SNeTerm aborts.  Any ideas how to isolate where in the
Wine atl.dll that is causing the problem?

SecureNetTerm!CDataSourceControl::`RTTI Complete Object Locator'
isiTerminal!ATL::CComControlBase::FireViewChange+0x2d
isiTerminal!ATL::IOleObjectImpl<CTerminal>::DoVerbInPlaceActivate+0x29
isiTerminal!ATL::IOleObjectImpl<CTerminal>::DoVerb+0x62
SecureNetTerm!COleControlSite::DoVerb+0x28
WARNING: Frame IP not in any known module. Following frames may be wrong.
0xfffffffb
SecureNetTerm!COleControlContainer::CreateControl+0x77
SecureNetTerm!COleControlContainer::CreateControl+0x5c
SecureNetTerm!COccManager::CreateDlgControl+0x21e
SecureNetTerm!COccManager::CreateDlgControls+0xfb
SecureNetTerm!COccManager::CreateDlgControls+0x55
0x1607fe8
SecureNetTerm!CWnd::WindowProc+0x22
SecureNetTerm!AfxCallWndProc+0x90
SecureNetTerm!AfxWndProc+0x34
user32+0x9c40a
user32+0x9cdcd
user32+0x9f5e0
user32+0x64960
user32+0x68c42

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #19 from Dan Kegel   2010-02-02 11:41:32 ---
We usually use things like WINEDEBUG=+relay,+atl and look at the log,
possibly adding TRACE statements in the component of interest.
You might consider adding a dummy implementation of AtlAxDialogBoxA
and see if that gets called right before the crash.

You can also use winedebug, occasionally that's helpful.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #20 from Kenneth Robinette   2010-02-02 12:20:21 ---
(In reply to comment #19)
> We usually use things like WINEDEBUG=+relay,+atl and look at the log,
> possibly adding TRACE statements in the component of interest.
> You might consider adding a dummy implementation of AtlAxDialogBoxA
> and see if that gets called right before the crash.
>
> You can also use winedebug, occasionally that's helpful.

I tried that before but it did not help.  I did place trace statements in the
fuction CComControlBase::FireViewChange right before the call to:

m_spInPlaceSite->InvalidateRect(NULL,TRUE);

was called, and right after.  It reached the trace statement before the call,
but not after.  So right after the m_spInPlaceSite->InvalidateRect() is where
the abort happens.  What I don't understand is the statement:

SecureNetTerm!CDataSourceControl::`RTTI Complete Object Locator'

That would indicate to me some vector table has been corrupted or not setup
correctly.  And there are no calls to the atl.dll after the FireViewChange()
call.

It just does not make sense.  If I use the default wine atl.dll, it aborts.  It
I override in winecfg and use the native atl.dll, all is well.

I was just hoping someone had seen this before.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #21 from Dan Kegel   2010-02-02 12:33:23 ---
Could you provide source (and executable and how to build)
for a minimal test case, i.e. a very small C++ app
that causes the problem?  Sometimes that's the fastest
route to understanding the cause.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #22 from Kenneth Robinette   2010-02-02 16:27:52 ---
(In reply to comment #21)
> Could you provide source (and executable and how to build)
> for a minimal test case, i.e. a very small C++ app
> that causes the problem?  Sometimes that's the fastest
> route to understanding the cause.

I wish I could, however I have no idea what is causing this.  Perhaps someone
else will come across this and I can get some more clues.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #23 from Dan Kegel   2010-02-02 16:50:49 ---
Can't you use divide and conquor?  Remove part of your app's sources,
see if the problem perists, if not, put them back.  Repeat until you
have a small program with the problem.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #24 from Kenneth Robinette   2010-02-02 16:52:14 ---
(In reply to comment #23)
> Can't you use divide and conquor?  Remove part of your app's sources,
> see if the problem perists, if not, put them back.  Repeat until you
> have a small program with the problem.

One thing I did note in the WINEDEBUG=+atl trace is AtlIPersistStreamInit_Load
is called and that funtion is a dummy one in Wine.  Could that be what is
really causing the problem and the source of the strange 'RTTI Complete Object
Locator' item in the WinDbg output?  The following is the WINEDEBUG output:

/home/zkrr01/.wine/drive_c/Program Files/InterSoft International,
Inc/SecureNetTerm $ WINDEBUG=+atl wine SecureNetTerm
fixme:ole:CoInitializeSecurity ((nil),-1,(nil),(nil),0,3,(nil),0,(nil)) - stub!
fixme:wbemprox:wbem_locator_ConnectServer 0x15d6f8, L"ROOT\CIMV2", (null),
(null), (null), 0x00000000, (null), (nil), 0x13bea84)
fixme:mountmgr:harddisk_ioctl unsupported ioctl 560000
fixme:wbemprox:wbem_locator_ConnectServer 0x15b660, L"ROOT\CIMv2", (null),
(null), (null), 0x00000000, (null), (nil), 0x14bea8c)
fixme:mountmgr:harddisk_ioctl unsupported ioctl 560000
fixme:wbemprox:wbem_locator_ConnectServer 0x15b660, L"ROOT\CIMv2", (null),
(null), (null), 0x00000000, (null), (nil), 0x13bea8c)
fixme:mountmgr:harddisk_ioctl unsupported ioctl 560000
fixme:wbemprox:wbem_locator_ConnectServer 0x15b6b0, L"ROOT\CIMv2", (null),
(null), (null), 0x00000000, (null), (nil), 0x14bea8c)
fixme:wbemprox:wbem_locator_ConnectServer 0x15b6b0, L"ROOT\WMI", (null),
(null), (null), 0x00000000, (null), (nil), 0x13bea84)
fixme:atl:AtlModuleInit SEMI-STUB (0x6801beb0 0x6801b6f0 0x68000000)
fixme:atl:AtlIPersistStreamInit_Load (0x32df20, 0x6801b958, 0x1b30178,
0x1b301e0)
fixme:atl:AtlModuleRegisterWndClassInfoA 0x6801beb0 0x6801b7a8 0x1b301dc
semi-stub
fixme:dbghelp:fetch_thread_info Couldn't open thread 32 (87)
wine: Unhandled page fault on read access to 0x00000098 at address 0x61d2d7
(thread 0009), starting debugger...

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #25 from Dan Kegel   2010-02-02 16:59:32 ---
Could be.   We really need a small program that uses ATL and shows
the problem, though.  Maybe you can find some examples on the net
that use similar functions and misbehave similarly.  Extracting
a test case from your real app is still a good option.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #26 from Kenneth Robinette   2010-02-02 19:05:01 ---
(In reply to comment #25)
> Could be.   We really need a small program that uses ATL and shows
> the problem, though.  Maybe you can find some examples on the net
> that use similar functions and misbehave similarly.  Extracting
> a test case from your real app is still a good option.

It appears that AtlIPersistStreamInit_Load() is the problem.  When I comment
out:

public IPersistStreamInitImpl<CTerminal>
COM_INTERFACE_ENTRY_IMPL(IPersistStreamInit)

in the active x control I am using, the problem goes away.  Of course the
activex control no longer functions as required.

I will continue to search the Internet for a simple example that will confirm.

Of course the final solution will require implementation of that function,
whioh I suspect it not a trival matter since its not alreay done.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #27 from Kenneth Robinette   2010-02-03 07:29:50 ---
Created an attachment (id=26041)
 --> (http://bugs.winehq.org/attachment.cgi?id=26041)
AtlIPersistStreamInit_Load abort fix

Instuctions on how to reproduce the problem with a sample application, a
workaround, and a suggested fix to implement the atl_main function
AtlModuleRegisterWndClassInfoA().

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #28 from Kenneth Robinette   2010-02-03 07:32:16 ---
Resolved SNetTerm load problem.  See attachment on how to duplicate problem and
fix.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Dan Kegel  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch, source
   Target Milestone|---                         |1.2.0

--- Comment #29 from Dan Kegel   2010-02-03 07:45:36 ---
Awesome work.  Is the code your own, i.e. free of any copied Microsoft
code?  If so, please send it to wine-patches as a patch
(see http://wiki.winehq.org/SubmittingPatches ).

core componenet, good analysis with proposed fix -> nominating for 1.2

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #30 from Kenneth Robinette   2010-02-03 08:16:26 ---
(In reply to comment #29)
> Awesome work.  Is the code your own, i.e. free of any copied Microsoft
> code?  If so, please send it to wine-patches as a patch
> (see http://wiki.winehq.org/SubmittingPatches ).
>
> core componenet, good analysis with proposed fix -> nominating for 1.2

All I did was duplicate the current Wine 1.1.37
AtlModuleRegisterWndClassInfoW() function and changed necessary stuff for Ansi
instead of Unicode.

If possible, I don't want to get involved with all the rules/regulations
associated with submitting a patch.  Have a wine expert review my code and
submit the patch.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #31 from Dan Kegel   2010-02-03 08:20:31 ---
OK, thanks.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #32 from Kenneth Robinette   2010-02-18 11:04:58 ---
(In reply to comment #31)
> OK, thanks.

Upon additional testing, if an application is compiled with
_ATL_STATIC_REGISTRY as outlined in my last attachment, it is not necessary to
to implement the ATLModuleRegisterWndClassInfoA() function in Wine.

I would recommend that the implementation of ATLModuleRegisterWndClassInfoA not
be a required fix for the Wine 1.2 release.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

André H.  changed:

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

--- Comment #33 from André H.   2010-05-25 12:15:01 ---
Should be fixed by commit 2d658e656f95070c410db3719ce5d03a4e609daf
please retest

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #34 from Kenneth Robinette   2010-05-25 12:32:13 ---
(In reply to comment #33)
> Should be fixed by commit 2d658e656f95070c410db3719ce5d03a4e609daf
> please retest
> From what I can tell, it did not resolve the problem I had.  See test program:

http://www.codeguru.com/cpp/w-d/dislog/commondialogs/article.php/c1959/

and my AtlPersisStreamInit Load abort fix attachment.

Looks like the static includes of the Microsoft ATL functions are still
required.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #35 from Alexandre Julliard   2010-06-14 13:06:30 ---
The test program works fine now. If you still have problems please try to
attach a small exe demonstrating the bug.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #36 from Kenneth Robinette   2010-06-14 13:41:17 ---
(In reply to comment #35)
> The test program works fine now. If you still have problems please try to
> attach a small exe demonstrating the bug.

I just tried running the demo listed in the attachment, composed of
ColorSpace.ocx and TestColorSpace.exe, and it aborted like it always have.  I
downloaded the test demo from:

http://www.codeguru.com/cpp/w-d/dislog/commondialogs/article.php/c1959/
 Select the "Download Demo Project" link.

You must first register the ColorSpace.ocx then run TestColorSpace.exe.  The
TestColorSpace.exe will not start, it will abort when it first references the
ColorSpace.ocx.  The abort is caused by missing ATL functions in WINE.

The test I ran was with wine-1.2-rc2.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #37 from André H.   2010-06-14 14:12:15 ---
(In reply to comment #36)
> The test I ran was with wine-1.2-rc2.

I guess AJ meant todays git, or at least rc3...

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #38 from Kenneth Robinette   2010-06-14 17:51:31 ---
(In reply to comment #37)
> (In reply to comment #36)
> > The test I ran was with wine-1.2-rc2.
>
> I guess AJ meant todays git, or at least rc3...

I just installed the latest Wine git, now when I try to install a program to
test this, I get an abort of the Installshield installer (which calls the
Microsoft msi installer).  If I go back to wine-1.2-rc2, the installer works. 
So it appears to be some type of regression between wine-1.2-rc2 and the latest
git.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #39 from Dan Kegel   2010-06-14 17:53:34 ---
Please file a new bug for the installer regression.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #40 from Kenneth Robinette   2010-06-15 08:02:33 ---
(In reply to comment #38)
> (In reply to comment #37)
> > (In reply to comment #36)
> > > The test I ran was with wine-1.2-rc2.
> >
> > I guess AJ meant todays git, or at least rc3...
>
> I just installed the latest Wine git, now when I try to install a program to
> test this, I get an abort of the Installshield installer (which calls the
> Microsoft msi installer).  If I go back to wine-1.2-rc2, the installer works. 
> So it appears to be some type of regression between wine-1.2-rc2 and the latest
> git.

Wine version "wine-1.2-rc3-43-gba9ec15" does fix the original problem, so bug
17697 is resolved.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Wylda  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |[e-mail]
         Resolution|                            |FIXED

--- Comment #41 from Wylda   2010-06-15 08:05:45 ---

Reported fixed.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

Alexandre Julliard  changed:

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

--- Comment #42 from Alexandre Julliard   2010-06-18 12:46:47 ---
Closing bugs fixed in 1.2-rc4.

[Bug 17697] SNetTerm won't start

http://bugs.winehq.org/show_bug.cgi?id=17697

--- Comment #43 from Wylda   2010-06-21 17:24:06 ---
(In reply to comment #39)
> Please file a new bug for the installer regression.

Just for info... I can't reproduce the installer problem. Tested under
wine-1.2-rc3, -rc4.