Recent SVN annoyances...

[Alex Ionescu]

Hi,

Some things that need to be fixed ASAP:

1) Gunnar's additions to ntdll's critical.c break cmd.exe. This is not 
his fault per-se, since his checks are legitimate. It seems cmd.exe is 
not intiailizing the CS before using it. Although I understand that his 
changes are needed, I would prefer devs to actually test their changes 
and fix the broken apps, and not just break them, since this is a 
regression in a way (much like Thomas's pseh fixes).

2) Thomas' recent patches flood debug output with an error about 
NtGdiCombineRgn. Once again, please either fix the broken apps which do 
this, or show the message only once. It makes debugging impossible.

Best regards,
Alex Ionescu

[Thomas Weidenmueller]

Alex Ionescu wrote:
> Thomas' recent patches flood debug output with an error about
>
> NtGdiCombineRgn. Once again, please either fix the broken apps which 
> do this, or show the message only once. It makes debugging impossible.

should be fixed.

Regards
Thomas

[Gunnar Dalsnes]

Alex Ionescu wrote:
> Hi,
>
> Some things that need to be fixed ASAP:
> Gunnar's additions to ntdll's critical.c break cmd.exe. This is not

I tested it (compile ros-on-roa) and it works fine for me.

[Thomas Weidenmueller]

Gunnar Dalsnes wrote:
> Alex Ionescu wrote:
> > Hi,
> >
> > Some things that need to be fixed ASAP:
> > Gunnar's additions to ntdll's critical.c break cmd.exe. This is not
>
> I tested it (compile ros-on-roa) and it works fine for me.

Fixed it already. Failing when OwnerThread is NULL in 
RtlEnterCriticalSection() is wrong and the message shouldn't be 
displayed either. I added a detailed comment on why it is wrong.

Regards,
Thomas

[Alex Ionescu]

Gunnar Dalsnes wrote:
> Alex Ionescu wrote:
> > Hi,
> >
> > Some things that need to be fixed ASAP:
> > Gunnar's additions to ntdll's critical.c break cmd.exe. This is not
>
> I tested it (compile ros-on-roa) and it works fine for me.
> _______________________________________________

If 10000000 messages scrolling by and returning STATUS_INVALID_PARAMETER 
to every CS acquire/release is "working fine", then yes, it's "working 
fine". Fact of the matter is that this is totally unacceptable and means 
cmd.exe is not initializing the CS.

Best regards,
Alex Ionescu

[Alex Ionescu]

Critical Sections are fixed again as are Region Combines.

Thank you Thomas.

Best regards,
Alex Ionescu

[Gunnar Dalsnes]

Alex Ionescu wrote:
> Gunnar Dalsnes wrote:
> > Alex Ionescu wrote:
> > > Hi,
> > >
> > > Some things that need to be fixed ASAP:
> > > Gunnar's additions to ntdll's critical.c break cmd.exe. This is not
> >
> > I tested it (compile ros-on-roa) and it works fine for me.
> > _______________________________________________
>
> If 10000000 messages scrolling by and returning STATUS_INVALID_PARAMETER 
> to every CS acquire/release is "working fine", then yes, it's "working 
> fine". Fact of the matter is that this is totally unacceptable and means 
> cmd.exe is not initializing the CS.

I dont get any such messages.
> Best regards,
> Alex Ionescu
> _______________________________________________
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev