Git version, XWin dies in cygwin but not in Windows

Git version, XWin dies in cygwin but not in Windows

[Michel Hummel]

I am testing the git version of the XWin server (I don't know if it is
the good place to talk about this version) and I am experiencing a
problem (May be it is also a problem on the official Xwin).
Sometime ( I can not make a reproducible test case) when the server
stops, the Xwin process disappears from Cygwin (as expected) but the
Windows process XWin.exe still be alive.

After some investigations (I'm not a good Windows hacker) it seems
that the process hangs  on the call to PostQuitMessage (0); of the
function ddxGiveUp of the file InitOutput.c

I can't tell why (Like I said, I'm not a good Windows hacker) but my
tests seems to show that delete this call  fixes the bug (may be there
is no link ).

Is it possible that this problem lies to the fact that the main window
is destroyed before the call to PostQuitMessage  (So the WM_QUIT
message can't be treated isn't it ?)

Well It's a question more than a notice.

Thanks,
Michel Hummel

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://x.cygwin.com/docs/
FAQ:                   http://x.cygwin.com/docs/faq/

Re: Git version, XWin dies in cygwin but not in Windows

[Jon TURNEY]

On 27/09/2010 16:43, Michel Hummel wrote:
> I am testing the git version of the XWin server (I don't know if it is
> the good place to talk about this version)

This is absolutely the right place :-)

I presume by 'git version' you mean the X.Org master tree.

> and I am experiencing a
> problem (May be it is also a problem on the official Xwin).
> Sometime ( I can not make a reproducible test case) when the server
> stops, the Xwin process disappears from Cygwin (as expected) but the
> Windows process XWin.exe still be alive.

There are currently quite a few patches applied on top of the X.Org releases 
to make the cygwin released version (cygwin releases are tagged in [1]), at 
least one of which is related to stability during shutdown [5].

If you want a tree with those patches forward ported to xserver 1.9, take a 
look at [2]

Sorry that the contributors guide documentation is somewhat out of date and 
doesn't contain this information.

> After some investigations (I'm not a good Windows hacker) it seems
> that the process hangs  on the call to PostQuitMessage (0); of the
> function ddxGiveUp of the file InitOutput.c
>
> I can't tell why (Like I said, I'm not a good Windows hacker) but my
> tests seems to show that delete this call  fixes the bug (may be there
> is no link ).
>
> Is it possible that this problem lies to the fact that the main window
> is destroyed before the call to PostQuitMessage  (So the WM_QUIT
> message can't be treated isn't it ?)

That shouldn't be the case, PostMessage() [3] functions are supposed to by 
asynchronous (unlike SendMessage() [4] which is synchronous, waiting for the 
message to be processed before returning)

[1] http://cgit.freedesktop.org/~yselkowitz/xserver/
[2] http://cgit.freedesktop.org/~jturney/xserver/log/?h=cygwin-1.9-testing
[3] http://msdn.microsoft.com/en-us/library/ms644945%28VS.85%29.aspx
[4] http://msdn.microsoft.com/en-us/library/ms644950%28VS.85%29.aspx
[5] 
http://cgit.freedesktop.org/~yselkowitz/xserver/commit/?h=cygwin-release-1.8&id=9cbbc1e8aefc6111f6ccdc73c061337508061996

-- 
Jon TURNEY
Volunteer Cygwin/X X Server maintainer

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://x.cygwin.com/docs/
FAQ:                   http://x.cygwin.com/docs/faq/