Conformance tests fail with 20031016 build

[Paul Millar]

Hi Sylvain,

On Thu, 30 Oct 2003, [iso-8859-1] Sylvain Petreolle wrote:
> > What settings are you using for wine?  I don't see this problem with
> > WRT, but I've only recently got it back up and running again.
> What are the settings you want to know ?

~/.wine/config would be good.


> > On a seperate note, some other strange things I've noticed:
> > 
> > One of Monday night's patches seems to have broken a conformance test
> > in listbox.c (line 123) but only for wine in win3.1-mode!  Win3.1-mode
> > also passes two todo tests (lines 129 and 134 both in the 
> > user/listbox.c).
> Got same results here. Every emulated OS passes except win31.

Good!  I'm glad its not just me ...

> > Also, does anyone else get orphaned zombie processes after running
> > the regression tests?  (which is moderately impressive: I thought init
> > was supposed to clean these up)
> 
> I get some zombies too, and the kernel notices it. This is what I have
> in /var/log/messages:
> Oct 30 10:44:18 wine kernel: application bug: wine(2120) has SIGCHLD
> set to SIG_IGN but calls wait().
> Oct 30 10:44:18 wine kernel: (see the NOTES section of 'man 2 wait').
> Workaround activated.

That's interesting.  My syslog doesn't have any message like that.  What
kernel are you running?  I'm using 2.6.0-test8 on quisquiliae.


> Talking about the kernel process test, it seems they are runned more
> than once, is it the normal way ?
> ../../../tools/runtest -q -P wine -M kernel32.dll -T ../../.. -p kernel32_test.exe.so process.c && touch process.ok
> tests/process.c: 1 tests executed, 0 marked as todo, 0 failures.
[12 idential lines follow]

Yep, I get that (except I get 14 lines in total) when I take the -q option
out.

Cheers,

Paul.

----
Paul Millar