Re: New maintainer for EMX?

On Thu, Jun 19, 2003 at 11:25:48AM +0200, Eberhard Mattes wrote:
> Ilya Zakharevich wrote:
> 
> >  f) a fix for socket handles not being inherited by childs;
> 
> I've never seen a bug report for this.

Why would anyone send a bug report when this is somewhere in TODO
lists?  ;-)

> > I also may be able to fix a couple of bugs in sprintf(%g) I know.

> Ditto.

Retracted: I was running with a debugging EMX dll's, and my build of
these DLLs was obsolete.  With the current EMX DLLs the perl test
suite (of the maintainance brach) finds no bugs in sprintf().  (It was
sprintf("%0g", 0.1) or some such.)

BTW, the only failing subtest on my computer is a failure of wait().
[sorry for being a little bit fuzzy now, I'm not on OS/2 computer now]
One starts a kid asyncroneously; kid kills itself with sig=3.  The
parent's wait() gets 3, but in a wrong part of the result short - not
in the signal slot, but in retval slot.

E.g., on Unix

 perl -wle 'system qw(perl -wle), q(kill 3, $$); printf q(%#x), $?'

prints 0x3.  IIRC, on OS/2 it prints 0x300 (system() is asyncroneous
on OS/2 to account for childs which do not intercept Ctrl-C).

This return value of wait() is contrary to EMX docs.  Should it be
expected?

Thanks,
Ilya