lm@snafu.Sun.COM (Larry McVoy) (04/04/89)
Hi-
I'm POSIX-ifying SunOS 4.1 and in the process have run into a few
bugs in X11R3 (I hate sunview so I drag around an X tape). I need to
know where to send them and when they'll be fixed. The fix date is
medium important - other wise we have things like
if (vp == sp->s_vp && !(sp->s_flags & SESS_NOTTY)) {
if (write_access(stp, 0) == 0)
return;
}
else {
/*
* kludge for X windows. X (display manager) does
* ioctl's on these guys and shouldn't be.
*/
register dev_t dev = major(vp->v_rdev);
if (dev != MOUSE && dev != KBD) {
u.u_error = ENOTTY;
return;
}
}
in the kernel. This particular one is because X does
ioctl( , TIOCSPGRP, &arg) on the keyboard and mouse. Or
at least that seems to be what's happening.
Larry McVoy, Lachman Associates. ...!sun!lm or lm@sun.comrws@EXPO.LCS.MIT.EDU (Bob Scheifler) (04/04/89)
I drag around an X tape). I need to know where to send them
(If you drag around at MIT X tape, then you should have by now read the
release notes, which tell you where to send bugs.) Send bug reports to
xbugs@expo.lcs.mit.edu, using the form in doc/bugs/bug-report.
and when they'll be fixed.
MIT will fix bugs based on their relative priority to other tasks/bugs that
we need to get done/fixed. I can't be more specific than that. If you are
working with Sun on SunOS 4.1, then presumably you can work through your
channels to deal with Sun on their product release of X.lm@snafu.Sun.COM (Larry McVoy) (04/07/89)
I wrote: Where do I send X bugs? And Bob wrote: >(If you drag around at MIT X tape, then you should have by now read the >release notes, which tell you where to send bugs.) Send bug reports to >xbugs@expo.lcs.mit.edu, using the form in doc/bugs/bug-report. >MIT will fix bugs based on their relative priority to other tasks/bugs that >we need to get done/fixed. I can't be more specific than that. If you are >working with Sun on SunOS 4.1, then presumably you can work through your >channels to deal with Sun on their product release of X. Silly me, I suppose I should read the release notes (still haven't and X runs file :-) Anyway, I just wanted people to know that X was not the problem - the problem was that fcntl(fd, F_SETOWN, getpid()) was implemented as ioctl(fd, TIOCSPGRP, &mypid) and as such fell victim to stricter POSIX checking than it used to. In particular, POSIX refused to believe that the mouse and kbd were controlling tty's (can't imagine why :-) The bug has been fixed, X is redeemed, and all is wonderful. Thanks to all who commented on this, Larry McVoy, Lachman Associates. ...!sun!lm or lm@sun.com