Re: src/winsup/cygwin ChangeLog thread.cc thread.h ...

["Robert Collins" <robert dot collins at itdomain dot com dot au>]

The pthread changes work without trouble, without the patch below...

Mon Sep 24 17:41:03 2001  Christopher Faylor <cgf@redhat.com>

        * fhandler.h (fhandler_pipe::hit_eof): New method.
        (writepipe_exists): New class element.
        (orig_pid): Ditto.
        (id): Ditto.
        (is_slow): Eliminate.
        * pipe.cc (fhandler_pipe::set_close_on_exec): Set inheritance on
        writepipe_exists, if it exists.
        (fhandler_pipe::hit_eof): New method, modelled after tty.
        (fhandler_pipe::dup): Duplicate writepipe_exists, if it exists.
        (make_pipe): Set up a dummy event for pipes on windows 9x.  The
        nonexistence of this event means that the write side of the
        pipe has closed.
        (_dup): Move to syscalls.cc
        (_dup2): Ditto.

        * dtable.cc (dtable::build_fhandler): Fill out set_names here,
if
        appropriate.
        * syscalls.cc (_open): Call set_names in build_fhandler.

And I note that more files were committed than listed in the
changelog?!?
(see http://www.cygwin.com/ml/cygwin-cvs/2001-q3/msg00264.html).

-/dtable.cc/1.54/Fri Sep 28 01:02:29 2001/-kk/
+/dtable.cc/1.53/Fri Sep 28 00:24:48 2001/-kk/

-/fhandler.h/1.83/Fri Sep 28 01:02:29 2001/-kk/
+/fhandler.h/1.82/Fri Sep 28 00:24:49 2001/-kk/

-/fhandler_tty.cc/1.49/Fri Sep 28 01:02:29 2001/-kk/
+/fhandler_tty.cc/1.48/Fri Sep 28 00:24:49 2001/-kk/

-/how-vfork-works.txt/1.3/Fri Sep 28 01:02:29 2001/-kk/
+/how-vfork-works.txt/1.2/Fri Sep 28 00:24:49 2001/-kk/

-/pipe.cc/1.19/Fri Sep 28 01:02:29 2001/-kk/
+/pipe.cc/1.18/Fri Sep 28 00:24:49 2001/-kk/

-/select.cc/1.47/Fri Sep 28 01:02:30 2001/-kk/
+/select.cc/1.46/Fri Sep 28 00:24:49 2001/-kk/

-/syscalls.cc/1.149/Fri Sep 28 01:02:30 2001/-kk/
+/syscalls.cc/1.148/Fri Sep 28 00:24:49 2001/-kk/


----- Original Message -----
From: "egor duda" <deo@logos-m.ru>
To: "Christopher Faylor" <cygwin-patches@cygwin.com>
Sent: Friday, September 28, 2001 4:08 AM
Subject: Re: src/winsup/cygwin ChangeLog thread.cc thread.h ...


> Hi!
>
> Thursday, 27 September, 2001 Christopher Faylor cgf@redhat.com wrote:
>
> >>I'm having some trouble with cvs+ssh with this patch .. though I'm
not
> >>sure why. For a little while I though it might be chris's tuesday
> >>sleep(1) change, because I was getting strange results from pspec>
I'm
> >>not sure though.
>
> CF> Huh?  What is my "sleep(1)" change?  The only change I made on
Tuesday was
> CF> to fhandler_tty_common::ready_for_read.  How would that affect
cvs?
>
> it does. i've seen the following at the end of strace (where ssh
> seemed to block in ReadFile() on empty pipe)
>
>   517 12076688 [main] ssh 7436 peek_socket: considering handle 0x24
>   242 12076930 [main] ssh 7436 peek_socket: adding read fd_set
/dev/tcp, fd 4
>   235 12077165 [main] ssh 7436 peek_socket: adding write fd_set
/dev/tcp, fd 4
>   291 12077456 [main] ssh 7436 peek_socket: WINSOCK_SELECT returned 1
>   434 12077890 [main] ssh 7436 set_bits: me 0xA013CC0, testing fd 4
(/dev/tcp)
>   281 12078171 [main] ssh 7436 set_bits: ready 1
>   242 12078413 [main] ssh 7436 peek_pipe: already ready
>   236 12078649 [main] ssh 7436 set_bits: me 0xA013C78, testing fd 0
(/dev/piper)
>   241 12078890 [main] ssh 7436 set_bits: ready 1
>   233 12079123 [main] ssh 7436 select_stuff::poll: returning 2
>   242 12079365 [main] ssh 7436 select_stuff::cleanup: calling cleanup
routines
>   248 12079613 [main] ssh 7436 select_stuff::~select_stuff: deleting
select records
>   694 12080307 [main] ssh 7436 _read: read (0, 0x24AD7C4, 8192)
blocking, sigcatchers 3
>   320 12080627 [main] ssh 7436 peek_pipe: /dev/piper, saw EOF
>   261 12080888 [main] ssh 7436 peek_pipe: saw eof on '/dev/piper'
>   238 12081126 [main] ssh 7436 fhandler_pipe::ready_for_read:
returning 1
>
> Egor.            mailto:deo@logos-m.ru ICQ 5165414 FidoNet
2:5020/496.19
>
>