snapshot 2013-01-23

snapshot 2013-01-23

[Achim Gratz]

I'm seeing strange behaviour from lftp scripts that use external commands; lftp
forks, but then hangs and needs to be killed.  This is only happening from
mintty, but not when lftp gets started from CMD:

sh -c "PATH=/bin lftp -c '\!echo bla'"

Tracing lftp in mintty yields the following after the fork:

   44  264494 [main] lftp 5764 sync_with_parent: signalling parent: performed
fork fixup
   26  264520 [main] lftp 5764 child_info::ready: signalled 0x250 that I was ready
  378  264522 [main] lftp 2924 child_info::sync: pid 5764, WFMO returned 0,
exit_code 0x103, res 1
   49  264571 [main] lftp 2924 sig_send: sendsig 0x8C, pid 2924, signal -41,
its_me 1
   28  264599 [main] lftp 2924 sig_send: wakeup 0x264
   35  264634 [main] lftp 2924 sig_send: Waiting for pack.wakeup 0x264
    5  264639 [sig] lftp 2924 wait_sig: signalling pack.wakeup 0x264
   45  264684 [main] lftp 2924 sig_send: returning 0x0 from sending signal -41
  207  264891 [main] lftp 2924 fork: 5764 = fork()
   32  264923 [main] lftp 2924 wait4: calling proc_subproc, pid 5764, options 2
   27  264950 [main] lftp 2924 proc_subproc: args: 6, 2284032
   25  264975 [main] lftp 2924 proc_subproc: wval->pid 5764, wval->options 2
   24  264999 [main] lftp 2924 checkstate: nprocs 1
   26  265025 [main] lftp 2924 stopped_or_terminated: considering pid 5764, pgid
3108, w->pid 5764
   22  265047 [main] lftp 2924 checkstate: no matching terminated children found
   27  265074 [main] lftp 2924 checkstate: returning -1
   25  265099 [main] lftp 2924 proc_subproc: only found non-terminated children
   24  265123 [main] lftp 2924 proc_subproc: finished processing
terminated/stopped child
   24  265147 [main] lftp 2924 proc_subproc: returning 1
 1514  266034 [main] lftp 5764 sig_send: sendsig 0x90, pid 5764, signal -41,
its_me 1
   27  266061 [main] lftp 5764 sig_send: wakeup 0x244
   29  266090 [main] lftp 5764 sig_send: Waiting for pack.wakeup 0x244
   25  266115 [sig] lftp 5764 wait_sig: signalling pack.wakeup 0x244
   30  266145 [main] lftp 5764 sig_send: returning 0x0 from sending signal -41
  177  266322 [main] lftp 5764 fork: 0 = fork()
   38  266360 [main] lftp 5764 getpid: 5764 = getpid()
   29  266389 [main] lftp 5764 setpgid: pid 5764, pgid 5764, res 0
   26  266415 [main] lftp 5764 getpid: 5764 = getpid()
   24  266439 [main] lftp 5764 kill0: kill (5764, 17)
   26  266465 [main] lftp 5764 sig_send: sendsig 0x90, pid 5764, signal 17, its_me 1
   25  266490 [main] lftp 5764 sig_send: wakeup 0x254
   29  266519 [main] lftp 5764 sig_send: Waiting for pack.wakeup 0x254
   31  266550 [sig] lftp 5764 fhandler_pty_slave::tcflush: tcflush(0) handle 0x14C
   30  266580 [sig] lftp 5764 fhandler_termios::bg_check: /dev/pty0, bg I/O pgid
5764, tpgid 3108, myctty ctty /dev/pty0
   25  266605 [sig] lftp 5764 tty_min::is_orphaned_process_group: checking pgid
5764, my sid -1, my parent 2924
  120  266725 [sig] lftp 5764 open_shared: name cygpid.3460, n 3460, shared
0x3D0000 (wanted 0x0), h 0x244, *m 6
   52  266777 [sig] lftp 5764 open_shared: name cygpid.5664, n 5664, shared
0x3F0000 (wanted 0x0), h 0x258, *m 6
   48  266825 [sig] lftp 5764 open_shared: name cygpid.2924, n 2924, shared
0x4E0000 (wanted 0x0), h 0x268, *m 6
   56  266881 [sig] lftp 5764 open_shared: name cygpid.5696, n 5696, shared
0x4F0000 (wanted 0x0), h 0x274, *m 6
   42  266923 [sig] lftp 5764 open_shared: name cygpid.2340, n 2340, shared
0x500000 (wanted 0x0), h 0x27C, *m 6
   26  266949 [sig] lftp 5764 pinfo::init: execed process windows pid 2340,
cygwin pid 3460
   49  266998 [sig] lftp 5764 open_shared: name cygpid.3460, n 3460, shared
0x500000 (wanted 0x0), h 0x27C, *m 6
   60  267058 [sig] lftp 5764 open_shared: name cygpid.4168, n 4168, shared
0x500000 (wanted 0x0), h 0x27C, *m 6
   25  267083 [sig] lftp 5764 __set_errno: void pinfo::init(pid_t, DWORD,
void*):314 setting errno 3
   42  267125 [sig] lftp 5764 open_shared: name cygpid.3108, n 3108, shared
0x500000 (wanted 0x0), h 0x27C, *m 6
   46  267171 [sig] lftp 5764 open_shared: name cygpid.5400, n 5400, shared
0x510000 (wanted 0x0), h 0x284, *m 6
   36  267207 [sig] lftp 5764 tty_min::is_orphaned_process_group: checking pid
3460 - has pgid 3460
   24  267231 [sig] lftp 5764 tty_min::is_orphaned_process_group: checking pid
5664 - has pgid 0
   23  267254 [sig] lftp 5764 tty_min::is_orphaned_process_group: checking pid
2924 - has pgid 3108
   21  267275 [sig] lftp 5764 tty_min::is_orphaned_process_group: checking pid
5764 - has pgid 5764
   30  267305 [sig] lftp 5764 open_shared: name cygpid.2924, n 2924, shared
0x520000 (wanted 0x0), h 0x288, *m 6
   28  267333 [sig] lftp 5764 tty_min::is_orphaned_process_group: ppid->pgid
3108, ppid->sid -1
  117  267450 [sig] lftp 5764 kill_pgrp: pid 5764, signal 21
   74  267524 [sig] lftp 5764 open_shared: name cygpid.3460, n 3460, shared
0x3D0000 (wanted 0x0), h 0x27C, *m 6
   42  267566 [sig] lftp 5764 open_shared: name cygpid.5664, n 5664, shared
0x3F0000 (wanted 0x0), h 0x274, *m 6
   44  267610 [sig] lftp 5764 open_shared: name cygpid.2924, n 2924, shared
0x4E0000 (wanted 0x0), h 0x268, *m 6
   47  267657 [sig] lftp 5764 open_shared: name cygpid.5696, n 5696, shared
0x4F0000 (wanted 0x0), h 0x258, *m 6
   43  267700 [sig] lftp 5764 open_shared: name cygpid.2340, n 2340, shared
0x500000 (wanted 0x0), h 0x244, *m 6
   25  267725 [sig] lftp 5764 pinfo::init: execed process windows pid 2340,
cygwin pid 3460
   33  267758 [sig] lftp 5764 open_shared: name cygpid.3460, n 3460, shared
0x500000 (wanted 0x0), h 0x244, *m 6
   53  267811 [sig] lftp 5764 open_shared: name cygpid.4168, n 4168, shared
0x500000 (wanted 0x0), h 0x244, *m 6
   25  267836 [sig] lftp 5764 __set_errno: void pinfo::init(pid_t, DWORD,
void*):314 setting errno 3
   33  267869 [sig] lftp 5764 open_shared: name cygpid.4168, n 4168, shared
0x500000 (wanted 0x0), h 0x244, *m 6
   25  267894 [sig] lftp 5764 __set_errno: void pinfo::init(pid_t, DWORD,
void*):314 setting errno 3
   39  267933 [sig] lftp 5764 open_shared: name cygpid.3108, n 3108, shared
0x500000 (wanted 0x0), h 0x244, *m 6
   43  267976 [sig] lftp 5764 open_shared: name cygpid.5400, n 5400, shared
0x510000 (wanted 0x0), h 0x28C, *m 6
   38  268014 [sig] lftp 5764 kill_pgrp: killing pid 5764, pgrp 5764, p->ctty
/dev/pty0, ctty /dev/pty0
   28  268042 [sig] lftp 5764 sig_send: sendsig 0x90, pid 5764, signal 21, its_me 1
   23  268065 [sig] lftp 5764 sig_send: wakeup 0x290
   24  268089 [sig] lftp 5764 sig_send: Waiting for pack.wakeup 0x290
28185096 28450243 [sig] lftp 2924 fhandler_pty_slave::tcflush: tcflush(0) handle
0x14C
   63 28450306 [sig] lftp 2924 fhandler_pty_slave::read: read(0, -1) handle 0x14C
  123 28450429 [sig] lftp 2924 fhandler_pty_slave::read: wait timed out,
time_to_wait 0
   24 28450453 [sig] lftp 2924 fhandler_pty_slave::read: -1=read(0, -1)
   19 28450472 [sig] lftp 2924 fhandler_pty_slave::tcflush: -1=tcflush(0)
   20 28450492 [sig] lftp 2924 sigpacket::process: signal 2 processing
   21 28450513 [sig] lftp 2924 init_cygheap::find_tls: sig 2
   20 28450533 [sig] lftp 2924 sigpacket::process: using tls 0x22CE64
  370 28450903 [sig] lftp 2924 sigpacket::process: signal 2, signal handler
0x6102FE00
   26 28450929 [sig] lftp 2924 sigpacket::setup_handler: controlled interrupt.
stackptr 0x22DA34, stack 0x22DA30, stackptr[-1] 0x41B306
   29 28450958 [sig] lftp 2924 proc_subproc: args: 5, 1
   23 28450981 [sig] lftp 2924 proc_subproc: clear waiting threads
   21 28451002 [sig] lftp 2924 checkstate: nprocs 1
   21 28451023 [sig] lftp 2924 stopped_or_terminated: considering pid 5764, pgid
5764, w->pid 5764
   22 28451045 [sig] lftp 2924 checkstate: no matching terminated children found
   21 28451066 [sig] lftp 2924 checkstate: returning -1
   21 28451087 [sig] lftp 2924 proc_subproc: waiting thread found no children
   22 28451109 [sig] lftp 2924 proc_subproc: finished clearing
   21 28451130 [sig] lftp 2924 proc_subproc: returning 1
  -16 28451114 [main] lftp 2924 wait4: 0 = cygwait (...)
   36 28451150 [sig] lftp 2924 _cygtls::interrupt_setup: armed signal_arrived
0x28C, signal 2
   39 28451189 [sig] lftp 2924 sigpacket::setup_handler: signal 2 delivered
   21 28451210 [sig] lftp 2924 sigpacket::process: returning 1
    3 28451213 [main] lftp 2924 set_process_mask_delta: oldmask 0x80000, newmask
0x80000, deltamask 0x0
   42 28451255 [main] lftp 2924 signal_exit: exiting due to signal 2
   24 28451279 [main] lftp 2924 signal_exit: about to call do_exit (2)
   23 28451302 [main] lftp 2924 do_exit: do_exit (2), exit_state 2
   23 28451325 [main] lftp 2924 void: 0x4451B0 = signal (20, 0x1)
   22 28451347 [main] lftp 2924 void: 0x43BC30 = signal (1, 0x1)
   21 28451368 [main] lftp 2924 void: 0x0 = signal (2, 0x1)
   21 28451389 [main] lftp 2924 void: 0x0 = signal (3, 0x1)
   28 28451417 [main] lftp 2924 fhandler_pty_slave::cleanup: /dev/pty0 closed,
usecount 4
   24 28451441 [main] lftp 2924 fhandler_base::close_with_arch: line 1125: 
/dev/pty0<0x61275368> usecount + -1 = 3
   23 28451464 [main] lftp 2924 fhandler_base::close_with_arch: not closing
archetype
   24 28451488 [main] lftp 2924 fhandler_pty_slave::cleanup: /dev/pty0 closed,
usecount 3
   22 28451510 [main] lftp 2924 fhandler_base::close_with_arch: line 1125: 
/dev/pty0<0x61275368> usecount + -1 = 2
   22 28451532 [main] lftp 2924 fhandler_base::close_with_arch: not closing
archetype
   26 28451558 [main] lftp 2924 fhandler_pty_slave::cleanup: /dev/pty0 closed,
usecount 2
   22 28451580 [main] lftp 2924 fhandler_base::close_with_arch: line 1125: 
/dev/pty0<0x61275368> usecount + -1 = 1
   22 28451602 [main] lftp 2924 fhandler_base::close_with_arch: not closing
archetype
   25 28451627 [main] lftp 2924 fhandler_base::close: closing
'/mnt/upload/mirror' handle 0x240
   26 28451653 [main] lftp 2924 fhandler_base::close: closing
'/mnt/upload/mirror/test.lftp' handle 0x24C
   25 28451678 [main] lftp 2924 init_cygheap::close_ctty: closing cygheap->ctty
0x61275368
   22 28451700 [main] lftp 2924 fhandler_base::close_with_arch: closing passed
in archetype, usecount 0
   23 28451723 [main] lftp 2924 fhandler_pty_slave::close: closing last open
/dev/pty0 handle
  149 28451872 [main] lftp 2924 fhandler_console::free_console: freed console, res 1
   27 28451899 [main] lftp 2924 fhandler_pty_common::close: pty0 <0x14C,0x154>
closing
   28 28451927 [main] lftp 2924 getpid: 2924 = getpid()
   29 28451956 [main] lftp 2924 proc_terminate: nprocs 1
   23 28451979 [main] lftp 2924 proc_subproc: args: 5, 1
   23 28452002 [main] lftp 2924 proc_subproc: clear waiting threads
   20 28452022 [main] lftp 2924 proc_subproc: finished clearing
   21 28452043 [main] lftp 2924 proc_subproc: returning 1
   23 28452066 [main] lftp 2924 cygthread::terminate_thread: thread 'waitproc',
id 0x758, inuse 1, stack_ptr 0x1F0AD40
  108 28452174 [main] lftp 2924 proc_terminate: leaving
   47 28452221 [main] lftp 2924 __to_clock_t: dwHighDateTime 0, dwLowDateTime 468003
   23 28452244 [main] lftp 2924 __to_clock_t: total 00000000 0000002E
   20 28452264 [main] lftp 2924 __to_clock_t: dwHighDateTime 0, dwLowDateTime 780005
   21 28452285 [main] lftp 2924 __to_clock_t: total 00000000 0000004E
   21 28452306 [main] lftp 2924 pinfo::exit: Calling ExitProcess n 0x2, exitcode
0x200



Regards,
Achim.


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

Re: snapshot 2013-01-23

[Achim Gratz]

Achim Gratz <Stromeko <at> NexGo.DE> writes:
> I'm seeing strange behaviour from lftp scripts that use external commands

Git apparently has similar problems when it is started from Emacs (via magit).


Regards,
Achim.


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

Re: snapshot 2013-01-23

[Achim Gratz]

Achim Gratz <Stromeko <at> NexGo.DE> writes:
> I'm seeing strange behaviour from lftp scripts that use external commands;

The problem with lftp goes away when I roll back to the snapshot 2013-01-07, the
one from 2013-01-18 is also bad.  I did not test any in-between snapshots.  I
don't know yet if the problem with Git was related or not.


Regards,
Achim.


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

Re: snapshot 2013-01-23

[Achim Gratz]

Achim Gratz <Stromeko <at> NexGo.DE> writes:
> Git apparently has similar problems when it is started from Emacs (via magit).

Part of the problem has been solved by rolling back to the 2013-01-07 snapshot.

I still get Git to hang up when doing a pull that results in a forced update on
a worktree that is mounted with "noacl" option.  Interestingly, when I do:

git gc --aggressive
git remote update
git pull

instead of simply:

git pull

then everything seems to work.  I do not have that problem with the same
worktree when instantiated on a volume where "noacl" does not have to be used,
so somehow it must have something to do with Git waiting for the results of a
chmod or similar change to the FS becoming visible (which then never happens
because the access rights are fake).


Regards,
Achim.


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

Re: snapshot 2013-01-23

[Achim Gratz]

Achim Gratz <Stromeko <at> NexGo.DE> writes:
> I do not have that problem with the same
> worktree when instantiated on a volume where "noacl" does not have to be used,
> so somehow it must have something to do with Git waiting for the results of a
> chmod or similar change to the FS becoming visible (which then never happens
> because the access rights are fake).

Based on that observation, I configured that Git repo to ignore filemodes
altogether (filemode=false) and everything seems to work as it should (still on
snapshot 2013-01-07, though).
 
Regards,
Achim.




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

Re: snapshot 2013-01-23

[Christopher Faylor]

On Mon, Jan 28, 2013 at 01:58:44PM +0000, Achim Gratz wrote:
>I'm seeing strange behaviour from lftp scripts that use external commands; lftp
>forks, but then hangs and needs to be killed.  This is only happening from
>mintty, but not when lftp gets started from CMD:
>
>sh -c "PATH=/bin lftp -c '\!echo bla'"
>
>Tracing lftp in mintty yields the following after the fork:

Please don't post strace data unless it is solicited.  A reproducer
is usually sufficient.

This particular problem should be fixed in the next snapshot.

cgf

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

Re: snapshot 2013-01-23

[Christopher Faylor]

On Wed, Jan 30, 2013 at 11:28:17AM +0000, Achim Gratz wrote:
>Achim Gratz <Stromeko <at> NexGo.DE> writes:
>> I do not have that problem with the same
>> worktree when instantiated on a volume where "noacl" does not have to be used,
>> so somehow it must have something to do with Git waiting for the results of a
>> chmod or similar change to the FS becoming visible (which then never happens
>> because the access rights are fake).
>
>Based on that observation, I configured that Git repo to ignore filemodes
>altogether (filemode=false) and everything seems to work as it should (still on
>snapshot 2013-01-07, though).

I can't duplicate this with a simple "git pull".  Could you provide the
exact sequence of steps needed to duplicate the problem?

cgf

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

Re: snapshot 2013-01-23

[Achim Gratz]

Christopher Faylor <cgf-use-the-mailinglist-please <at> cygwin.com> writes:
> I can't duplicate this with a simple "git pull".  Could you provide the
> exact sequence of steps needed to duplicate the problem?

In origin, roll back and then advance HEAD (by rewriting or amending some
commits), then try to pull this in the downstream repo so that it needs to do a
forced update.  The downstream repo is on an FS that I have to mount with
"noacl" because I am not allowed to change any ACL there:

//server/repo on /mnt/repo type netapp (binary,noacl,user,bind)

The hang apparently happens when the HEAD rewind is about to happen.

Originally I couldn't even read the ACL on that volume, but it seems that this
has been changed that I complained about it.  IT did tell me they'd do nothing,
but it looks they changed their mind, without telling me of course.  This change
must have happened quite recently and it is quite possible that this was when
the failures started (which did not happen again since I've changed the Git
config to include filemode=false for this repo).

Anyway, I now think it's not related to any snapshot, even if it might warrant
further investigation.


Regards,
Achim.


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

Re: snapshot 2013-01-23

[Achim Gratz]

Christopher Faylor <cgf-use-the-mailinglist-please <at> cygwin.com> writes:
> This particular problem should be fixed in the next snapshot.

Fix confirmed.


Regards,
Achim.




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