GNU bug report logs - #21617
Emacs compilation mode prevents child processes from spawning/forking daemons

Previous Next

To reply to this bug, email your comments to 21617 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Paul Pogonyshev <pogonyshev <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Emacs compilation mode prevents child processes from spawning/forking
 daemons
Date: Mon, 5 Oct 2015 10:58:33 +0200

[Message part 1 (text/plain, inline)]

This can be reproduced with Gradle, but you don't need a Gradle project for
this. When running from a normal console:

    $ gradle
    Starting a new Gradle Daemon for this build (subsequent builds will be
faster).

    :help
    ...
    Total time: 2.014 secs

Now if I repeat it, the daemon is reused, so the first message is gone and
the "build" goes much faster:

    $ gradle
    :help
    ...
    Total time: 0.495 secs

By comparison, when run from Emacs M-x compile, Gradle starts daemon every
time, so all builds look like the first example above. However, if daemon
already exists (e.g. created by console run above), it will not be killed.

File '~/.gradle/daemon/VERSION/daemon-PID.out.log' for daemons spawned
after M-x compile ends with this line:

    Daemon vm is shutting down... The daemon has exited normally or was
terminated in response to a user interrupt.

When daemon is allowed to exist (e.g. when run from a normal console), log
receives this message instead:

    ... Daemon is idle, sleeping until state change or idle timeout at ...

Additionally, I checked several other ways to run process from Emacs:

    * M-x eshell: daemon is killed, as with M-x compile
    * M-x shell: daemon works properly
    * M-!: daemon works properly

So, it seems that Emacs has a way to keep daemons alive / not kill them,
but it is not used in compilation mode.

Paul

[Message part 2 (text/html, inline)]

Message #8 received at 21617 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Paul Pogonyshev <pogonyshev <at> gmail.com>
Cc: 21617 <at> debbugs.gnu.org
Subject: Re: bug#21617: Emacs compilation mode prevents child processes
 from	spawning/forking daemons
Date: Mon, 05 Oct 2015 12:25:36 +0300

> Date: Mon, 5 Oct 2015 10:58:33 +0200
> From: Paul Pogonyshev <pogonyshev <at> gmail.com>
> 
> By comparison, when run from Emacs M-x compile, Gradle starts daemon every
> time, so all builds look like the first example above. However, if daemon
> already exists (e.g. created by console run above), it will not be killed.
> 
> File '~/.gradle/daemon/VERSION/daemon-PID.out.log' for daemons spawned after
> M-x compile ends with this line:
> 
> Daemon vm is shutting down... The daemon has exited normally or was terminated
> in response to a user interrupt.
> 
> When daemon is allowed to exist (e.g. when run from a normal console), log
> receives this message instead:
> 
> ... Daemon is idle, sleeping until state change or idle timeout at ...
> 
> Additionally, I checked several other ways to run process from Emacs:
> 
> * M-x eshell: daemon is killed, as with M-x compile
> * M-x shell: daemon works properly
> * M-!: daemon works properly
> 
> So, it seems that Emacs has a way to keep daemons alive / not kill them, but it
> is not used in compilation mode.

I just tried

  M-x compile RET C-a C-k emacs -Q -daemon RET

and that started a daemon which stayed up in the air and didn't exit
or get killed.  If you see the same on your system, then Gradle,
whatever that is, is indeed somehow involved in this, because without
it everything works as you expected.

Also, you didn't tell which version of Emacs you used (I used the
latest development sources).

Message #11 received at 21617 <at> debbugs.gnu.org (full text, mbox):

From: Paul Pogonyshev <pogonyshev <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 21617 <at> debbugs.gnu.org
Subject: Re: bug#21617: Emacs compilation mode prevents child processes from
 spawning/forking daemons
Date: Mon, 5 Oct 2015 11:43:48 +0200

[Message part 1 (text/plain, inline)]

> I just tried
>
>   M-x compile RET C-a C-k emacs -Q -daemon RET
>
> and that started a daemon which stayed up in the air and didn't exit
> or get killed.

Yeah, it also works here. I can only guess that Gradle daemon is terminated
if
stdout is closed before the main process is shut down or something like
that.

> If you see the same on your system, then Gradle,
> whatever that is, is indeed somehow involved in this, because without
> it everything works as you expected.

It's a pretty standard build tool (Make of sorts) in a Java world. They
will of
course tell me (for the second time, I already asked for support for this
problem
before) it's an Emacs' fault, because outside it Gradle works fine.

> Also, you didn't tell which version of Emacs you used (I used the
> latest development sources).

GNU Emacs 25.0.50.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.10.9) of
2015-08-27,
a fairly recent trunk.

Paul

[Message part 2 (text/html, inline)]

Message #14 received at 21617 <at> debbugs.gnu.org (full text, mbox):

From: Noam Postavsky <npostavs <at> users.sourceforge.net>
To: 21617 <at> debbugs.gnu.org
Cc: Eli Zaretskii <eliz <at> gnu.org>, Paul Pogonyshev <pogonyshev <at> gmail.com>
Subject: Re: bug#21617: Emacs compilation mode prevents child processes from
 spawning/forking daemons
Date: Tue, 14 Jun 2016 17:51:39 -0400

This sounds like http://debbugs.gnu.org/cgi/bugreport.cgi?bug=21866
and http://debbugs.gnu.org/cgi/bugreport.cgi?bug=16019, Emacs sends
SIGHUP to all children of pty subprocesses. The "user interrupt"
mentioned above could be the SIGHUP.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.