Re: compile mode and sh ( ) &

[Dan Jacobson]

>>>>> "RMS" == Richard Stallman <rms@gnu.org> writes:

RMS>     This is a problem with Emacs closing the compilation buffer and killing
RMS>     all sub-processes after the compilation process finished (It is done
RMS>     by `call-process' [C function] so it is not easily solved).

RMS> `compile' uses asynchronous processes when possible,
RMS> and uses `call-process' only as a fall-back on systens where
RMS> asynchronous processes are not supported.  What system is this,
RMS> and does it support asynchronous processes?

$ uname -a
Linux localhost.localdomain 2.2.17-21mdk #1 Thu Oct 5 13:16:08 CEST 2000 i586 
unknown
i.e. "Mandrake 7.2"

RMS> But I think that asynchronous processes may have the same
RMS> characteristic.  It is clear that Emacs closes the pipe as soon as the
RMS> first-level subprocess exits, and that would stop any further
RMS> output from second-level subprocesses.  It may be that the second-level
RMS> subprocesses are actually killed, though I have been unable to
RMS> see where that happens.

RMS> We could consider changing this, but what criterion would
RMS> Emacs use for determining that the subprocess has finished?

sounds technical.  good luck in finding the bug.
-- 
http://www.geocities.com/jidanni/ Tel+886-4-25854780