[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 2/3] Don't print "process finished" into the stderr buffer.

From: Eli Zaretskii
Subject: Re: [PATCH 2/3] Don't print "process finished" into the stderr buffer.
Date: Tue, 10 Apr 2018 07:59:05 +0300
User-agent: K-9 Mail for Android

On April 10, 2018 7:14:39 AM GMT+03:00, Stephen Leake <address@hidden> wrote:
> Eli Zaretskii <address@hidden> writes:
> >> From: Philipp Stephani <address@hidden>
> >> Date: Sat, 07 Apr 2018 20:21:40 +0000
> >> Cc: address@hidden, address@hidden
> >> 
> >>  > * src/process.c (syms_of_process): Add 'ignore' symbol.
> >>  > (Fmake_process): Use it as sentinel for the standard error pipe
> >>  > process.
> >> 
> >>  Why would we want that, and by default on top of that?  Please
> give at
> >>  least some rationale behind this change.
> >> 
> >> Neither the manual not the docstring for `make-process' specify
> that
> >> Emacs prints "Process foo stderr
> >> finished" at the end of the standard error buffer, so that message
> >> shouldn't be printed.
> >
> > That assumes that the documentation is correct and the code isn't;
> it
> > could be the other way around.
> >
> > But I'm guessing that the current behavior was unexpected for some
> > reason, and that's why you looked in the documentation.  If the
> guess
> > is correct, could you describe why it was unexpected/unwanted?
> I had to work around that behavior as well, in DVC (Emacs front end
> for
> some CM tools). When you are parsing the output of a process, that
> string is unexpected.

Are we talking about the same thing?  Inserting the exit status into the process
buffer by the default sentinel is veteran Emacs behavior; if you don't want 
you are supposed to define your own sentinel.

Philipp suggested to avoid this *only* for buffers collecting stderr output,
and *only* when make-process defines :stderr.  That's a much more specialized
situation, which is also quite new.  By contrast, what you are saying seems to
suggest changes in general behavior of make-process that IMO are a no-starter
at this time.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]