stderr vs start-process

stderr vs start-process

[Stephen Leake]

The help string for `start-process' does not say what happens to
stderr for the spawned process. It just says:

     Process output goes at end of that buffer, unless you specify
     an output stream or filter function to handle the output.

I tested this (on Windows XP) and it seems both stderr and stdout go
to that buffer. I tried reading the C code in process.c, but quickly
got lost.

I suggest that the help string be modified to:

     Process output (both standard error and standard output) goes at
     end of that buffer, unless you specify an output stream or filter
     function to handle the output.

Is there a way to separate standard error from standard output for the
spawned process? That would probably be useful in some situations.

-- 
-- Stephe

Re: stderr vs start-process

[Johan Bockgård]

Stephen Leake <stephen_leake@member.fsf.org> writes:

> Is there a way to separate standard error from standard output for
> the spawned process? That would probably be useful in some
> situations.

(info "(elisp)Output from Processes")

   It is impossible to separate the standard output and standard error
   streams of the subprocess, because Emacs normally spawns the
   subprocess inside a pseudo-TTY, and a pseudo-TTY has only one
   output channel. If you want to keep the output to those streams
   separate, you should redirect one of them to a file--for example,
   by using an appropriate shell command.

-- 
Johan Bockgård

Re: stderr vs start-process

[Eli Zaretskii]

> From: Stephen Leake <stephen_leake@member.fsf.org>
> Date: Tue, 17 Apr 2007 20:13:20 -0400
> 
> The help string for `start-process' does not say what happens to
> stderr for the spawned process. It just says:
> 
>      Process output goes at end of that buffer, unless you specify
>      an output stream or filter function to handle the output.
> 
> I tested this (on Windows XP) and it seems both stderr and stdout go
> to that buffer. I tried reading the C code in process.c, but quickly
> got lost.
> 
> I suggest that the help string be modified to:
> 
>      Process output (both standard error and standard output) goes at
>      end of that buffer, unless you specify an output stream or filter
>      function to handle the output.

Done.

> Is there a way to separate standard error from standard output for the
> spawned process? That would probably be useful in some situations.

I added a sentence about this to the doc string.