Re: split up process.c [was: Re: Changes in process.c]

[Eli Zaretskii <eliz@gnu.org>]

> Cc: emacs-devel@gnu.org
> From: Dan Nicolaescu <dann@gnu.org>
> Date: Thu, 08 Jul 2010 19:05:29 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Cc: emacs-devel@gnu.org
> >> From: Dan Nicolaescu <dann@gnu.org>
> >> Date: Thu, 08 Jul 2010 14:52:39 -0400
> >> 
> >> Why don't we split the part for supporting MS-DOS into a different
> >> file: process-no-subprocesses.c (or some better name)
> >> 
> >> That makes the file easier to read, less clunky, and problems easier
> >> to catch with a simple grep.
> >
> > It will still leave two implementations of the same code.  
> 
> That's better than what we currently have.

I did something more radical (revno 100767): unify the two branches of
process.c.

There's now only one function, wait_reading_process_output, which has
2 different implementations.  (I could easily have a single function
with two different bodies conditioned by `subprocesses', or I could
move the second implementation to msdos.c, if people prefer that.  But
both alternatives looked no cleaner, and the latter would even make
more maintenance headaches, IMO.)

Other than this single function, the rest is unified, at the cost of a
few "#ifdef subprocesses" here and there.  Not surprisingly, I found
and fixed a few bugs along the way...