Re: [commit] Add add_setshow_enum_cmd, use in mips

[Mark Kettenis <mark.kettenis@xs4all.nl>]

   Date: Wed, 10 Nov 2004 22:48:40 +0200
   From: "Eli Zaretskii" <eliz@gnu.org>

   > Before committing a patch as I did, I ask the question:
   > 
   > ``Could a review significantly alter the change?''
   > 
   > Conversely, when reviewing a change, I ask:
   > 
   > ``Is the objection going to significantly alter the change?''
   > 
   > Here, I cloned a pre-existing interface, adding another variant.  Anyone 
   > on this list with the problem I hit would have come up with an identical 
   > change.
   > 
   > Waiting a week would have achieved what?

   Let me turn the table, Andrew, and ask you what possible harm could be
   caused by posting the patch for review?  At best, no one would have
   responded and you'd be committing the patch a week later.  No harm
   done; case closed.  At worst, someone _would_ have responded, and the
   patch would have been committed 7 days, instead of a week, later,
   after some short discussion.  Again, no harm done.

Sorry, but I have to disagree here.  For the (unfortunately) limited
number of people that contribute several patches in a week this is a
significant problem.  When I'm working on a particular area I often
find myself making multiple changes to the same file.  If I have to
post a patch and wait a week before I can check it in, I have two
options:

1. Juggle with the patches for a week, risking an accidental commit of
   stuff belonging to a different patch to the same file, or dropping
   a patch completely in the process.

2. Postpone further work on that part of GDB until the week is over
   and the patch has been committed.

Neither option is good for GDB.

Mark