[RFC] bisect extension for mercurial

[Matt Mackall]

On Thu, Sep 29, 2005 at 07:53:09PM +0200, Benoit Boissinot wrote:
> On Thu, Sep 29, 2005 at 02:55:11PM +0200, Benoit Boissinot wrote:
> > On Wed, Sep 28, 2005 at 01:51:46PM -0700, Bryan O'Sullivan wrote:
> > > On Wed, 2005-09-28 at 20:37 +0200, Benoit Boissinot wrote:
> > > 
> > > > hg bisect-init # start bisecting
> > > > hg bisect-bad # mark current revision as broken
> > > > hg bisect-good [<rev>] # mark <rev> as working
> > > > ... the bisect code finds a new revision to try
> > > > ... see if it works
> > > > hg bisect-good # if it worked
> > > > hg bisect-bad # it doesn't work
> > > 
> > > This clutters the command name space too much.
> > > 
> > > Please just use one command, "bisect", and make each thing a subcommand
> > > of it.
> > > 
> > > 	<b
> > > 
> > Attached is a new version which uses "hg bisect <subcommand>"
> > 
> I have put an updated version (rewrote the help code) on
> http://perso.ens-lyon.fr/benoit.boissinot/hbisect.py

Folks, please always send an actual patch in the standard format. My
setup is heavily optimized for accepting and queueing patches. Alt-g
in my mailer puts it in the queue for later batch processing, and I
have some other tools that automate that processing.

Accepting whole files means manually saving the file, scping it, hg
add, manually setting author name for commit, and constructing or
cut-and-pasting a commit description and generally disrupts my whole
batching process.

-- 
Mathematics is the supreme nostalgia of our time.