Advantages of MQ over vanilla hg

[Theodore Tso]

On Sun, Mar 11, 2007 at 02:13:46PM +0100, Thomas Arendsen Hein wrote:
> > Put it another way, if you didn't have an upstream tree (i.e. you are
> > just maintaining your own isolated tree and whatever other people
> > choose to do with that downstream is irrelevant), then you wouldn't
> > bother with MQ, you would just use vanilla hg?
> 
> Correct, with your use case you probably don't get benefits from
> using mq.
> 
> The only exception might be if you want to rewrite older changesets,
> e.g. to remove private information before publishing your
> repository. But you can defer the usage of mq until then.

An upstream maintainer might also want to use mq (or quilt) if they
they want to test and fix up a patch series before casting them into
stone.  Sometimes I find it's easier to edit patch files (to remove
trailing whitespace, fix up the commit comments, etc.), all at one go,
and being able to do "hg qpop -a ; <hack on patch files>; hg qpush -a"
is highly convenient.

I suppose I should just be hard-nosed and just reject the whole patch
series until the submitter gets it right, sometimes it's easier just
to fix it up myself (although admittedly that doesn't scale in the
long term).

Regards,

						- Ted