diff to tag includes .hgtags
Kevin Smith
yarcs at qualitycode.com
Tue Jun 28 15:12:03 UTC 2005
Bryan O'Sullivan wrote:
> On Tue, 2005-06-28 at 08:45 -0400, Kevin Smith wrote:
>
>
>> So let's pretend
>>you are the mercurial maintainer and I have posted some proposed
>>contributions on a pull server somewhere. You can't just pull the two
>>changes you like and ignore my other garbage, right? So what do you
>>actually do?
>>
>>My guess is that you would set up a local mirror of my repo, and pull
>>everything I have. Then you would use export to grab my good patches,
>>and import to stuff them into your mercurial repo.
>
>
> The only way this scales is if the people doing the pushes actually give
> clean trees to push. In the case of the kernel, Linus has historically
> refused to pull changes that aren't completely clean, so the onus has
> been on contributors to provide One Clean Patch per logical unit of
> work.
Do you mean "clean trees" as in they contain nothing but the official
tree plus patches that should all be accepted? Or do you mean trees that
might contain some weird patches, but whatever proposed patches the
include are themselves clean (that is, the patches specifically proposed
for inclusion in the mainline each encapsulate one clean unit of work)?
Or both?
If we assume realistic human nature, and therefore guess that people
won't always provide clean trees (whatever the definition), then what is
the scalable alternative process? Is it emailing patches around? That
just seems so primative, although it does have the benefit of allowing
all mailing list readers to review patches as they are submitted.
I have not worked in a project where most of the work is submitted via
emailed patches, so this is all new to me. The good news is that I'm not
tainted by thinking about how such a project worked with CVS or BK. The
bad news is that I am strongly tainted by being immersed in centralized
CVS development for so long. I REALLY want to break out and start using
fully decentralized processes, but without some best practices docs it's
really hard to know how to structure the workflow.
Thanks,
Kevin
More information about the Mercurial
mailing list