[PATCH 0 of 2] help: RST formatting

Olav Reinert seroton10 at gmail.com
Sun May 20 12:32:27 CDT 2012


On 18. maj 2012, at 00.40, Matt Mackall wrote:
> And that's just one of numerous reasons why such a big patch is a
> non-starter. Others include lack of bisectability, and the inability of
> the reviewer to accept good bits and send others back for rework.
> 
> Again, this really needs to be broken into several steps.

OK.

My motivation for making the patches the way they are was to submit coherent changes that introduce a tangible and meaningful change of behaviour to the system.  But I'm learning that that just isn't what you care about (the most).

I take it you don't mind receiving a string of small (reviewable and bisectable) patches containing only refactoring to prepare the introduction of change?

> But if you say "no really, I just did the minimum work to move the code
> and didn't slip in any other changes because it seemed like a convenient
> time to do it" then I'll probably believe you.

If the patch file format was able to directly represent moving blocks of text between files (instead of removing some lines in one, and maybe adding the same lines elsewhere), I wouldn't need to make you believe anything, it would be evident.

I'm left wondering what the preferred way to submit such a patch is.



More information about the Mercurial-devel mailing list