[PATCH 0 of 3 V2] move cache files to a collected cache folder .hg/caches/

Matt Mackall mpm at selenic.com
Mon Jan 3 16:24:06 CST 2011


On Mon, 2011-01-03 at 00:28 +0100, Jason Harris wrote:

> We are spending too more time discussing this... What you say below is true in
> that I would also of course go with a copener approach when you introduce the
> specialization as you are saying below. But generally I would specialize *only*
> when I have actually added code like you mention below... 
> 
> But I just want someone who can push this patch to comment and I'll fix the
> patch the way they want and be done with it...

And here I was waiting for the dust to settle. My thoughts:

- the general idea of a cache directory is good
- tests should be fixed in the same patch that introduces the changes
for bisectability
- running check-code is strongly encouraged
- running the test suite is also encouraged (and check-code is run by
the test suite so it's easy to spot when you haven't done it!)
- you've got the right idea about ordering patches so that optional bits
are at the end
- I have mixed feelings about introducing another opener object, as this
is moving us into fairly fine semantic distinctions (compare opener vs
sopener and opener vs copener)

So let's see a fixed up patch 1+2 and we'll consider patch 3 after that.

-- 
Mathematics is the supreme nostalgia of our time.




More information about the Mercurial-devel mailing list