[PATCH 0 of 7 F2 series] fncache2 repo format

[Adrian Buehlmann]

On 2012-10-09 22:17, Bryan O'Sullivan wrote:
> On Tue, Oct 9, 2012 at 9:29 AM, Adrian Buehlmann <adrian at cadifra.com
> <mailto:adrian at cadifra.com>> wrote:
> 
> 
>     It might be interesting to know more about that. Did it fail on hg
>     verify? Can you do a "clone --pull" and then do a verify again?
> 
> 
> It actually failed during the update phase of "clone --pull".

Interesting. With what error message?

And a 'clone --pull -U' succeeds and leaves a repo that passes verify?

>     I still can't figure out what could be wrong. Which perhaps doesn't mean
>     much. There's no proof that my fncache2 code is bug free.
> 
> 
> Yeah, I face a similar bug with the fncache C code.

Not to the same extent ;). For fncache2, there exist no user-repos yet.

But fncache2 would create another C-code / Python-code pair. Albeit a
simpler one.

>     I'd recommend you clone your repo again with V3.
> 
> 
> I'll try. It takes hours to run a clone --pull, though ;-(

Yep. Takes time for any interesting repo.

>     Reviewing this is a lot simpler
> 
>     than hashmangle C function (where I stall each time I try to swallow it
>     and no one else has tried to have a look at so far).
> 
> 
> I'm working on the reviewability of that function. Should have something
> to show tomorrow.

Don't forget the tests.

For the current repo format we have the benefit of a reference
implementation in Python. And the Python code has survived a non-zero,
unknown, number of users throwing file names at it.