[PATCH 1 of 3 v4] store: implement fncache path mangling code in C

Adrian Buehlmann adrian at cadifra.com
Fri Sep 7 14:45:40 CDT 2012


On 2012-09-07 21:25, Bryan O'Sullivan wrote:
> On Fri, Sep 7, 2012 at 12:07 PM, Adrian Buehlmann <adrian at cadifra.com
> <mailto:adrian at cadifra.com>> wrote:
> 
>     Bryan, can I ask you what "slop" means?
> 
> 
> In this context, it means "leftovers". It serves the same purpose as
> this code in
> store.py: http://selenic.com/repo/hg/file/3ee5d3c372fa/mercurial/store.py#l218

Ok. Thanks. I found the word in the English <-> German dictionary. I
first falsely thought it was an abbreviation. My bad.

The word is good.

> 
>     If I would have had to write such kind of code, I would be almost 100%
>     sure *I* would have (unintentionally) inserted at least 5 bugs in it.
> 
> 
> Me too, since I fixed about that many while writing it, mainly due to
> all the extra test cases you added to test-hybridencode.py.
> 
> I'm not certain there are no remaining bugs, but the current code is
> quite a lot closer to correct.

I feel the same. But it also feels a bit like an attack what you are
doing here.

Don't get me wrong: I know you have very good intentions and I know you
are absolutely not careless.

What I still don't understand is, why is it worth the risk (and the
effort) here?

> I'll be writing a fuzz tester to exercise this code further.

Ok. That sounds great.


More information about the Mercurial-devel mailing list