Home
Reading
Searching
Subscribe
Sponsors
Statistics
Posting
Contact
Spam
Lists
Links
About
Hosting
Filtering
Features Download
Marketing
Archives
FAQ
Blog
 
Gmane
From: Adrian Buehlmann <adrian <at> cadifra.com>
Subject: Re: [PATCH 0 of 3 v4] "final" fncache-in-C code
Newsgroups: gmane.comp.version-control.mercurial.devel
Date: Friday 7th September 2012 09:53:30 UTC (over 5 years ago)
On 2012-09-07 01:31, Bryan O'Sullivan wrote:
> This version should (hopefully) build under MSVC, no longer uses fixed
> size buffers internally, and passes all of Adrian's new hybridencode
> tests.

I'm still -1 on doing hashed paths in C.

I think I now have demonstrated that the risk is pretty high, given how
broken your first attempt was.

The vast majority of the paths are unhashed, so I still don't think it
is worth doing the hashed code path in C.

I spent some time adding a couple of testcases in order to defend the
status quo, but I doubt it will be sufficient.

Doing basicencode() in C is ok by me now, but the hashencode() function
should IMHO stay in Python-land.

In particular, I hate the hashmangle C function.
 
CD: 3ms