[PATCH] Increase performance on Windows by up to 2x

[Matt Mackall]

On Thu, 2009-03-26 at 16:26 -0700, Bryan O'Sullivan wrote:
> On Thu, Mar 26, 2009 at 4:21 PM, TK Soh <teekaysoh at gmail.com> wrote:
>  
>         Just curious. Any benchmarks to show where the gains are?
> 
> Mostly on the read path, since reads are buffered and don't require
> whatever the Windows equivalent of a system call is. Mercurial issues
> a number of small read requests on every revlog it reads, so the
> buffering is a big win there.

Well it shouldn't. We should have one read to grab the index and one
read to grab the relevant chunk of the log. In the interleaved case, we
should have one read total per log because we should be reading the
entire thing at once.

-- 
http://selenic.com : development and support for Mercurial and Linux