[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH]? svn status performance (win32)

From: Branko Čibej <brane_at_xbc.nu>
Date: 2004-10-26 00:12:41 CEST

Martin Hauner wrote:

> Branko Čibej wrote:
>
>> This one I don't like on philosophical grounds. If we do make
>> special-file checks platform-specific -- something I don't like,
>> either -- then the ifdef belongs inside svn_io_check_special_path,
>> not in its callers.
>
> I don't really like the #ifdef but it was the easiest for testing.
> A better way would be a has_special_files function which returns false
> for
> Win32 and true for anything else and then add a has_special_files to
> io_check_special to save the apr_stat call.

It shouldn't be a platform-specific #ifdef but a feature-specific one,
as you note. Once that's taken care of, it doesn't matter if it's an if
or an #ifdef. The main problem was adding conditionals to the call site
instead of in the function itself.

>> My guess is that you meant to make the ..._p2 function private in
>> libsvn_wc and have those functions that already do the check and then
>> call svn_wc_text_modified_p call the private function instead.
>
> Yes, much better this way. :) Would it create an acceptable patch?

Well, if I were you I'd repeat the performance measurement now that the
first patch has been applied. As I said, I measure a 60% (real time)
speedup with that patch, so it's not exactly obvious that this change
would make a sgignificant difference.

-- Brane

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Oct 26 00:12:49 2004

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.