You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Greg Stein <gs...@lyra.org> on 2002/06/06 23:18:06 UTC

Re: svn commit: rev 2106 - trunk/subversion/libsvn_fs trunk/subversion/libsvn_ra_local trunk/subversion/svnlook trunk/subversion/clients/cmdline trunk/subversion/tests/libsvn_fs

On Thu, Jun 06, 2002 at 03:18:59PM -0700, Blair Zajac wrote:
> gstein@tigris.org wrote:
>...
> > +++ trunk/subversion/tests/libsvn_fs/fs-test.c  Thu Jun  6 15:30:36 2002
> > @@ -4292,7 +4292,7 @@
> >       crappy.  We need to keep this buffer fairly large so we don't run
> >       out of memory doing undeltification of large files into tiny
> >       buffers.  Issue #465.  */
> > -  char buf[100000];
> > +  char *buf = apr_palloc (pool, 100000);
> 
> I think this change causes make check's run-fs-tests.sh to use an
> inordinate amount of memory when it didn't use to, over 80% on my
> 256 MByte Linux box on and takes forever to run.

It shouldn't. There is one caller to get_file_digest(), and it clears the
pool on each iteration.

That said, it is hard to argue with empirical evidence :-). But that just
means there is *another* problem to fix. Not to restore the old code.
Allocating 100k on the stack is just plain wrong. Big Badness(tm).

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org