Re: svn memory leak -- better example

On Sun, 2004-07-18 at 19:59, Michael wrote:
> Here is a better example of the memory leak.

For a commit, we deliberately retain a bit of information for each file
being committed, so that we can send file diffs at the end of a commit
after conflict detection. We could perhaps reduce the amount of
information retained per file with some careful coding (we already go to
some effort to avoid allocating an 8K memory pool for each file), but
there will always be some number of files such that a commit over all
those files causes an unacceptable amount of memory to be used.

I don't know if this kind of memory use is expected from cleanup.
Entries caching may be relevant, though I think there's something more
than that going on.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jul 19 19:46:22 2004