On Thu, 15 Jul 2004 16:54:18 -0500, Robert Guthrie <rguthrie@pobox.com> wrote:
> Well, my experiment was with
> 1. a repository that I created for the purpose of testing
> 2. a working copy that I checked out under userid "rguthr"
> 3. a copy of the huge file in question, which I copied with the "rguthr"
> user.
>
> But, just to make sure I'm not crazy....
>
> rguthr $> ls -l . tempdir
> .:
> total 0
> drwxr-s--- 3 rguthr cmt 256 Jul 15 16:46 tempdir/
>
> tempdir:
> total 23754400
> -rw-r----- 1 rguthr cmt 0 Jul 15 16:46 emptyfile1
> -rw-r----- 1 rguthr cmt 0 Jul 15 16:46 emptyfile2
> -rw-r----- 1 rguthr cmt 12162160632 Jul 15 16:42 reallybigfile
>
> rguthr $> svn stat
> svn: Can't read directory 'tempdir': Partial results are valid but
> processing is incomplete
> rguthr $> rm tempdir/reallybigfile
> rguthr $> svn stat # no response, because svn:ignore '*' on tempdir
> rguthr $> whoami
> rguthr
>
> ... so the empty files, and just about any other reasonablly sized ones
> seem not to cause this problem. Just outlandishly huge ones.
> Unfortunately, we end up with really huge files all the time here, but
> really, this shouldn't be a problem for subversion, should it?
Did you build APR with large file support? APR 0.9 doesn't default to
this. Make sure that the following extra flag gets passed to the
compiler when building APR:
-D_FILE_OFFSET_BITS=64
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Jul 16 18:21:29 2004