Argh!

Gwynne and I talked briefly about changing the svn directory layout to
put trunk and branches at the very top to make things more cvs-like.
She was saying that this is not how svn is usually organized, but
looking around I see other large projects doing that.

eg. http://websvn.kde.org/

I know re-running the 21-hour conversion script would suck, but we will
have to live with this stuff for years and I am not looking forward to
having to write:

for $i in `svn ls http://svn.php.net/repository/pecl`; svn co $i/trunk

for the next many years just to get a copy of all of pecl trunk.

It wouldn't make pulling branches any harder.  It would just be:

svn co http://svn.php.net/repository/branches/pecl/apc/APC_3_0

vs.

svn co http://svn.php.net/repository/pecl/apc/branches/APC_3_0

-Rasmus