Re: CVS update: subversion/subversion/libsvn_fs dag.c

Yoshiki Hayashi <yoshiki@xemacs.org> writes:

> > +dag_node_t *svn_fs__dag_dup (dag_node_t *node,
>
> > +svn_error_t *svn_fs__dag_open (dag_node_t **child_p,
>
> Well, you might want to consider my patch for these two. ;-)

Doh! I totally forgot about this!

I looked back through my mail history and found your patches.
Regarding the svn_fs__dag_dup one, I'm not sure that your (or my, for
that matter) implementation of that is complete. Specifically, the
fact that the function does nothing more than alloc() and memcpy()
seems too trivial. My guess, which I'm hoping Jim will judge, is that
the real point is to reproduce a dag_node_t entirely in the given
trail->pool, which means that memcpy()'s are no good (since the
pointers will still refer to memory potentially alloc'ed in a
different pool). But. Perhaps I'm just "making a mountain out of a
molehill," to quote my mother's favorite phrase.
Received on Sat Oct 21 14:36:23 2006