Re: memory pool problem in merge

2011-11-21 Thread Stefan Sperling
On Mon, Nov 21, 2011 at 01:00:08PM +, Philip Martin wrote: > Stefan Sperling writes: > > > I think your patch makes sense as-is. This code creates the conflicted_paths > > table and should ensure that anything referenced from the table has the > > same lifetime as the table itself. > > I'm h

Re: memory pool problem in merge

2011-11-21 Thread Philip Martin
Stefan Sperling writes: > I think your patch makes sense as-is. This code creates the conflicted_paths > table and should ensure that anything referenced from the table has the > same lifetime as the table itself. I'm happy with that for tree_conflict. I think tree_conflict_on_add should do the

Re: memory pool problem in merge

2011-11-21 Thread Stefan Sperling
On Mon, Nov 21, 2011 at 12:12:56PM +, Philip Martin wrote: > Philip Martin writes: > > > This hack fixes the problem: > > > > Index: subversion/libsvn_client/merge.c > > === > > --- subversion/libsvn_client/merge.c(revi

Re: memory pool problem in merge

2011-11-21 Thread Philip Martin
Philip Martin writes: > This hack fixes the problem: > > Index: subversion/libsvn_client/merge.c > === > --- subversion/libsvn_client/merge.c (revision 1203766) > +++ subversion/libsvn_client/merge.c (working copy) > @@ -593,6 +5