> -----Original Message-----
> From: Greg Stein [mailto:gst...@gmail.com]
> Sent: woensdag 23 mei 2012 4:54
> To: dev@subversion.apache.org
> Subject: Re: svn commit: r1341715 - in /subversion/trunk/subversion:
> libsvn_wc/wc-queries.sql libsvn_wc/wc_db.c
tests/libsvn_wc/wc-queries-test.c
>
> On Tue, May 22, 2012 at 9:31 PM, <rhuij...@apache.org> wrote:
> >...
> > +++ subversion/trunk/subversion/libsvn_wc/wc_db.c Wed May 23 01:31:09
> 2012
> > @@ -4947,17 +4947,29 @@ populate_targets_tree(svn_wc__db_wcroot_
> > const char *changelist = APR_ARRAY_IDX(changelist_filter, i,
> > const char *);
> >
> > - SVN_ERR(svn_sqlite__get_statement(&stmt, wcroot->sdb,
stmt_idx));
> > + SVN_ERR(svn_sqlite__get_statement(&stmt, wcroot->sdb,
> > +
STMT_INSERT_TARGET_WITH_CHANGELIST));
> > SVN_ERR(svn_sqlite__bindf(stmt, "iss", wcroot->wc_id,
> > local_relpath, changelist));
> > SVN_ERR(svn_sqlite__update(&sub_affected, stmt));
> >
> > + /* If the root is matched by the changelist, we don't have to
match
> > + the children. As that tells us the root is a file */
> > + if (!sub_affected && depth > svn_depth_empty)
> > + {
> > + SVN_ERR(svn_sqlite__get_statement(&stmt, wcroot->sdb,
stmt_idx));
> > + SVN_ERR(svn_sqlite__bindf(stmt, "iss", wcroot->wc_id,
> > + local_relpath, changelist));
> > + SVN_ERR(svn_sqlite__update(&sub_affected, stmt));
> > + }
> > +
> > affected_rows += sub_affected;
> > }
> > }
> > else /* No changelist filtering */
> > {
> > int stmt_idx;
> > + int sub_affected;
> >
> > switch (depth)
> > {
> > @@ -4983,9 +4995,19 @@ populate_targets_tree(svn_wc__db_wcroot_
> > break;
> > }
> >
> > - SVN_ERR(svn_sqlite__get_statement(&stmt, wcroot->sdb, stmt_idx));
> > + SVN_ERR(svn_sqlite__get_statement(&stmt, wcroot->sdb,
> > + STMT_INSERT_TARGET));
> > SVN_ERR(svn_sqlite__bindf(stmt, "is", wcroot->wc_id,
local_relpath));
> > - SVN_ERR(svn_sqlite__update(&affected_rows, stmt));
> > + SVN_ERR(svn_sqlite__update(&sub_affected, stmt));
> > + affected_rows += sub_affected;
> > +
> > + if (depth > svn_depth_empty)
> > + {
>
> Why is sub_affected not used here? (like above) ... should a comment
> be added to explain?
This just inserts the directory itself. The previous statement (which has a
comment) will only insert a node when it is not a directory. Directories
still can't be part of a changelist.
I just used the result 'did we insert a file' as a quick win to avoid
looking for changelists on children of files.
Bert
>
> >...
>
> Cheers,
> -g
