You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Greg Stein <gs...@gmail.com> on 2009/10/07 21:50:14 UTC
Re: svn commit: r39625 - trunk/subversion/libsvn_wc
On Sat, Sep 26, 2009 at 12:31, Hyrum K. Wright <hy...@hyrumwright.org> wrote:
>...
> +++ trunk/subversion/libsvn_wc/update_editor.c Sat Sep 26 09:31:53 2009 (r39625)
> @@ -1830,23 +1830,27 @@ set_copied_callback(const char *local_ab
> apr_pool_t *scratch_pool)
> {
> struct set_copied_baton_t *b = walk_baton;
> - const svn_wc_entry_t *entry;
> + svn_wc__db_status_t status;
> + svn_wc__db_kind_t kind;
>
> if (strcmp(local_abspath, b->added_subtree_root_path) == 0)
> return SVN_NO_ERROR; /* Don't touch the root */
>
> - SVN_ERR(svn_wc__get_entry(&entry, b->eb->db, local_abspath, FALSE,
> - svn_node_unknown, FALSE,
> - scratch_pool, scratch_pool));
> + SVN_ERR(svn_wc__db_read_info(&status, &kind, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL, NULL, NULL, NULL,
> + NULL, NULL, NULL, NULL, NULL,
> + b->eb->db, local_abspath,
> + scratch_pool, scratch_pool));
This loses the SHOW_HIDDEN == FALSE concept. Either keep it in, or I'd
like to know why you can ignore it...
(specifically, I think this might affect depth=exclude? maybe the walk
doesn't visit parent stubs?)
>...
Of course, it would be Even Better to translate that copied=true flag
into its wc_db meaning. Without tracing thru the entry-writing code,
I'm not really sure.
Cheers,
-g
> - if (entry->kind == svn_node_dir)
> + if (kind == svn_wc__db_kind_dir)
> {
> /* We don't want to mark a deleted PATH as copied. If PATH
> is added without history we don't want to make it look like
> it has history. If PATH is replaced we don't want to make
> it look like it has history if it doesn't. Only if PATH is
> schedule normal do we need to mark it as copied. */
> - if (entry->schedule == svn_wc_schedule_normal)
> + if (status == svn_wc__db_status_normal)
> {
> svn_wc_entry_t tmp_entry;
>
> @@ -1867,7 +1871,7 @@ set_copied_callback(const char *local_ab
> it has history. If PATH is replaced we don't want to make
> it look like it has history if it doesn't. Only if PATH is
> schedule normal do we need to mark it as copied. */
> - if (entry->schedule == svn_wc_schedule_normal)
> + if (status == svn_wc__db_status_normal)
> {
> svn_wc_entry_t tmp_entry;
>
> @@ -1875,7 +1879,9 @@ set_copied_callback(const char *local_ab
> tmp_entry.copied = TRUE;
> SVN_ERR(svn_wc__entry_modify2(b->eb->db,
> local_abspath,
> - entry->kind,
> + kind == svn_wc__db_kind_dir
> + ? svn_node_dir
> + : svn_node_file,
> FALSE,
> &tmp_entry,
> SVN_WC__ENTRY_MODIFY_COPIED,
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=495&dsMessageId=2400651
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2404705
RE: svn commit: r39625 - trunk/subversion/libsvn_wc
Posted by Bert Huijben <rh...@sharpsvn.net>.
> -----Original Message-----
> From: Greg Stein [mailto:gstein@gmail.com]
> Sent: woensdag 7 oktober 2009 23:50
> To: dev@subversion.tigris.org
> Subject: Re: svn commit: r39625 - trunk/subversion/libsvn_wc
>
> On Sat, Sep 26, 2009 at 12:31, Hyrum K. Wright <hy...@hyrumwright.org>
> wrote:
> >...
> > +++ trunk/subversion/libsvn_wc/update_editor.c Sat Sep 26 09:31:53
> 2009 (r39625)
> > @@ -1830,23 +1830,27 @@ set_copied_callback(const char *local_ab
> > apr_pool_t *scratch_pool)
> > {
> > struct set_copied_baton_t *b = walk_baton;
> > - const svn_wc_entry_t *entry;
> > + svn_wc__db_status_t status;
> > + svn_wc__db_kind_t kind;
> >
> > if (strcmp(local_abspath, b->added_subtree_root_path) == 0)
> > return SVN_NO_ERROR; /* Don't touch the root */
> >
> > - SVN_ERR(svn_wc__get_entry(&entry, b->eb->db, local_abspath, FALSE,
> > - svn_node_unknown, FALSE,
> > - scratch_pool, scratch_pool));
> > + SVN_ERR(svn_wc__db_read_info(&status, &kind, NULL, NULL, NULL,
> > + NULL, NULL, NULL, NULL, NULL, NULL,
> NULL,
> > + NULL, NULL, NULL, NULL, NULL, NULL,
> NULL,
> > + NULL, NULL, NULL, NULL, NULL,
> > + b->eb->db, local_abspath,
> > + scratch_pool, scratch_pool));
>
> This loses the SHOW_HIDDEN == FALSE concept. Either keep it in, or I'd
> like to know why you can ignore it...
>
> (specifically, I think this might affect depth=exclude? maybe the walk
> doesn't visit parent stubs?)
>
> >...
>
> Of course, it would be Even Better to translate that copied=true flag
> into its wc_db meaning. Without tracing thru the entry-writing code,
> I'm not really sure.
One of the reasons I avoided changing this function is that it will break on switched paths.
I think this tree conflict handling helper needs a redesign, because you don't want local copies of switched paths, etc.
Bert
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2404708