You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Bert Huijben <be...@qqmail.nl> on 2011/07/07 13:06:08 UTC

RE: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn


> -----Original Message-----
> From: philip@tigris.org [mailto:philip@tigris.org]
> Sent: donderdag 7 juli 2011 12:56
> To: issues@subversion.tigris.org
> Subject: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn
> 
> http://subversion.tigris.org/issues/show_bug.cgi?id=3951
> 
> 
> 
> 
> 
> 
> ------- Additional comments from philip@tigris.org Thu Jul  7 03:56:05
-0700
> 2011 -------
> As issue #3948 the client gets both props via serf/1.6 and only the new
prop
> in
> other cases.  When merging the props update_editor.c:close_directory calls
> props.c:svn_wc__merge_props, which calls
> props.c:apply_single_prop_change which
> calls props.c:apply_single_generic_prop_change, and there we have:
> 
> 1394      /* If working_val is the same as new_val already then there is
> 1395       * nothing to do */
> 1396      if (working_val && new_val
> 1397          && svn_string_compare(working_val, new_val))
> 1398        {
> 1399           set_prop_merge_state(state, svn_wc_notify_state_merged);
> 1400        }

This exact block was added by stsp in r1136063, about a month ago.
[[
Fix issue #3919. During a merge of a property, add a check against the
incoming new property value and the working copy value. If they
already match, then the merge trivially succeeds.

* subversion/libsvn_wc/props.c
  (apply_single_generic_prop_change): Do nothing if the incoming new
  property value already matches the working value.

Patch by: Brian Neal <bg...@gmail.com>
          me
]]

I think we shouldn't set the state to merged here, (but of course allow the
merge without conflict).

	Bert


RE: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn

Posted by Bert Huijben <be...@qqmail.nl>.

> -----Original Message-----
> From: Philip Martin [mailto:philip.martin@wandisco.com]
> Sent: donderdag 7 juli 2011 16:46
> To: Bert Huijben
> Cc: dev@subversion.apache.org; issues@subversion.tigris.org
> Subject: Re: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn
> 
> "Bert Huijben" <be...@qqmail.nl> writes:
> 
> >> -----Original Message-----
> >> From: philip@tigris.org [mailto:philip@tigris.org]
> >> Sent: donderdag 7 juli 2011 12:56
> >> To: issues@subversion.tigris.org
> >> Subject: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn
> >>
> >> http://subversion.tigris.org/issues/show_bug.cgi?id=3951
> >>
> >>
> >>
> >>
> >>
> >>
> >> ------- Additional comments from philip@tigris.org Thu Jul  7 03:56:05
> > -0700
> >> 2011 -------
> >> As issue #3948 the client gets both props via serf/1.6 and only the new
> > prop
> >> in
> >> other cases.  When merging the props update_editor.c:close_directory
> calls
> >> props.c:svn_wc__merge_props, which calls
> >> props.c:apply_single_prop_change which
> >> calls props.c:apply_single_generic_prop_change, and there we have:
> >>
> >> 1394      /* If working_val is the same as new_val already then there
is
> >> 1395       * nothing to do */
> >> 1396      if (working_val && new_val
> >> 1397          && svn_string_compare(working_val, new_val))
> >> 1398        {
> >> 1399           set_prop_merge_state(state, svn_wc_notify_state_merged);
> >> 1400        }
> >
> > This exact block was added by stsp in r1136063, about a month ago.
> > [[
> > Fix issue #3919. During a merge of a property, add a check against the
> > incoming new property value and the working copy value. If they
> > already match, then the merge trivially succeeds.
> >
> > * subversion/libsvn_wc/props.c
> >   (apply_single_generic_prop_change): Do nothing if the incoming new
> >   property value already matches the working value.
> >
> > Patch by: Brian Neal <bg...@gmail.com>
> >           me
> > ]]
> >
> > I think we shouldn't set the state to merged here, (but of course allow
the
> > merge without conflict).
> 
> This is a bit of a special case when base==working==incoming.  Normally
> we don't see incoming==base.  I am experimenting with this patch:

Via mod-dav we have a few cases where we receive unchanged properties.
(I think Paul Burba knows more about this than anybody would like to know)

	Bert


Re: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn

Posted by Philip Martin <ph...@wandisco.com>.
"Bert Huijben" <be...@qqmail.nl> writes:

>> -----Original Message-----
>> From: philip@tigris.org [mailto:philip@tigris.org]
>> Sent: donderdag 7 juli 2011 12:56
>> To: issues@subversion.tigris.org
>> Subject: [Issue 3951] serf fail prop_tests.py 3 with 1.6 mod_dav_svn
>> 
>> http://subversion.tigris.org/issues/show_bug.cgi?id=3951
>> 
>> 
>> 
>> 
>> 
>> 
>> ------- Additional comments from philip@tigris.org Thu Jul  7 03:56:05
> -0700
>> 2011 -------
>> As issue #3948 the client gets both props via serf/1.6 and only the new
> prop
>> in
>> other cases.  When merging the props update_editor.c:close_directory calls
>> props.c:svn_wc__merge_props, which calls
>> props.c:apply_single_prop_change which
>> calls props.c:apply_single_generic_prop_change, and there we have:
>> 
>> 1394      /* If working_val is the same as new_val already then there is
>> 1395       * nothing to do */
>> 1396      if (working_val && new_val
>> 1397          && svn_string_compare(working_val, new_val))
>> 1398        {
>> 1399           set_prop_merge_state(state, svn_wc_notify_state_merged);
>> 1400        }
>
> This exact block was added by stsp in r1136063, about a month ago.
> [[
> Fix issue #3919. During a merge of a property, add a check against the
> incoming new property value and the working copy value. If they
> already match, then the merge trivially succeeds.
>
> * subversion/libsvn_wc/props.c
>   (apply_single_generic_prop_change): Do nothing if the incoming new
>   property value already matches the working value.
>
> Patch by: Brian Neal <bg...@gmail.com>
>           me
> ]]
>
> I think we shouldn't set the state to merged here, (but of course allow the
> merge without conflict).

This is a bit of a special case when base==working==incoming.  Normally
we don't see incoming==base.  I am experimenting with this patch:

Index: subversion/libsvn_wc/props.c
===================================================================
--- subversion/libsvn_wc/props.c	(revision 1143768)
+++ subversion/libsvn_wc/props.c	(working copy)
@@ -1393,7 +1393,8 @@
   if (working_val && new_val
       && svn_string_compare(working_val, new_val))
     {
-       set_prop_merge_state(state, svn_wc_notify_state_merged);
+      if (! svn_string_compare(old_val, working_val))
+        set_prop_merge_state(state, svn_wc_notify_state_merged);
     }
   /* If working_val is the same as old_val... */
   else if (working_val && old_val



-- 
uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com