List Info

Thread: New - Handle sparsely-populated directories, or portions of a tree not checked out due




New - Handle sparsely-populated directories, or portions of a tree not checked out due
user name
2007-07-13 16:49:10
http://subversion.tigris.org/issues/show_bug.cgi?id=2827

                 Issue #|2827
                 Summary|Handle sparsely-populated
directories, or portions of 
                        |a tree not checked out due to
insufficient authz
               Component|subversion
                 Version|trunk
                Platform|All
                     URL|
              OS/Version|All
                  Status|NEW
       Status whiteboard|
                Keywords|
              Resolution|
              Issue type|ENHANCEMENT
                Priority|P2
            Subcomponent|libsvn_client
             Assigned to|pburba
             Reported by|dlr






------- Additional comments from dlrtigris.org Fri Jul 13
14:49:10 -0700 2007 -------
Handle sparsely-populated directories, or portions of a tree
not
checked out due to lack of authorization.

* Inheritance of merge info through sparsely-populated
directories
  could potentially be handled via explicit merge info (e.g.
to
  avoid recording merge info for paths not checked out).

* There's little we can do about paths not checked out
because of
  insufficient authorization.  Since the client isn't
allowed to
  know these paths exist, we can't avoid potentially
inherited merge
  info!

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle sparsely-populated directories, or portions of a tree not checked out due to i
user name
2007-07-13 16:49:24
http://subversion.tigris.org/issues/show_bug.cgi?id=2827




User dlr changed the following:

                What    ld value 
               |New value
============================================================
====================
                Keywords|                         
|merge-tracking
------------------------------------------------------------
--------------------
        Target milestone|---                       |1.5
------------------------------------------------------------
--------------------




------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle sparsely-populated directories, or portions of a tree not checked out due to i
user name
2007-07-17 11:43:07
http://subversion.tigris.org/issues/show_bug.cgi?id=2827







------- Additional comments from kfogeltigris.org Tue Jul 17 09:43:07 -0700 2007 -------
It's not quite clear to me how this issue relates to the
sparse-directories
work, just FYI.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle sparsely-populated directories, or portions of a tree not checked out due to i
user name
2007-07-17 12:39:56
http://subversion.tigris.org/issues/show_bug.cgi?id=2827







------- Additional comments from markphiptigris.org Tue Jul 17 10:39:56 -0700 2007 -------
Karl, this is a merge tracking task.  Suppose you use the
sparse-directories feature to create a WC with only certain
parts in it.  Then you do a merge into this working copy. 
If we do nothing special, the merge tracking code would make
it look as if a complete merge was performed, when in
reality, it only merged into the parts you had checked out.

So this issue is about coming up with a recording system for
the merge info property that makes it clear which parts of
the WC received the merge.  In theory, someone else with a
complete WC ought to be able to run merge later and have the
parts that were previously merged skipped, while now
bringing in the parts that were not merged with the previous
attempt.

Personally, I think if we can do this great, but I do not
think it should be a release blocker.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle sparsely-populated directories, or portions of a tree not checked out due to i
user name
2007-07-17 12:43:36
http://subversion.tigris.org/issues/show_bug.cgi?id=2827







------- Additional comments from kfogeltigris.org Tue Jul 17 10:43:36 -0700 2007 -------
Oh, I just didn't see the keyword, NM.  Thanks.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle merge info for sparsely-populated directories, or portions of a tree not check
user name
2007-07-17 17:56:33
http://subversion.tigris.org/issues/show_bug.cgi?id=2827




User dlr changed the following:

                What    ld value 
               |New value
============================================================
====================
OtherIssuesDependingOnTh|                          |2824
                      is|                          |
------------------------------------------------------------
--------------------
                 Summary|Handle sparsely-populated |Handle
merge info for spar
                        |directories, or portions
o|sely-populated directories
                        |f a tree not checked out d|, or
portions of a tree no
                        |ue to insufficient authz  |t
checked out due to insuf
                        |                          |ficient
authz
------------------------------------------------------------
--------------------




------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle merge info for sparsely-populated directories, or portions of a tree not check
user name
2007-08-21 20:05:19
http://subversion.tigris.org/issues/show_bug.cgi?id=2827




User dlr changed the following:

                What    ld value 
               |New value
============================================================
====================
                  Status|NEW                       |STARTED
------------------------------------------------------------
--------------------




------- Additional comments from dlrtigris.org Tue Aug 21
18:05:19 -0700 2007 -------
Paul, what effect does r26229 have on this issue?  The
commit log mentions:

Issue #2839 - Support non-inheritable mergeinfo revision
ranges.
Issue #2823 - Account for merge info differences for
switched directories.

...but not this issue.  Thanks!

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle merge info for sparsely-populated directories, or portions of a tree not check
user name
2007-08-21 20:19:39
http://subversion.tigris.org/issues/show_bug.cgi?id=2827







------- Additional comments from pburbatigris.org Tue Aug 21 18:19:39 -0700 2007 -------
Dan - r26229 doesn't fix this issue specifically, but it
does put in place the
"tool" (i.e. non-inheritable mergeinfo ranges) we
need to easily fix the sparse
directory case.  I'm not entirely sure about the authz case
yet...but the spares
checkout case I hope to fix this week.  I suppose this has
always been two
separate but closely related issues, I can split it if you
think that makes
sense at this point.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle merge info for sparsely-populated directories
user name
2007-08-27 07:40:24
http://subversion.tigris.org/issues/show_bug.cgi?id=2827




User pburba changed the following:

                What    ld value 
               |New value
============================================================
====================
                 Summary|Handle merge info for spar|Handle
merge info for spar
                        |sely-populated
directories|sely-populated directories
                        |, or portions of a tree no|
                        |t checked out due to insuf|
                        |ficient authz             |
------------------------------------------------------------
--------------------




------- Additional comments from pburbatigris.org Mon Aug 27 05:40:23 -0700 2007 -------
Split the part of this issue related to "portions of a
tree not checked out due
to lack of authorization" to a separate issue 2893.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


Handle merge info for sparsely-populated directories
user name
2007-08-29 14:54:19
http://subversion.tigris.org/issues/show_bug.cgi?id=2827







------- Additional comments from pburbatigris.org Wed Aug 29 12:54:19 -0700 2007 -------
Test added r26376.
Functional spec described r26378.
Partial fix r26379.

Bah.  I say "partial" because I just realized that
the deletion (and probably
addition) of missing children is still not handled
correctly, for example:

1) Given a working copy path WC_PATH checked out to --depth
files or --depth empty.

2) Given an operative merge of -r8:10 from URL to WC_PATH
where some immediate
directory child of WC_PATH, say WC_PATH/DIR_CHILD is deleted
in r10 and a
property addition is made to WC_PATH in r9.

After the merge, WC_PATH will end up non-inheritable
mergeinfo for -r8:10 (i.e.
'/SOURCE:9-10*').  But what happens when WC_PATH is updated
to --depth infinity
or immediates?  WC_PATH/DIR_CHILD is now present but we
can't re-merge -r8:10 to
WC_PATH to remove it because of WC_PATH's mergeinfo.

We could make the above merge only add non-inheritable
mergeinfo for r9 and not
add r10 at all.  But what if r10 is both a deletion of
WC_PATH/DIR_CHILD *and* a
prop change to WC_PATH?

Not seeing a solution to this at the moment...my thought is
to spin it off as a
separate P3/1.5-consider issue.

------------------------------------------------------------
---------
To unsubscribe, e-mail: issues-unsubscribesubversion.tigris.org
For additional commands, e-mail: issues-helpsubversion.tigris.org


[1-10] [11]

about | contact  Other archives ( Real Estate discussion Medical topics )