You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "Hyrum K. Wright" <hy...@mail.utexas.edu> on 2008/12/10 18:58:03 UTC
1.5.5 next Wednesday (12/17)
1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
what's in the release, so if there are items in STATUS you want in the release,
poke, prod, campaign or bribe people to vote on them.
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
Re: 1.5.5 next Wednesday (12/17)
Posted by Paul Burba <pt...@gmail.com>.
On Wed, Dec 10, 2008 at 1:58 PM, Hyrum K. Wright
<hy...@mail.utexas.edu> wrote:
> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> what's in the release, so if there are items in STATUS you want in the release,
> poke, prod, campaign or bribe people to vote on them.
>
> -Hyrum
There are several merge tracking related groups in STATUS right now.
It's probably too much to expect them all to be reviewed, but if
anyone wants looks at just one, here are (IMHO) the highest to lowest
priority:
* r34091, r34094, r34244, r34256, r34306
Various svn merge --reintegrate improvements. Even if you hate
reintegrate, wouldn't it be nice if it was actually usable
some of the time?
* r34385, r34393
Fix broken merge when the merge target's natural history includes
resurrections. Merges that don't error out are a good thing.
* r34432, r34669
Follow-up to r32975, which is already backported and had a bug, see
http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=145757.
Fixing regressions is also good.
* r34560, r34562
Fix a bug which occurs when a merge adds a subtree as an immediate child
of the merge target when the merge target has non-inheritable mergeinfo
and is also the current working directory. If all these planets align
properly then invalid mergeinfo is created on the added subtree.
Ok, this is probably quite rare, but it is a bug.
* r34547
Fix issue #3323 'Mergeinfo deleted by a merge should disappear'
Also fairly rare, but not dire.
* r30257, r33024, r33026
Don't try to filter natural history from added mergeinfo during
reverse merges. It was suggested in IRC that this should be
de-nominated...I won't go that far, but it makes the bottom of
this list.
Paul
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983038
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Julian Foad <ju...@btopenworld.com>.
On Wed, 2008-12-17 at 18:47 +0000, Stefan Sperling wrote:
> On Thu, Dec 11, 2008 at 03:03:12PM -0600, Steve Suehs wrote:
> > I don't like being the merge gimp. I have other things to do.
>
> I've created a proper backport of this change, including all
> related follow-up changes made by Julian, here:
> https://svn.collab.net/repos/svn/branches/1.5.x-r30756
>
> David, Julian, or anyone else, if you could add +1's to STATUS
> quickly we could still get this change into 1.5.5.
OK, added my vote. One more vote needed.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985995
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Dec 18, 2008 at 03:07:06PM +0000, David O'Shea wrote:
> On 18/12/2008 14:50, Stefan Sperling wrote:
> > On Wed, Dec 17, 2008 at 03:43:04PM -0600, Steve Suehs wrote:
> >> I'm in the Austin, TX area. Where do you want the beer?
> >
> > In Cork City, Ireland, please.
>
> Just goes to show - you can't tell anything from an email address! I
> never expected ".de" to be close to me!
Yay, another thirsty Irish friend! :)
Steve, we'll need at least one more barrel so!
David, to clear up the confusion:
I am in fact from Berlin, but currently studying in Cork.
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986823
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by David O'Shea <da...@s3group.com>.
On 18/12/2008 14:50, Stefan Sperling wrote:
> On Wed, Dec 17, 2008 at 03:43:04PM -0600, Steve Suehs wrote:
>> I'm in the Austin, TX area. Where do you want the beer?
>
> In Cork City, Ireland, please.
Just goes to show - you can't tell anything from an email address! I
never expected ".de" to be close to me!
David.
--
The information contained in this e-mail and in any attachments is confidential and is designated solely for the attention of the intended recipient(s). If you are not an intended recipient, you must not use, disclose, copy, distribute or retain this e-mail or any part thereof. If you have received this e-mail in error, please notify the sender by return e-mail and delete all copies of this e-mail from your computer system(s).
Please direct any additional queries to: communications@s3group.com.
Thank You.
Silicon and Software Systems Limited. Registered in Ireland no. 378073.
Registered Office: South County Business Park, Leopardstown, Dublin 18
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986722
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Stefan Sperling <st...@elego.de>.
On Wed, Dec 17, 2008 at 03:43:04PM -0600, Steve Suehs wrote:
> I'm in the Austin, TX area. Where do you want the beer?
In Cork City, Ireland, please.
Please supply enough quantities so I can share with all my
thirsty Irish friends!
42 large barrels OK for you?
Thanks,
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986709
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Steve Suehs <cs...@drillinginfo.com>.
I'm in the Austin, TX area. Where do you want the beer?
-s
On Wed, 2008-12-17 at 18:47 +0000, Stefan Sperling wrote:
> On Thu, Dec 11, 2008 at 03:03:12PM -0600, Steve Suehs wrote:
> > I don't like being the merge gimp. I have other things to do.
>
> I've created a proper backport of this change, including all
> related follow-up changes made by Julian, here:
> https://svn.collab.net/repos/svn/branches/1.5.x-r30756
>
> David, Julian, or anyone else, if you could add +1's to STATUS
> quickly we could still get this change into 1.5.5.
>
> Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986084
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Dec 11, 2008 at 03:03:12PM -0600, Steve Suehs wrote:
> I don't like being the merge gimp. I have other things to do.
I've created a proper backport of this change, including all
related follow-up changes made by Julian, here:
https://svn.collab.net/repos/svn/branches/1.5.x-r30756
David, Julian, or anyone else, if you could add +1's to STATUS
quickly we could still get this change into 1.5.5.
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985940
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Steve Suehs <cs...@drillinginfo.com>.
Thanks. We would really appreciate it. Right now, I'm the merge gimp
for the company because I'm the only one with a patched svn client.
We have multiple teams running into this, and I'm the only one who can
merge for them. While we are implementing a commit hook to prevent
further introduction of inconsistent eol's, we're still stuck with about
3 teams (5 branches) with the problem, both merging trunk out to branch
and just now, merging back into trunk.
I don't like being the merge gimp. I have other things to do.
-s
On Thu, 2008-12-11 at 12:56 -0800, David Glasser wrote:
> If that gets nominated, I'll vote for it (it was my patch), but I seem
> to recall failing to get anybody interested in reviewing it when I
> submitted it as a patch pre-commit, let alone backport review.
>
> --dave
>
> On Thu, Dec 11, 2008 at 11:54 AM, Steve Suehs <cs...@drillinginfo.com> wrote:
> > http://subversion.tigris.org/issues/show_bug.cgi?id=3262
> >
> >
> > --- subversion/libsvn_wc/merge.c (revision 34661)
> > +++ subversion/libsvn_wc/merge.c (working copy)
> > @@ -245,7 +245,7 @@
> > pool));
> > SVN_ERR(svn_subst_copy_and_translate3(old_target,
> > tmp_new,
> > - eol, eol ? FALSE : TRUE,
> > + eol, TRUE,
> > NULL, FALSE,
> > FALSE, pool));
> >
> >
> > On Wed, 2008-12-10 at 10:58 -0800, Hyrum K. Wright wrote:
> >> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> >> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> >> what's in the release, so if there are items in STATUS you want in the release,
> >> poke, prod, campaign or bribe people to vote on them.
> >>
> >> -Hyrum
> >>
> >> ------------------------------------------------------
> >> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
> >
> > ------------------------------------------------------
> > http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983006
> >
>
>
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983035
Re: [PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by David Glasser <gl...@davidglasser.net>.
If that gets nominated, I'll vote for it (it was my patch), but I seem
to recall failing to get anybody interested in reviewing it when I
submitted it as a patch pre-commit, let alone backport review.
--dave
On Thu, Dec 11, 2008 at 11:54 AM, Steve Suehs <cs...@drillinginfo.com> wrote:
> http://subversion.tigris.org/issues/show_bug.cgi?id=3262
>
>
> --- subversion/libsvn_wc/merge.c (revision 34661)
> +++ subversion/libsvn_wc/merge.c (working copy)
> @@ -245,7 +245,7 @@
> pool));
> SVN_ERR(svn_subst_copy_and_translate3(old_target,
> tmp_new,
> - eol, eol ? FALSE : TRUE,
> + eol, TRUE,
> NULL, FALSE,
> FALSE, pool));
>
>
> On Wed, 2008-12-10 at 10:58 -0800, Hyrum K. Wright wrote:
>> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
>> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
>> what's in the release, so if there are items in STATUS you want in the release,
>> poke, prod, campaign or bribe people to vote on them.
>>
>> -Hyrum
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983006
>
--
glasser@davidglasser.net | langtonlabs.org | flickr.com/photos/glasser/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983037
[PATCH] for 3262 Re: 1.5.5 next Wednesday (12/17)
Posted by Steve Suehs <cs...@drillinginfo.com>.
http://subversion.tigris.org/issues/show_bug.cgi?id=3262
--- subversion/libsvn_wc/merge.c (revision 34661)
+++ subversion/libsvn_wc/merge.c (working copy)
@@ -245,7 +245,7 @@
pool));
SVN_ERR(svn_subst_copy_and_translate3(old_target,
tmp_new,
- eol, eol ? FALSE : TRUE,
+ eol, TRUE,
NULL, FALSE,
FALSE, pool));
On Wed, 2008-12-10 at 10:58 -0800, Hyrum K. Wright wrote:
> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> what's in the release, so if there are items in STATUS you want in the release,
> poke, prod, campaign or bribe people to vote on them.
>
> -Hyrum
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=983006
Re: 1.5.5 next Wednesday (12/17)
Posted by David Glasser <gl...@davidglasser.net>.
* r34506
Fix a pair of bugs in the BDB backend's node-origins cache-miss fallback
logic, one of which can cause a SEGFAULT, the other of which can result
in an incorrect return value.
Votes:
+1: cmpilato, glasser
Segfaults! Data corruption! Relatively small changes (one of which
is just "do the thing that the comment says it does")! Perhaps
somebody should review and vote for it!
--dave
On Wed, Dec 10, 2008 at 10:58 AM, Hyrum K. Wright
<hy...@mail.utexas.edu> wrote:
> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> what's in the release, so if there are items in STATUS you want in the release,
> poke, prod, campaign or bribe people to vote on them.
>
> -Hyrum
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
--
glasser@davidglasser.net | langtonlabs.org | flickr.com/photos/glasser/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982580
Re: 1.5.5 next Wednesday (12/17)
Posted by Paul Burba <pt...@gmail.com>.
On Thu, Dec 11, 2008 at 9:44 AM, Michael Susser <hd...@googlemail.com> wrote:
> I really would like to see r34432 in 1.5.5 :-) Anyone else?
Michael,
I know Kamesh is looking at this. I'm working on a regression test
(which honestly I should have done already) to better demonstrate the
problem.
Paul
> Regards,
> Michael
>
> * r34432
> Follow-up to r32975, which is already backported and had a bug, see
> http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=145757.
> Notes:
> There is a very minor whitespace conflict in merge_props_changed().
> Justification:
> r32975 fixed a bug which occurred when a merge added explicit mergeinfo
> to a subtree that had no explicit mergeinfo prior to the merge. But
> this fix itself had a bug which could result in an error during a
> --dry-run merge, r34432 fixes this.
> Votes:
> +1: pburba
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982864
Re: 1.5.5 next Wednesday (12/17)
Posted by Michael Susser <hd...@googlemail.com>.
I really would like to see r34432 in 1.5.5 :-) Anyone else?
Regards,
Michael
* r34432
Follow-up to r32975, which is already backported and had a bug, see
http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=145757.
Notes:
There is a very minor whitespace conflict in merge_props_changed().
Justification:
r32975 fixed a bug which occurred when a merge added explicit mergeinfo
to a subtree that had no explicit mergeinfo prior to the merge. But
this fix itself had a bug which could result in an error during a
--dry-run merge, r34432 fixes this.
Votes:
+1: pburba
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982861
Re: 1.5.5 on 12/18 (was Re: 1.5.5 next Wednesday (12/17))
Posted by Julian Foad <ju...@btopenworld.com>.
Kamesh Jayachandran wrote:
> Hyrum K. Wright wrote:
> > Postponing 'til tomorrow (12/18) for a couple of reasons:
> > * A Ruby bindings test failure
> > (see http://svn.haxx.se/dev/archive-2008-12/0500.shtml)
> > * Allow people more time to review and vote for the reintegrate-improvements
> > branch. It only lacks one vote, and last I knew, several people were looking
> > at it. If you are one of those people, please let us know.
> >
> > Assuming all this gets taken care of, I'll roll tomorrow around 12pm CST (1800 UTC).
> >
>
> I am in between the review of reintegrate-improvements.
I have been unable to comprehend it all. The parts I've understood look
OK. I tried it out and found a funny error report:
$ svn merge --reintegrate $SVN_REPOS/branches/1.5.x-reintegrate-improvements .
/home/julianfoad/src/subversion-grey/subversion/libsvn_ra_neon/util.c:592: (apr_err=175002)
svn: REPORT of '/repos/svn/!svn/bc/34802': 200 OK (http://svn.collab.net)
which Paul is looking at now. Mail thread:
<http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986670>.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986712
Re: 1.5.5 on 12/18 (was Re: 1.5.5 next Wednesday (12/17))
Posted by Kamesh Jayachandran <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hyrum K. Wright wrote:
> Hyrum K. Wright wrote:
>> Hyrum K. Wright wrote:
>>> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
>>> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
>>> what's in the release, so if there are items in STATUS you want in the release,
>>> poke, prod, campaign or bribe people to vote on them.
>> Status update: still on track for rolling 1.5.5 tomorrow. There are currently a
>> number of merge-related issues still in STATUS which only lack one vote. Folks
>> familiar with that code are encouraged to review. (Those that aren't are
>> encouraged to become so. :) )
>
> Postponing 'til tomorrow (12/18) for a couple of reasons:
> * A Ruby bindings test failure
> (see http://svn.haxx.se/dev/archive-2008-12/0500.shtml)
> * Allow people more time to review and vote for the reintegrate-improvements
> branch. It only lacks one vote, and last I knew, several people were looking
> at it. If you are one of those people, please let us know.
>
> Assuming all this gets taken care of, I'll roll tomorrow around 12pm CST (1800 UTC).
>
I am in between the review of reintegrate-improvements.
With regards
Kamesh Jayachandran
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD4DBQFJSfV+3WHvyO0YTCwRAr2RAJiREs2bSD2VhYGvX0Wmvx3AYUUnAJ46B6EZ
VT7OCEXh72vuhm1oB//Czw==
=RM2h
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986362
1.5.5 on 12/18 (was Re: 1.5.5 next Wednesday (12/17))
Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
Hyrum K. Wright wrote:
> Hyrum K. Wright wrote:
>> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
>> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
>> what's in the release, so if there are items in STATUS you want in the release,
>> poke, prod, campaign or bribe people to vote on them.
>
> Status update: still on track for rolling 1.5.5 tomorrow. There are currently a
> number of merge-related issues still in STATUS which only lack one vote. Folks
> familiar with that code are encouraged to review. (Those that aren't are
> encouraged to become so. :) )
Postponing 'til tomorrow (12/18) for a couple of reasons:
* A Ruby bindings test failure
(see http://svn.haxx.se/dev/archive-2008-12/0500.shtml)
* Allow people more time to review and vote for the reintegrate-improvements
branch. It only lacks one vote, and last I knew, several people were looking
at it. If you are one of those people, please let us know.
Assuming all this gets taken care of, I'll roll tomorrow around 12pm CST (1800 UTC).
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=986298
Re: Please review for backport the issue #3275 changes (Was: Re:
1.5.5 next Wednesday (12/17))
Posted by Julian Foad <ju...@btopenworld.com>.
On Tue, 2008-12-16 at 10:25 -0500, Mark Phippard wrote:
> On Tue, Dec 16, 2008 at 10:21 AM, C. Michael Pilato <cm...@collab.net> wrote:
> > Konstantin Kolinko wrote:
> >> 2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
> >>> Hyrum K. Wright wrote:
> >>> Status update: still on track for rolling 1.5.5 tomorrow.
> >>>
> >>
> >> Will the fix for issue #3275 be included in 1.5.5?
> >> It is a P1 one.
> >
> > I agree that this is an important bugfix. Unfortunately, if we don't get
> > other eyes on it soon, it will not be included in the release.
>
> I do not think we should do the release until all of the issues in
> STATUS are cleared. I do not mean that as a general policy statement,
> just that of the bugs that are still in STATUS, several of them are P1
> nature. The --reintegrate bug fix goes a long way towards solving the
> #1 problem with the 1.5 release. I cannot see why we would do another
> bug fix release without this fix now that it exists.
Mark,
I'll do some reviewing now.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985094
Re: Please review for backport the issue #3275 changes (Was: Re:
1.5.5 next Wednesday (12/17))
Posted by David Ferguson <fe...@gmail.com>.
As an admin of a mutli-site SVN shop, please consider including the fix to
#3275 in 1.5.5. We are stuck on 1.5.0!
thanks,
David
On Tue, Dec 16, 2008 at 10:42 AM, Hyrum K. Wright <
hyrum_wright@mail.utexas.edu> wrote:
> Mark Phippard wrote:
> > On Tue, Dec 16, 2008 at 10:21 AM, C. Michael Pilato <cm...@collab.net>
> wrote:
> >> Konstantin Kolinko wrote:
> >>> 2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
> >>>> Hyrum K. Wright wrote:
> >>>> Status update: still on track for rolling 1.5.5 tomorrow.
> >>>>
> >>> Will the fix for issue #3275 be included in 1.5.5?
> >>> It is a P1 one.
> >> I agree that this is an important bugfix. Unfortunately, if we don't
> get
> >> other eyes on it soon, it will not be included in the release.
> >
> > I do not think we should do the release until all of the issues in
> > STATUS are cleared. I do not mean that as a general policy statement,
> > just that of the bugs that are still in STATUS, several of them are P1
> > nature. The --reintegrate bug fix goes a long way towards solving the
> > #1 problem with the 1.5 release. I cannot see why we would do another
> > bug fix release without this fix now that it exists.
>
> If you want specific revisions in the next patch release, then encourage
> people
> to review the items in STATUS.
>
> The entire point of doing time-based releases (patch or feature) is that we
> can
> just release what we've got when the time hits. There are always
> exceptions to
> the rule, but I think they should be few and far between. We need to think
> hard
> before making them. If developers (myself included) aren't willing to
> review
> items which have a known impact to our user base, we've got deeper issues
> at
> play, and no amount of postponement will fix them.
>
> It's been no secret for the *last month* that 1.5.5 was rolling on 12/17.
> The
> reintegrate improvements branch was nominated on 12/2, two weeks ago. I
> find it
> difficult to believe that postponing the release an indefinite amount of
> time
> will speed review, when nothing has happened on that item in the past two
> weeks.
> (To the contrary, actually, I find that hard release dates actually
> encourage
> increased review.)
>
> In the end, I'm just a reflection of the community, and enough people say
> "hold
> the release", I'll do so. But know that I'm taking that action under
> protest. :)
>
> -Hyrum
>
> PS - Mark, nice to see you've got power back.
>
> ------------------------------------------------------
>
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985046
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985054
Re: Please review for backport the issue #3275 changes (Was: Re:
1.5.5 next Wednesday (12/17))
Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
Mark Phippard wrote:
> On Tue, Dec 16, 2008 at 10:21 AM, C. Michael Pilato <cm...@collab.net> wrote:
>> Konstantin Kolinko wrote:
>>> 2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
>>>> Hyrum K. Wright wrote:
>>>> Status update: still on track for rolling 1.5.5 tomorrow.
>>>>
>>> Will the fix for issue #3275 be included in 1.5.5?
>>> It is a P1 one.
>> I agree that this is an important bugfix. Unfortunately, if we don't get
>> other eyes on it soon, it will not be included in the release.
>
> I do not think we should do the release until all of the issues in
> STATUS are cleared. I do not mean that as a general policy statement,
> just that of the bugs that are still in STATUS, several of them are P1
> nature. The --reintegrate bug fix goes a long way towards solving the
> #1 problem with the 1.5 release. I cannot see why we would do another
> bug fix release without this fix now that it exists.
If you want specific revisions in the next patch release, then encourage people
to review the items in STATUS.
The entire point of doing time-based releases (patch or feature) is that we can
just release what we've got when the time hits. There are always exceptions to
the rule, but I think they should be few and far between. We need to think hard
before making them. If developers (myself included) aren't willing to review
items which have a known impact to our user base, we've got deeper issues at
play, and no amount of postponement will fix them.
It's been no secret for the *last month* that 1.5.5 was rolling on 12/17. The
reintegrate improvements branch was nominated on 12/2, two weeks ago. I find it
difficult to believe that postponing the release an indefinite amount of time
will speed review, when nothing has happened on that item in the past two weeks.
(To the contrary, actually, I find that hard release dates actually encourage
increased review.)
In the end, I'm just a reflection of the community, and enough people say "hold
the release", I'll do so. But know that I'm taking that action under protest. :)
-Hyrum
PS - Mark, nice to see you've got power back.
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985046
Re: Please review for backport the issue #3275 changes (Was: Re:
1.5.5 next Wednesday (12/17))
Posted by Mark Phippard <ma...@gmail.com>.
On Tue, Dec 16, 2008 at 10:21 AM, C. Michael Pilato <cm...@collab.net> wrote:
> Konstantin Kolinko wrote:
>> 2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
>>> Hyrum K. Wright wrote:
>>> Status update: still on track for rolling 1.5.5 tomorrow.
>>>
>>
>> Will the fix for issue #3275 be included in 1.5.5?
>> It is a P1 one.
>
> I agree that this is an important bugfix. Unfortunately, if we don't get
> other eyes on it soon, it will not be included in the release.
I do not think we should do the release until all of the issues in
STATUS are cleared. I do not mean that as a general policy statement,
just that of the bugs that are still in STATUS, several of them are P1
nature. The --reintegrate bug fix goes a long way towards solving the
#1 problem with the 1.5 release. I cannot see why we would do another
bug fix release without this fix now that it exists.
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985039
Please review for backport the issue #3275 changes (Was: Re: 1.5.5
next Wednesday (12/17))
Posted by "C. Michael Pilato" <cm...@collab.net>.
Konstantin Kolinko wrote:
> 2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
>> Hyrum K. Wright wrote:
>> Status update: still on track for rolling 1.5.5 tomorrow.
>>
>
> Will the fix for issue #3275 be included in 1.5.5?
> It is a P1 one.
I agree that this is an important bugfix. Unfortunately, if we don't get
other eyes on it soon, it will not be included in the release.
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985037
Re: 1.5.5 next Wednesday (12/17)
Posted by Konstantin Kolinko <kn...@gmail.com>.
2008/12/16 Hyrum K. Wright <hy...@mail.utexas.edu>:
> Hyrum K. Wright wrote:
> Status update: still on track for rolling 1.5.5 tomorrow.
>
Will the fix for issue #3275 be included in 1.5.5?
It is a P1 one.
Best regards,
Konstantin Kolinko
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985035
Re: patch for 3262 (Re: 1.5.5 next Wednesday (12/17))
Posted by David Glasser <gl...@davidglasser.net>.
Sorry, I am barely an active developer these days... my participation
has been limited to "making sure bugs aren't added to FSFS" and "vote
for things when somebody pokes me to". Doesn't look like this has
been backported yet... can somebody else give it the review that it
failed to get even pre-commit so we can backport it?
This bug basically prevents you from doing merges in some situations
involving svn:eol-style; the fix appears to allow them.
--dave
On Tue, Dec 16, 2008 at 9:24 AM, Steve Suehs <cs...@drillinginfo.com> wrote:
> David Glasser,
>
> Do we need to convince a third person to review this one-line patch, or
> is it everyone? How do we get this nominated?
>
> -s
>
> On Tue, 2008-12-16 at 08:10 -0600, Hyrum K. Wright wrote:
>> Hyrum K. Wright wrote:
>> > 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
>> > roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
>> > what's in the release, so if there are items in STATUS you want in the release,
>> > poke, prod, campaign or bribe people to vote on them.
>>
>> Status update: still on track for rolling 1.5.5 tomorrow. There are currently a
>> number of merge-related issues still in STATUS which only lack one vote. Folks
>> familiar with that code are encouraged to review. (Those that aren't are
>> encouraged to become so. :) )
>>
>> -Hyrum
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=984993
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985110
>
--
glasser@davidglasser.net | langtonlabs.org | flickr.com/photos/glasser/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985219
patch for 3262 (Re: 1.5.5 next Wednesday (12/17))
Posted by Steve Suehs <cs...@drillinginfo.com>.
David Glasser,
Do we need to convince a third person to review this one-line patch, or
is it everyone? How do we get this nominated?
-s
On Tue, 2008-12-16 at 08:10 -0600, Hyrum K. Wright wrote:
> Hyrum K. Wright wrote:
> > 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> > roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> > what's in the release, so if there are items in STATUS you want in the release,
> > poke, prod, campaign or bribe people to vote on them.
>
> Status update: still on track for rolling 1.5.5 tomorrow. There are currently a
> number of merge-related issues still in STATUS which only lack one vote. Folks
> familiar with that code are encouraged to review. (Those that aren't are
> encouraged to become so. :) )
>
> -Hyrum
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=984993
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=985110
Re: 1.5.5 next Wednesday (12/17)
Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
Hyrum K. Wright wrote:
> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> what's in the release, so if there are items in STATUS you want in the release,
> poke, prod, campaign or bribe people to vote on them.
Status update: still on track for rolling 1.5.5 tomorrow. There are currently a
number of merge-related issues still in STATUS which only lack one vote. Folks
familiar with that code are encouraged to review. (Those that aren't are
encouraged to become so. :) )
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=984993
RE: 1.5.5 next Wednesday (12/17)
Posted by Julian Foad <ju...@btopenworld.com>.
Bert, that looks like useful info. Please could you put it in a new
email thread or in the issue tracker. I think this is (part of) issue
#3119 <http://subversion.tigris.org/issues/show_bug.cgi?id=3119>, also
referenced recently in email thread
<http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982066> "Maven scm tag does not work with Subversion 1.5.x".
- Julian
On Thu, 2008-12-11 at 01:33 +0100, Bert Huijben wrote:
> > -----Original Message-----
> > From: Mark Phippard [mailto:markphip@gmail.com]
> > Sent: Thursday, December 11, 2008 12:13 AM
> > To: kmradke@rockwellcollins.com
> > Cc: dev@subversion.tigris.org
> > Subject: Re: 1.5.5 next Wednesday (12/17)
> >
> > On Wed, Dec 10, 2008 at 6:08 PM, <km...@rockwellcollins.com> wrote:
> > >
> > > Anyone know if the 1.5 regression that causes a:
> > >
> > > svn cp WC URL
> > >
> > > operation to fail with a "file already exists" error
> > > when performed over http when WC has modified files has
> > > been fixed? I'm not seeing any closure to this thread:
> > >
> > > http://svn.haxx.se/users/archive-2008-10/0829.shtml
> > >
> > > I didn't see an issue being filed or anything in 1.5.x STATUS,
> > > but I have run into this particular problem recently and it
> > > worked in 1.4 (and possibly as late as 1.5.1)
> >
> > AFAIK, even 1.6 is in danger of shipping with that bug still in place.
> > I have not seen anyone attempt to fix it.
>
> For other developers looking at this issue:
>
> r31692 (which was merged to 1.5.x in r31799) is the only change between
> 1.5.0 and 1.5.1 in neon that can explain the regression starting in 1.5.1.
> It touches precisely the code Karl quotes in his mail.
>
> Modified: trunk/subversion/libsvn_ra_neon/commit.c
> URL:
> http://svn.collab.net/viewvc/svn/trunk/subversion/libsvn_ra_neon/commit.c?pa
> threv=31692&r1=31691&r2=31692
> ============================================================================
> ==
> --- trunk/subversion/libsvn_ra_neon/commit.c Tue Jun 10 12:59:16 2008
> (r31691)
> +++ trunk/subversion/libsvn_ra_neon/commit.c Tue Jun 10 13:20:28 2008
> (r31692)
> @@ -1036,8 +1036,8 @@ static svn_error_t * commit_add_file(con
> svn_ra_neon__resource_t *res;
> svn_error_t *err = svn_ra_neon__get_starting_props(&res,
> file->cc->ras,
> - file->rsrc->url,
> NULL,
> - workpool);
> +
> file->rsrc->wr_url,
> + NULL, workpool);
> if (!err)
> {
> /* If the PROPFIND succeeds the file already exists */
>
> =====================================================================
> [[
> Fix issue #2939 over ra_neon: bogus existence check.
>
> See reproduction recipe in the issue to verify this fix -- it uses
> svnmucc. Scenario can't be tested using the normal 'svn' client,
> since it involves doing a directory replacement within a single
> transaction. svnmucc and gvn can create the scenario, however,
> resulting is a revision which (until now) couldn't be svnsync'd over
> http!
>
> * subversion/libsvn_ra_neon/commit.c (commit_add_file):
> do PROPFIND existence check on the *working* url (i.e. the path
> within the transaction), not the public url (which the path in HEAD).
> ]]
>
> [[
> Merge r31692, 31695 from trunk:
>
> * r31692, r31695
> Fix issue #2939 over ra_neon/serf: bogus existence check.
> Votes:
> +1: epg, lgo, Arfrever
> ]]
>
> I would guess that reverting this change will bring us back to the old
> behavior, but it will break issue #2939.
>
>
> Karl's diagnosis continues in mod_dav_svn, but the comment above this block
> of code says:
>
> /* If the parent directory existed before this commit then there may be a
> file with this URL already. We need to ensure such a file does not
> exist, which we do by attempting a PROPFIND. Of course, a
> PROPFIND *should* succeed if this "add" is actually the second
> half of a "replace".
>
> [....]
>
> Looking at this neon change, it looks like it starts checking whether the
> file exists in the transaction and stops checking whether it exists before
> the commit. (I'm trying to understand what the consequences are, but the old
> behavior matches what I would expect reading the comments around this code).
>
> The serf change for the same issue is completely different than the neon
> version. (I don't think r31695 needs further investigation)
>
> Bert
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982800
RE: 1.5.5 next Wednesday (12/17)
Posted by Bert Huijben <be...@vmoo.com>.
> -----Original Message-----
> From: Mark Phippard [mailto:markphip@gmail.com]
> Sent: Thursday, December 11, 2008 12:13 AM
> To: kmradke@rockwellcollins.com
> Cc: dev@subversion.tigris.org
> Subject: Re: 1.5.5 next Wednesday (12/17)
>
> On Wed, Dec 10, 2008 at 6:08 PM, <km...@rockwellcollins.com> wrote:
> >
> > Anyone know if the 1.5 regression that causes a:
> >
> > svn cp WC URL
> >
> > operation to fail with a "file already exists" error
> > when performed over http when WC has modified files has
> > been fixed? I'm not seeing any closure to this thread:
> >
> > http://svn.haxx.se/users/archive-2008-10/0829.shtml
> >
> > I didn't see an issue being filed or anything in 1.5.x STATUS,
> > but I have run into this particular problem recently and it
> > worked in 1.4 (and possibly as late as 1.5.1)
>
> AFAIK, even 1.6 is in danger of shipping with that bug still in place.
> I have not seen anyone attempt to fix it.
For other developers looking at this issue:
r31692 (which was merged to 1.5.x in r31799) is the only change between
1.5.0 and 1.5.1 in neon that can explain the regression starting in 1.5.1.
It touches precisely the code Karl quotes in his mail.
Modified: trunk/subversion/libsvn_ra_neon/commit.c
URL:
http://svn.collab.net/viewvc/svn/trunk/subversion/libsvn_ra_neon/commit.c?pa
threv=31692&r1=31691&r2=31692
============================================================================
==
--- trunk/subversion/libsvn_ra_neon/commit.c Tue Jun 10 12:59:16 2008
(r31691)
+++ trunk/subversion/libsvn_ra_neon/commit.c Tue Jun 10 13:20:28 2008
(r31692)
@@ -1036,8 +1036,8 @@ static svn_error_t * commit_add_file(con
svn_ra_neon__resource_t *res;
svn_error_t *err = svn_ra_neon__get_starting_props(&res,
file->cc->ras,
- file->rsrc->url,
NULL,
- workpool);
+
file->rsrc->wr_url,
+ NULL, workpool);
if (!err)
{
/* If the PROPFIND succeeds the file already exists */
=====================================================================
[[
Fix issue #2939 over ra_neon: bogus existence check.
See reproduction recipe in the issue to verify this fix -- it uses
svnmucc. Scenario can't be tested using the normal 'svn' client,
since it involves doing a directory replacement within a single
transaction. svnmucc and gvn can create the scenario, however,
resulting is a revision which (until now) couldn't be svnsync'd over
http!
* subversion/libsvn_ra_neon/commit.c (commit_add_file):
do PROPFIND existence check on the *working* url (i.e. the path
within the transaction), not the public url (which the path in HEAD).
]]
[[
Merge r31692, 31695 from trunk:
* r31692, r31695
Fix issue #2939 over ra_neon/serf: bogus existence check.
Votes:
+1: epg, lgo, Arfrever
]]
I would guess that reverting this change will bring us back to the old
behavior, but it will break issue #2939.
Karl's diagnosis continues in mod_dav_svn, but the comment above this block
of code says:
/* If the parent directory existed before this commit then there may be a
file with this URL already. We need to ensure such a file does not
exist, which we do by attempting a PROPFIND. Of course, a
PROPFIND *should* succeed if this "add" is actually the second
half of a "replace".
[....]
Looking at this neon change, it looks like it starts checking whether the
file exists in the transaction and stops checking whether it exists before
the commit. (I'm trying to understand what the consequences are, but the old
behavior matches what I would expect reading the comments around this code).
The serf change for the same issue is completely different than the neon
version. (I don't think r31695 needs further investigation)
Bert
>
> --
> Thanks
>
> Mark Phippard
> http://markphip.blogspot.com/
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageI
> d=982600
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982623
Re: 1.5.5 next Wednesday (12/17)
Posted by Mark Phippard <ma...@gmail.com>.
On Wed, Dec 10, 2008 at 6:08 PM, <km...@rockwellcollins.com> wrote:
>
> Anyone know if the 1.5 regression that causes a:
>
> svn cp WC URL
>
> operation to fail with a "file already exists" error
> when performed over http when WC has modified files has
> been fixed? I'm not seeing any closure to this thread:
>
> http://svn.haxx.se/users/archive-2008-10/0829.shtml
>
> I didn't see an issue being filed or anything in 1.5.x STATUS,
> but I have run into this particular problem recently and it
> worked in 1.4 (and possibly as late as 1.5.1)
AFAIK, even 1.6 is in danger of shipping with that bug still in place.
I have not seen anyone attempt to fix it.
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982600
Re: 1.5.5 next Wednesday (12/17)
Posted by km...@rockwellcollins.com.
Anyone know if the 1.5 regression that causes a:
svn cp WC URL
operation to fail with a "file already exists" error
when performed over http when WC has modified files has
been fixed? I'm not seeing any closure to this thread:
http://svn.haxx.se/users/archive-2008-10/0829.shtml
I didn't see an issue being filed or anything in 1.5.x STATUS,
but I have run into this particular problem recently and it
worked in 1.4 (and possibly as late as 1.5.1)
Kevin R.
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982595
Re: 1.5.5 next Wednesday (12/17)
Posted by Steve Suehs <cs...@drillinginfo.com>.
May we please have a fix for
http://subversion.tigris.org/issues/show_bug.cgi?id=3262
in subversion 1.5.5?
It appears that an application of r30756 on the trunk to the 1.5.x
branch would fix it. I admit to not having run regression tests.
--- subversion/libsvn_wc/merge.c (revision 34661)
+++ subversion/libsvn_wc/merge.c (working copy)
@@ -245,7 +245,7 @@
pool));
SVN_ERR(svn_subst_copy_and_translate3(old_target,
tmp_new,
- eol, eol ? FALSE : TRUE,
+ eol, TRUE,
NULL, FALSE,
FALSE, pool));
On Wed, 2008-12-10 at 10:58 -0800, Hyrum K. Wright wrote:
> 1.5.x STATUS is still pretty full, but I'm going to move forward with plans to
> roll 1.5.5 next Wednesday, Dec. 17. What's been merged by that time will be
> what's in the release, so if there are items in STATUS you want in the release,
> poke, prod, campaign or bribe people to vote on them.
>
> -Hyrum
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982393
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982573