Next 1.6 release?

[kmradke-2]

I'd like to be a little selfish and wonder if it would be useful to release a
1.6.7 version faster than normal.  (Especially since the release schedule
may be affected by the amount of holiday time near the end of the year.)

We are really getting hurt by the external problems that were fixed in r40152
and merged into 1.6.x in r40219.  Without this change we can't move a large
number of our users from 1.5, and I would rather not be forced to compile
custom patched versions for all the clients we use.
(command line x 4 platforms, TortoiseSVN, Subclipse)

Looking at STATUS, there are some other things already nominated (but need
votes!!!) that also seem useful:

* Fix for #3432
* Win 7 exception handler support
* Fix for #3489
* Win32 performance improvements
* future-proof patch
* Fix for #3519

I don't see it nominated, but the svn+ssh changes for windows may also
fix some server resource problems others were seeing (We don't use svn+ssh,
so I can't comment.)

Is it too soon to be talking about this?

Thanks!
Kevin R.

[Julian Foad]

kmradke@... wrote:

> I'd like to be a little selfish and wonder if it would be useful to release a
> 1.6.7 version faster than normal.  (Especially since the release schedule
> may be affected by the amount of holiday time near the end of the year.)
>
> We are really getting hurt by the external problems that were fixed in r40152
> and merged into 1.6.x in r40219.  Without this change we can't move a large
> number of our users from 1.5, and I would rather not be forced to compile
> custom patched versions for all the clients we use.
> (command line x 4 platforms, TortoiseSVN, Subclipse)
>
> Looking at STATUS, there are some other things already nominated (but need
> votes!!!) that also seem useful:
>
> * Fix for #3432
> * Win 7 exception handler support
> * Fix for #3489
> * Win32 performance improvements
> * future-proof patch
> * Fix for #3519

For those who don't have a numerically-keyed hash of issues in their
brain:

* Fix for #3432:
  "Merge can record mergeinfo from natural history gaps"

* Fix for #3489:
  "Filenames with @ cannot be committed to the repository unless an @ is
added at the end."

* future-proof patch
  "Make 1.6.7 (and subsequent Subversion releases) recognize Subversion
1.7+ working copies."

* Fix for #3519:
  "ra_serf using Label header unsupported by mod_dav_svn"


> I don't see it nominated, but the svn+ssh changes for windows may also
> fix some server resource problems others were seeing (We don't use svn+ssh,
> so I can't comment.)
>
> Is it too soon to be talking about this?

+1 to the general notion of getting a 1.6.7 release out as soon as we
have some useful fixes in it.

- Julian

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2414453

[Hyrum K. Wright-4]

On Nov 4, 2009, at 8:48 AM, Julian Foad wrote:

> kmradke@... wrote:
>> I'd like to be a little selfish and wonder if it would be useful to  
>> release a
>> 1.6.7 version faster than normal.  (Especially since the release  
>> schedule
>> may be affected by the amount of holiday time near the end of the  
>> year.)
>>
>> We are really getting hurt by the external problems that were fixed  
>> in r40152
>> and merged into 1.6.x in r40219.  Without this change we can't move  
>> a large
>> number of our users from 1.5, and I would rather not be forced to  
>> compile
>> custom patched versions for all the clients we use.
>> (command line x 4 platforms, TortoiseSVN, Subclipse)
>>
>> Looking at STATUS, there are some other things already nominated  
>> (but need
>> votes!!!) that also seem useful:
>>
>> * Fix for #3432
>> * Win 7 exception handler support
>> * Fix for #3489
>> * Win32 performance improvements
>> * future-proof patch
>> * Fix for #3519
>
> For those who don't have a numerically-keyed hash of issues in their
> brain:
>
> * Fix for #3432:
>  "Merge can record mergeinfo from natural history gaps"
>
> * Fix for #3489:
>  "Filenames with @ cannot be committed to the repository unless an @  
> is
> added at the end."
>
> * future-proof patch
>  "Make 1.6.7 (and subsequent Subversion releases) recognize Subversion
> 1.7+ working copies."
>
> * Fix for #3519:
>  "ra_serf using Label header unsupported by mod_dav_svn"
>
>
>> I don't see it nominated, but the svn+ssh changes for windows may  
>> also
>> fix some server resource problems others were seeing (We don't use  
>> svn+ssh,
>> so I can't comment.)
>>
>> Is it too soon to be talking about this?
>
> +1 to the general notion of getting a 1.6.7 release out as soon as we
> have some useful fixes in it.

I've not yet looked at the the stuff that's currently been merged for  
1.6.7, but there are several items in STATUS that are worth a  
release.  I'm hesitant to do one *too* quickly (<4 weeks), but I'm  
happy to cut a release in the next few weeks if the demand is there,  
and it sounds like it is.

As usual, the roadblock for getting stuff into the release is the  
review needed in STATUS.  Please take some time to take a look.

-Hyrum

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2414481

[Bob Denny]

Kevin R:
> I don't see it nominated, but the svn+ssh changes for windows may also
> fix some server resource problems others were seeing (We don't use svn+ssh,
> so I can't comment.)

I do, and I can comment :-) Please consider this for the "quick-release" as
well. It has been running here and a couple of other sites with no problems,
after severe problems with accumulating remote processes without the patch.

  -- Bob

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2414948

[Greg Stein-4]

On Wed, Nov 4, 2009 at 12:45, Hyrum K. Wright <hyrum@...> wrote:

> On Nov 4, 2009, at 8:48 AM, Julian Foad wrote:
>
>> kmradke@... wrote:
>>> I'd like to be a little selfish and wonder if it would be useful to
>>> release a
>>> 1.6.7 version faster than normal.  (Especially since the release
>>> schedule
>>> may be affected by the amount of holiday time near the end of the
>>> year.)
>>>
>>> We are really getting hurt by the external problems that were fixed
>>> in r40152
>>> and merged into 1.6.x in r40219.  Without this change we can't move
>>> a large
>>> number of our users from 1.5, and I would rather not be forced to
>>> compile
>>> custom patched versions for all the clients we use.
>>> (command line x 4 platforms, TortoiseSVN, Subclipse)
>>>
>>> Looking at STATUS, there are some other things already nominated
>>> (but need
>>> votes!!!) that also seem useful:
>>>
>>> * Fix for #3432
>>> * Win 7 exception handler support
>>> * Fix for #3489
>>> * Win32 performance improvements
>>> * future-proof patch
>>> * Fix for #3519
>>
>> For those who don't have a numerically-keyed hash of issues in their
>> brain:
>>
>> * Fix for #3432:
>>  "Merge can record mergeinfo from natural history gaps"
>>
>> * Fix for #3489:
>>  "Filenames with @ cannot be committed to the repository unless an @
>> is
>> added at the end."
>>
>> * future-proof patch
>>  "Make 1.6.7 (and subsequent Subversion releases) recognize Subversion
>> 1.7+ working copies."
>>
>> * Fix for #3519:
>>  "ra_serf using Label header unsupported by mod_dav_svn"
>>
>>
>>> I don't see it nominated, but the svn+ssh changes for windows may
>>> also
>>> fix some server resource problems others were seeing (We don't use
>>> svn+ssh,
>>> so I can't comment.)
>>>
>>> Is it too soon to be talking about this?
>>
>> +1 to the general notion of getting a 1.6.7 release out as soon as we
>> have some useful fixes in it.
>
> I've not yet looked at the the stuff that's currently been merged for
> 1.6.7, but there are several items in STATUS that are worth a
> release.  I'm hesitant to do one *too* quickly (<4 weeks), but I'm
> happy to cut a release in the next few weeks if the demand is there,
> and it sounds like it is.
>
> As usual, the roadblock for getting stuff into the release is the
> review needed in STATUS.  Please take some time to take a look.

In the Apache Incubator, there has been a request for us to produce a
release. I'm going to point them at the 1.6.7 release process, so I'd
like to request that it starts in (say) a couple weeks.

Cheers,
-g

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415010

[Arfrever Frehtes Taifersar Arahesis-2]

2009-11-06 08:41:05 Greg Stein napisał(a):

> On Wed, Nov 4, 2009 at 12:45, Hyrum K. Wright <hyrum@...> wrote:
> > On Nov 4, 2009, at 8:48 AM, Julian Foad wrote:
> >
> >> kmradke@... wrote:
> >>> I'd like to be a little selfish and wonder if it would be useful to
> >>> release a
> >>> 1.6.7 version faster than normal.  (Especially since the release
> >>> schedule
> >>> may be affected by the amount of holiday time near the end of the
> >>> year.)
> >>>
> >>> We are really getting hurt by the external problems that were fixed
> >>> in r40152
> >>> and merged into 1.6.x in r40219.  Without this change we can't move
> >>> a large
> >>> number of our users from 1.5, and I would rather not be forced to
> >>> compile
> >>> custom patched versions for all the clients we use.
> >>> (command line x 4 platforms, TortoiseSVN, Subclipse)
> >>>
> >>> Looking at STATUS, there are some other things already nominated
> >>> (but need
> >>> votes!!!) that also seem useful:
> >>>
> >>> * Fix for #3432
> >>> * Win 7 exception handler support
> >>> * Fix for #3489
> >>> * Win32 performance improvements
> >>> * future-proof patch
> >>> * Fix for #3519
> >>
> >> For those who don't have a numerically-keyed hash of issues in their
> >> brain:
> >>
> >> * Fix for #3432:
> >>  "Merge can record mergeinfo from natural history gaps"
> >>
> >> * Fix for #3489:
> >>  "Filenames with @ cannot be committed to the repository unless an @
> >> is
> >> added at the end."
> >>
> >> * future-proof patch
> >>  "Make 1.6.7 (and subsequent Subversion releases) recognize Subversion
> >> 1.7+ working copies."
> >>
> >> * Fix for #3519:
> >>  "ra_serf using Label header unsupported by mod_dav_svn"
> >>
> >>
> >>> I don't see it nominated, but the svn+ssh changes for windows may
> >>> also
> >>> fix some server resource problems others were seeing (We don't use
> >>> svn+ssh,
> >>> so I can't comment.)
> >>>
> >>> Is it too soon to be talking about this?
> >>
> >> +1 to the general notion of getting a 1.6.7 release out as soon as we
> >> have some useful fixes in it.
> >
> > I've not yet looked at the the stuff that's currently been merged for
> > 1.6.7, but there are several items in STATUS that are worth a
> > release.  I'm hesitant to do one *too* quickly (<4 weeks), but I'm
> > happy to cut a release in the next few weeks if the demand is there,
> > and it sounds like it is.
> >
> > As usual, the roadblock for getting stuff into the release is the
> > review needed in STATUS.  Please take some time to take a look.
>
> In the Apache Incubator, there has been a request for us to produce a
> release. I'm going to point them at the 1.6.7 release process, so I'd
> like to request that it starts in (say) a couple weeks.

Change of license probably needs to be backported.

--
Arfrever Frehtes Taifersar Arahesis

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415156

[Greg Stein-4]

On Fri, Nov 6, 2009 at 12:28, Arfrever Frehtes Taifersar Arahesis
<Arfrever.FTA@...> wrote:

> 2009-11-06 08:41:05 Greg Stein napisał(a):
>> On Wed, Nov 4, 2009 at 12:45, Hyrum K. Wright <hyrum@...> wrote:
>> > On Nov 4, 2009, at 8:48 AM, Julian Foad wrote:
>> >
>> >> kmradke@... wrote:
>> >>> I'd like to be a little selfish and wonder if it would be useful to
>> >>> release a
>> >>> 1.6.7 version faster than normal.  (Especially since the release
>> >>> schedule
>> >>> may be affected by the amount of holiday time near the end of the
>> >>> year.)
>> >>>
>> >>> We are really getting hurt by the external problems that were fixed
>> >>> in r40152
>> >>> and merged into 1.6.x in r40219.  Without this change we can't move
>> >>> a large
>> >>> number of our users from 1.5, and I would rather not be forced to
>> >>> compile
>> >>> custom patched versions for all the clients we use.
>> >>> (command line x 4 platforms, TortoiseSVN, Subclipse)
>> >>>
>> >>> Looking at STATUS, there are some other things already nominated
>> >>> (but need
>> >>> votes!!!) that also seem useful:
>> >>>
>> >>> * Fix for #3432
>> >>> * Win 7 exception handler support
>> >>> * Fix for #3489
>> >>> * Win32 performance improvements
>> >>> * future-proof patch
>> >>> * Fix for #3519
>> >>
>> >> For those who don't have a numerically-keyed hash of issues in their
>> >> brain:
>> >>
>> >> * Fix for #3432:
>> >>  "Merge can record mergeinfo from natural history gaps"
>> >>
>> >> * Fix for #3489:
>> >>  "Filenames with @ cannot be committed to the repository unless an @
>> >> is
>> >> added at the end."
>> >>
>> >> * future-proof patch
>> >>  "Make 1.6.7 (and subsequent Subversion releases) recognize Subversion
>> >> 1.7+ working copies."
>> >>
>> >> * Fix for #3519:
>> >>  "ra_serf using Label header unsupported by mod_dav_svn"
>> >>
>> >>
>> >>> I don't see it nominated, but the svn+ssh changes for windows may
>> >>> also
>> >>> fix some server resource problems others were seeing (We don't use
>> >>> svn+ssh,
>> >>> so I can't comment.)
>> >>>
>> >>> Is it too soon to be talking about this?
>> >>
>> >> +1 to the general notion of getting a 1.6.7 release out as soon as we
>> >> have some useful fixes in it.
>> >
>> > I've not yet looked at the the stuff that's currently been merged for
>> > 1.6.7, but there are several items in STATUS that are worth a
>> > release.  I'm hesitant to do one *too* quickly (<4 weeks), but I'm
>> > happy to cut a release in the next few weeks if the demand is there,
>> > and it sounds like it is.
>> >
>> > As usual, the roadblock for getting stuff into the release is the
>> > review needed in STATUS.  Please take some time to take a look.
>>
>> In the Apache Incubator, there has been a request for us to produce a
>> release. I'm going to point them at the 1.6.7 release process, so I'd
>> like to request that it starts in (say) a couple weeks.
>
> Change of license probably needs to be backported.

Nah. I think svncorp should make the release (off the 1.6.x branch as
usual, with no license backport or other changes to the file headers).
We'll just have the Incubator peeps observe our processes.

Cheers,
-g

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415176