You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@couchdb.apache.org by Jan Lehnardt <ja...@apache.org> on 2010/04/03 19:53:13 UTC
1.0 Code Management
Hi dev@,
The next bigger thing up is our 1.0.0 release (yay!) and I'd like to get
clarification on how we are going to use SVN for this. So far, we
cut new releases from trunk, put them in a branch and kept that
up with bug fixes.
This time 0.11.x is meant to be the feature freeze for 1.0.0. Does that
mean we are going to use the 0.11.x branch as the basis for 1.0.0 or
do we want to keep using trunk?
I'd be fine with whatever we decide here, but I do want to get consensus.
A conservative view would mandate that we use the 0.11.x branch to add
bug fixes until it is shaped into a form that we want to call 1.0.0 and then
branch that into the 1.0.x branch that we keep for maintenance.
This would allow us to keep using trunk for new features as usual. This
also means, we have to maintain two to three branches (0.11.x, trunk and
eventually 1.0.x). This is a lot of bookkeeping.
One could argue that we should focus on 1.0.0 anyway and shouldn't add
new features to trunk until we're done, but I don't want to prescribe for
anyone what code to write when (that said, we can easily create another
branch for new features that we can merge back into trunk after 1.0.0 or each
developer can maintain a (semi-)-private git(hub) branch until after 1.0.0)
If we keep using trunk for 1.0.0 we should revisit if any patches that landed
in trunk after we branched 0.11.x are in fact meant for 1.0.0. I'm happy to
produce a list if there are any concerns. If there no concerns I'm happy to
ignore this point :)
I think the easiest way forward (as in easiest for us developers) is to
use trunk as the basis for 1.0 and agree to put new features elsewhere
until we branch 1.0.x.
But I'm happy to use 0.11.x as the basis and help managing all the
merging that needs being done, if we decide to do that.
What do you think?
Cheers
Jan
--
KISS
Re: 1.0 Code Management
Posted by Benoit Chesneau <bc...@gmail.com>.
On Wed, Apr 7, 2010 at 3:27 PM, Noah Slater <ns...@me.com> wrote:
>
> On 7 Apr 2010, at 14:16, Jan Lehnardt wrote:
>
>>> Btw I still can't change state of a bug, anyone can add to my user
>>> rhe needed rights?
>>
>> Noah, Paul, Damien, any idea?
>
> I bumped your perms.
>
Thanks.
- benoît
Re: 1.0 Code Management
Posted by Noah Slater <ns...@me.com>.
On 7 Apr 2010, at 14:16, Jan Lehnardt wrote:
>> Btw I still can't change state of a bug, anyone can add to my user
>> rhe needed rights?
>
> Noah, Paul, Damien, any idea?
I bumped your perms.
Re: 1.0 Code Management
Posted by Jan Lehnardt <ja...@apache.org>.
On 7 Apr 2010, at 14:58, Benoit Chesneau wrote:
> I am for a 1.0 branch too. Are all bugs marked already ? One thing
> that would be good is to dispatch bugs without owners actually, so it
> would help to not work on smth already handled somewhere in the
> offline world.
I think a run through JIRA for anyone interested is worth it. I'm planning
on it in the next weeks.
> Btw I still can't change state of a bug, anyone can add to my user
> rhe needed rights?
Noah, Paul, Damien, any idea?
Cheers
Jan
--
Re: 1.0 Code Management
Posted by Benoit Chesneau <bc...@gmail.com>.
On Sat, Apr 3, 2010 at 7:53 PM, Jan Lehnardt <ja...@apache.org> wrote:
> Hi dev@,
>
> The next bigger thing up is our 1.0.0 release (yay!) and I'd like to get
> clarification on how we are going to use SVN for this. So far, we
> cut new releases from trunk, put them in a branch and kept that
> up with bug fixes.
>
> This time 0.11.x is meant to be the feature freeze for 1.0.0. Does that
> mean we are going to use the 0.11.x branch as the basis for 1.0.0 or
> do we want to keep using trunk?
>
> I'd be fine with whatever we decide here, but I do want to get consensus.
>
> A conservative view would mandate that we use the 0.11.x branch to add
> bug fixes until it is shaped into a form that we want to call 1.0.0 and then
> branch that into the 1.0.x branch that we keep for maintenance.
>
> This would allow us to keep using trunk for new features as usual. This
> also means, we have to maintain two to three branches (0.11.x, trunk and
> eventually 1.0.x). This is a lot of bookkeeping.
>
> One could argue that we should focus on 1.0.0 anyway and shouldn't add
> new features to trunk until we're done, but I don't want to prescribe for
> anyone what code to write when (that said, we can easily create another
> branch for new features that we can merge back into trunk after 1.0.0 or each
> developer can maintain a (semi-)-private git(hub) branch until after 1.0.0)
>
> If we keep using trunk for 1.0.0 we should revisit if any patches that landed
> in trunk after we branched 0.11.x are in fact meant for 1.0.0. I'm happy to
> produce a list if there are any concerns. If there no concerns I'm happy to
> ignore this point :)
>
> I think the easiest way forward (as in easiest for us developers) is to
> use trunk as the basis for 1.0 and agree to put new features elsewhere
> until we branch 1.0.x.
>
> But I'm happy to use 0.11.x as the basis and help managing all the
> merging that needs being done, if we decide to do that.
>
> What do you think?
>
> Cheers
> Jan
> --
> KISS
>
>
>
I am for a 1.0 branch too. Are all bugs marked already ? One thing
that would be good is to dispatch bugs without owners actually, so it
would help to not work on smth already handled somewhere in the
offline world.
Btw I still can't change state of a bug, anyone can add to my user
rhe needed rights?
- benoit
Re: 1.0 Code Management
Posted by Filipe David Manana <fd...@gmail.com>.
Hi,
I'll vote for basing 1.0.0 on the 0.11.x branch.
Otherwise, create a new branch for post 1.0.0 features and base 1.0.0 on
trunk?
cheers
On Sat, Apr 3, 2010 at 6:53 PM, Jan Lehnardt <ja...@apache.org> wrote:
> Hi dev@,
>
> The next bigger thing up is our 1.0.0 release (yay!) and I'd like to get
> clarification on how we are going to use SVN for this. So far, we
> cut new releases from trunk, put them in a branch and kept that
> up with bug fixes.
>
> This time 0.11.x is meant to be the feature freeze for 1.0.0. Does that
> mean we are going to use the 0.11.x branch as the basis for 1.0.0 or
> do we want to keep using trunk?
>
> I'd be fine with whatever we decide here, but I do want to get consensus.
>
> A conservative view would mandate that we use the 0.11.x branch to add
> bug fixes until it is shaped into a form that we want to call 1.0.0 and
> then
> branch that into the 1.0.x branch that we keep for maintenance.
>
> This would allow us to keep using trunk for new features as usual. This
> also means, we have to maintain two to three branches (0.11.x, trunk and
> eventually 1.0.x). This is a lot of bookkeeping.
>
> One could argue that we should focus on 1.0.0 anyway and shouldn't add
> new features to trunk until we're done, but I don't want to prescribe for
> anyone what code to write when (that said, we can easily create another
> branch for new features that we can merge back into trunk after 1.0.0 or
> each
> developer can maintain a (semi-)-private git(hub) branch until after 1.0.0)
>
> If we keep using trunk for 1.0.0 we should revisit if any patches that
> landed
> in trunk after we branched 0.11.x are in fact meant for 1.0.0. I'm happy to
> produce a list if there are any concerns. If there no concerns I'm happy to
> ignore this point :)
>
> I think the easiest way forward (as in easiest for us developers) is to
> use trunk as the basis for 1.0 and agree to put new features elsewhere
> until we branch 1.0.x.
>
> But I'm happy to use 0.11.x as the basis and help managing all the
> merging that needs being done, if we decide to do that.
>
> What do you think?
>
> Cheers
> Jan
> --
> KISS
>
>
>
--
Filipe David Manana,
fdmanana@gmail.com
"Reasonable men adapt themselves to the world.
Unreasonable men adapt the world to themselves.
That's why all progress depends on unreasonable men."
Re: 1.0 Code Management
Posted by J Chris Anderson <jc...@gmail.com>.
On Apr 6, 2010, at 7:00 PM, Randall Leeds wrote:
> I know I would like to throw some patches down in the coming weeks and while
> i'd love for some to hit 1.0 I recognize the need for extreme conservatism
> with respect to the 1.0 branch as adoption picks up for a stable release.
>
> +1 on cutting 1.0 from the 0.11.x branch. In my head 1.0 === 0.11.1.
>
I agree on branching 1.0.x from 0.11.x, and not restricting trunk.
There is some stuff in trunk (and a lot of stuff on Jira) that should go into 0.11.1. I plan to do some patch vetting and applying over the next couple of weeks.
I can imagine 0.11.1 being released before 1.0 -- maybe 0.11.1 really will be treated as a release candidate for 1.0, so we can get a lot of eyes on it, and branch 1.0 from the 0.11.x branch after we've had a chance to make sure our 0.11.x bugfixes don't have any surprises.
The one feature I'd like to see in 1.0 (that we've been talking about for a while) is the _replicator db. It'd be great to get that out for testing as part of 0.11.1 as well.
I've set aside some time in the next few weeks to work on the _replicator db. I'll be sure to describe my approach etc with this list so there aren't surprises. I think it's a pretty straightforward feature (if you want to discuss this now, please reply in a new thread, let's not muddy this one).
Chris
> On Apr 6, 2010 5:33 PM, "Adam Kocoloski" <ko...@apache.org> wrote:
>
> On Apr 5, 2010, at 6:16 PM, Tim Smith wrote:
>
>> On Sat, Apr 3, 2010 at 11:53 AM, Jan Lehnardt <jan@a...
> I'd also like to cut 1.0 from the 0.11 branch and not have any restrictions
> on what lands in trunk. Best,
>
> Adam
Re: 1.0 Code Management
Posted by Jan Lehnardt <ja...@apache.org>.
On 7 Apr 2010, at 04:00, Randall Leeds wrote:
> I know I would like to throw some patches down in the coming weeks and while
> i'd love for some to hit 1.0 I recognize the need for extreme conservatism
> with respect to the 1.0 branch as adoption picks up for a stable release.
The earlier you send the patches to more time they have to simmer :)
Cheers
Jan
--
>
> +1 on cutting 1.0 from the 0.11.x branch. In my head 1.0 === 0.11.1.
>
> On Apr 6, 2010 5:33 PM, "Adam Kocoloski" <ko...@apache.org> wrote:
>
> On Apr 5, 2010, at 6:16 PM, Tim Smith wrote:
>
>> On Sat, Apr 3, 2010 at 11:53 AM, Jan Lehnardt <jan@a...
> I'd also like to cut 1.0 from the 0.11 branch and not have any restrictions
> on what lands in trunk. Best,
>
> Adam
Re: 1.0 Code Management
Posted by Randall Leeds <ra...@gmail.com>.
I know I would like to throw some patches down in the coming weeks and while
i'd love for some to hit 1.0 I recognize the need for extreme conservatism
with respect to the 1.0 branch as adoption picks up for a stable release.
+1 on cutting 1.0 from the 0.11.x branch. In my head 1.0 === 0.11.1.
On Apr 6, 2010 5:33 PM, "Adam Kocoloski" <ko...@apache.org> wrote:
On Apr 5, 2010, at 6:16 PM, Tim Smith wrote:
> On Sat, Apr 3, 2010 at 11:53 AM, Jan Lehnardt <jan@a...
I'd also like to cut 1.0 from the 0.11 branch and not have any restrictions
on what lands in trunk. Best,
Adam
Re: 1.0 Code Management
Posted by Adam Kocoloski <ko...@apache.org>.
On Apr 5, 2010, at 6:16 PM, Tim Smith wrote:
> On Sat, Apr 3, 2010 at 11:53 AM, Jan Lehnardt <ja...@apache.org> wrote:
>> A conservative view would mandate that we use the 0.11.x branch to add
>> bug fixes until it is shaped into a form that we want to call 1.0.0 and then
>> branch that into the 1.0.x branch that we keep for maintenance.
>>
>> This would allow us to keep using trunk for new features as usual. This
>> also means, we have to maintain two to three branches (0.11.x, trunk and
>> eventually 1.0.x). This is a lot of bookkeeping.
>
> I think it's best to keep trunk == where development happens.
> Bookkeeping is inevitable, and I think it doesn't serve anyone to
> create a new development branch named something other than 'trunk' and
> then use it like we should be using trunk.
>
> I'd be happy to help with the bookkeeping, if that would be useful.
>
> Tim
> --
> If you're not part of the solution, you're part of the precipitate
I'd also like to cut 1.0 from the 0.11 branch and not have any restrictions on what lands in trunk. Best,
Adam
Re: 1.0 Code Management
Posted by Tim Smith <ti...@couch.io>.
On Sat, Apr 3, 2010 at 11:53 AM, Jan Lehnardt <ja...@apache.org> wrote:
> A conservative view would mandate that we use the 0.11.x branch to add
> bug fixes until it is shaped into a form that we want to call 1.0.0 and then
> branch that into the 1.0.x branch that we keep for maintenance.
>
> This would allow us to keep using trunk for new features as usual. This
> also means, we have to maintain two to three branches (0.11.x, trunk and
> eventually 1.0.x). This is a lot of bookkeeping.
I think it's best to keep trunk == where development happens.
Bookkeeping is inevitable, and I think it doesn't serve anyone to
create a new development branch named something other than 'trunk' and
then use it like we should be using trunk.
I'd be happy to help with the bookkeeping, if that would be useful.
Tim
--
If you're not part of the solution, you're part of the precipitate