6 Apr
2011
6 Apr
'11
4:47 p.m.
Hi all,
for 1, if we set 1A as a rule, nothing prevents somebody from doing 1B
(if they want to allow other people to preview their changes before they
are merged in master) or 1C if they want code review, and somebody else
pulling their changes in master, right?
So I go for 1A as a rule, it's what we have now, it's easy to
understand, better suited to our manpower and allows to have all the new
stuff in one place as soon as possible (which is also a matter of
(wo)manpower).
I see the fact that we can have 1B and 1C as a plus of using git instead
of svn, as benefits, which is great.
Now, about 2, I tend to agree with Fabio's proposal. I read that article
and I find it very easily understandable and matching our model.
(see below, I have intext comments)
On Mon, 2011-04-04 at 16:59 +0200, Sergiu Dumitriu wrote:
On 04/04/2011 03:32 PM, Fabio Mancinelli wrote:
I think I agree with this, master could be latest, meaning latest code
that was committed (not released). But this doesn't really matter, it's
what it contains that matters.
As for the contents, if I understood right, the "stable" branch that you
proposed, Sergiu, is not the same as the "master" branch in Fabio's
proposal, since the idea of the master branch is that it only gets
commits on releases, once per release, and _that is it_, as for the
stable branch, it gets commits in RC stage, when stuff need to be
thoroughly tested. In short, while master is a "history of stables",
stable is "the future stable release".
Now, I am also very interested in the lifetime of these branches, which
I think we should clearify. From my pov, tags are very important,
I always want to see the code that was released in a specific version. I
was told that in git a tag is not a copy like in svn, it's a bookmark in
a branch. Just to make things clear, I expect this bookmark to be
permanent.
I like the way these are handled in the model described by Fabio's link,
since it's very clear that all the tags will be in master.
However, there is one usecase we cannot implement with that model:
We release 3.0, merge it in master and tag
We release 3.1, merge it in master and tag
We release 3.0.1 (from a hotfix branch created from the 3.0 tag), merge
it where? and tag where?
So we should find an answer to this and then I'm fine with this model, we
can go ahead.
Hi Sergiu,
I've read your mail and I have some quick remarks/questions:
1) Concerning part 1, I like the fact of working on a fork and sending
pull requests (solution C), though it seems a bit overkill for
committers who "really knows what they are doing".
In my particular case I would gladly fork the project in order to have
the chance to submit pull requests whenever I think my code needs a
review.
I think that sending pull requests is only possible if you are working
on a fork, so if we want to use github as an effective code-review
platform we should always work with forks.
No, you can send "pull requests" even on committed code. In that case
the request must be manually closed. So even with direct commits it's
possible to ask for code reviews via pull requests.
The overhead of working with a fork could be
worked around by
selectively pushing stuff to the main repo.
For example, if I am working on something I am confident with, I will
push my changes to the main repo (that on https://github.com/xwiki)
and then pull them in the fork.
If I need a code review, I will push to my fork and issue a pull request.
I think this is doable with git, since you can have multiple remotes
but I am not an expert so I don't know if this could lead to "hellish"
workflows for the developers.
Yes, it's doable.
"git push" pushes to the "origin" repo.
"git push myown" pushes to a remote repo called "myown".
What's critical is being careful where each commit goes.
2) Concerning part 2, it's not clear to me
what is the lifetime of the branches?
Here I am implicitly advertising (again)
http://nvie.com/posts/a-successful-git-branching-model/ (I like it a
lot because it's simple and cristal clear :))
Yes, I found this one around the time it appeared, and I also liked it a
lot. It's similar to D, but from their branch names they use the
"master" branch as the most stable one, and not as the development
branch. This is useful for forking and checking out "stable" sources,
since it will point to the latest release. But personally, as a
developer, I prefer to have "master" as the most unstable branch, with the latest code in it. It's all just naming in the
end, not very important.
this is not doable for us, as one usecase is that at RC1 time, we move
the currently released version in a branch from which we will release
and the trunk becomes version + 1 milestone 1, so that people can still
develop. If we do this than everybody must keep their code on their
machines until release is done.
I think that your point D is very similar to it,
but it's not clear to
me how the release branch is handled.
From my point of view nobody is allowed to commit in the stable
branch, except the release manager when doing a release. When a
release starts a release branch is created from the development
branch. The release manager prepares the release in that branch and
when she's done she merges everything in stable and develop and then
delete the release branch.
Yes, exactly.
Feature branches appear and disappear as
developers work on single
features. Every feature branch is deleted after it's merged in
develop.
develop is the new "SNAPSHOT".
Yes, but feature branches don't have to be deleted that fast. A feature
branch could be left in place for users to try and merge a feature in
their own version, or to do bugfixing for a short period in it. It will
be decided case by case.
> So to sum up, in principle I am
>
> +1 for 1C provided that git allows to bypass the pull request by
> pushing in the main repo from a fork and that this doesn't generate
> too much hassle for a developer
> +1 for 2D provided that I have well understood the semantics of these
> branches :)
>
> -Thanks,
> Fabio
>
> On Sat, Apr 2, 2011 at 7:31 PM, Sergiu Dumitriu<sergiu(a)xwiki.com> wrote:
>> Hi devs,
>>
>> Since we're moving to git / GitHub, it's time to re-evaluate the
>> development / git usage strategy.
>>
>> = Short version:
>>
>> 1/ Where do developers commit:
>> A. Always in the master branch
>> B. In feature branches in the official repo, merged into the master
>> branch when ready
>> C. In their personal forks, requesting pulls in the official repo with a
>> mandatory code review from another developer
>>
>> 2/ How to move code from development to release:
>> A. Commit and release from the master branch >> B. Develop in feature branches, merge them in
the master when ready,
>> release from master
same, for roughly the same reason. Ok, people would commit in feature
branches, but one would be able to see how all this stuff looks together
unless they create their clone and pull all these branches, etc, in
short there'd be no way to create a nightly build of what is in
development for the "trunk branch" in milestone.
>> C. Develop in feature branches, merge them in
a development branch
>> (master) for polishing, merge them in a release branch when done-done
>> D. Develop in feature branches, merge them often in the development
>> branch (master) for snapshot testing, move into the stabilization
>> (pre-release) branch for polishing, move into the release branch when
>> done-done.
For C and D I have the same observation:
I don't think discussing things in terms of "for polishing" or
"done-done" is ok, I thought the idea is to get the things in as sooner
as possible to allow testing as soon as possible. Having this idea of a
very stable clean branch and various stages to get there, first needs a
lot of manpower, which we shouldn't assume we have.
Then, I cannot imagine an unflawed process through which things
would move correctly through these stages: either they will get old in a
branch from where nobody pulls them because they're too buggy or
something, or they will be pushed through too early because we're late
and it has to make this release or something like that.
I would think that the answer to the release quality issue is more time
to polish, and better discipline about it. More branches to get features
through will not fix the issue. It would help organizing QA if we had
more manpower, but it will not fix the problem if we don't.
I also strongly prefer one place where everything is since removing and
adding and removing and adding components and features and things over
and over can create integration issues continuously.
So in short I am for keeping somewhat the current model, knowing that I
have nothing against changing it in the future to a more gitty way of
doing things, if we decide we need it.
Thanks,
Anca
> E. Develop
in feature branches, merge them in a development branch for
> snapshot testing, move into a stabilization branch for the Next release,
> which becomes the Current release branch after the previous release is
> done, and which becomes a Maintenance branch after the new release is
> performed. Only bugfixes go in the Current release branch.
>
> I vote 1C and 2D.
>
> = Long version
>
> The common practice with Subversion is to have as few branches as
> possible, usually a trunk and a few maintenance branches, or
> development+stable+maintenance. This is a consequence of the perceived
> difficulty of merging changes between branches in svn, and the high cost
> of keeping multiple branches checked out.
>
> On the other hand, the git philosophy is to use branches as much as
> possible. Two core elements are "feature branches" and "forks".
>
> A feature branch is a branch where one feature is being developed,
> separated from the trunk and all the other features. While working on
> it, the developer "rebases" the branch on top of the trunk to keep his
> branch up to date with the trunk, and at the end "merges" the feature
> branch into the trunk. This way in-development features are kept out of
> the main trunk, but still allowing changes to be committed someplace
> public (no local uncommitted code anymore).
>
> A central element of GitHub is the ability to "fork" a repository. This
> means that a user clones a project in a personal repository where he can
> commit changes. He can later ask the maintainer of the original to
> "pull" those changes back into the original repository. This is the
> preferred way of contributing patches on GitHub.
>
>
>
> == Commit/Development-related strategies
>
> A. One central repository, one trunk (subversion-like)
>
> Developers clone the official trunks repository, prepare commits
> locally, then push back to the official repository. It's the same
> strategy that we're using now, except that we can also have an offline
> local repository.
>
>
> B. One central repository, feature branches
>
> Developers clone the official trunks repository, prepare commits locally
> in feature branches, then push back to the official repository in
> feature branches as well. When a feature is considered stable, it is
> merged into the master branch. Small bugfixes and improvements go
> directly in the master branch.
>
> B1. Also use specific helper branches
>
> Security fixes also go into a "security" branch so that users can
> cherry-pick them into older tags to build a custom patched version.
>
> Retired features can go into a "retired" branch so that users can
> re-include that feature in a custom build if they need it.
>
>
> C. One aggregated repository, pulling from developers
>
> Developers fork the official repositories, work on their fork (in
> feature branches as well), then make pull requests for integrating their
> work into the trunk. The rule would be that another developer has to do
> the pull after a code review (mandatory code reviews). This means more
> bugs spotted before committing, but also more work/time needed from the
> committers.
>
> We can relax the rule so that obvious bugfixes can be pulled by the same
> developer making the pull request.
>
>
> Personally, I prefer C, since it ensures better quality since at least
> two eyes see each line of code.
>
>
> == Integration/Release-related strategies
>
> Currently, we're developing on the trunk, and we're releasing from it
> during short breaks from live development. This is highly dangerous, and
> imposes a certain rhythm, with fast bursts of development right after a
> release, and imposed slowdown as the next release approaches (no work on
> new features after the last milestone).
>
> Short releases from a development branch is inline with agile
> development, but personally I find it too dangerous.
>
> Most big projects always keep the main development at least one branch
> away from the release branch.
>
> One example is the Linux kernel. While a kernel release lasts about 3
> months, like our own releases, almost all of the code that goes into a
> release has been developed before the merge window opens. This means
> that after a kernel version is released, Linus opens a two-weeks merge
> window during which he accepts pull requests for existing, working,
> complete code. The next ~10 weeks are spent testing the new kernel and
> integrating bugfixes, while developers prepare the features for the next
> kernel version. This ensures that a released kernel has as few bugs as
> possible. They can afford to do that since there are hundreds or
> thousands of contributors. Still, this is entirely opposite to our way
> of working: after a release we barely start writing the code to go in
> the new release, and we get code in at the last minute (especially me).
>
> Another example that I'd like to present is the new proposed strategy
> for Mozilla Firefox:
> http://mozilla.github.com/process-releases/draft/development_overview/
> Basically, the propose using 4 branches, from development to release,
> where code enters on the lowest branch, and moves up towards a release
> as it stabilizes and becomes release-ready. They use 6-weeks release
> cycles, and only stable-enough features get promoted from one branch to
> the next when a new cycle starts. This process ensures quality as well.
>
>
> I'd like to move closer to one of these two strategies, so that our
> releases are more polished. The mechanism for ensuring quality that
> we're currently using is to have an "investigation" phase during the
> previous release, which is supposed to help define the exact goals, so
> that during the current release the development should go smoothly
> towards that "idea goal". Unfortunately, this doesn't work that well.
> Without the code in place, investigations may miss important
> details/limitations that will shift the development in another
> direction. Or it can happen that the time is too short to fully
> implement something, so we can either release a very "in progress"
> feature, or decide near the end that it's not enough time to implement
> everything and focus on polishing what's already available to have a
> "partial" feature, but polished enough not to reek of low quality.
>
> The main problem here is that we're mixing feature- and time-based
> releases, with mandatory features that must find their way into a
> release, and a fixed deadline to make the release. This means that
> features have 8-10 weeks to be fully implemented, polished, tested,
> validated. And that doesn't always happen.
>
> So, here are some possible integration strategies:
>
> A. Master development (like now)
>
> All development is done in the master branch, from which we branch a few
> hours/days before the release, so that the master remains clear for
> development.
>
> B. Feature branches
>
> All new feature development is done in a separate branch for each
> feature, and we merge it in the trunk once it's considered done (or very
> close). When a release date comes, we release with the completed
> features, whatever those are. We don't force a merge of an incomplete
> feature just because it's in the roadmap if it's not stable enough.
>
> C. Feature+Development+Release branches
>
> All development is done in feature branches, but they get merged on the
> master branch more often to have test builds; the release branch is
> separate and it integrates features when they are considered ready. This
> has the advantage over B. that automated builds expose all the
> development features.
>
> D. Feature+Development+Stable+Release branches
>
> This is similar to the new Mozilla strategy. Developers merge their work
> in the Development branch very often. Users and other developers can
> contribute here as well, and preview the upcoming features. When they
> are close to finalization, they are also merged to the stable branch,
> where UX, QA and feature owners can test and improve the feature,
> preparing it for release. Once it's considered ready, it is merged into
> the release branch, where QA does a final thorough test. Releases happen
> from this branch.
>
> E. Feature+Development+Next+Release
>
> This is similar to D, with the exception that done features go into the
> next release, while the current release is staging. When the release is
> done, Release moves into Bugfix, Next becomes Release, and we create a
> new Next branch and start pulling in it. This would work well if we had
> very short release cycles (2 weeks), but it's not worth the effort for
> our current 3-month releases, since a feature would stagnate too much
> before being released. And it would also work if we had more beta testers.
>
>
> We can also impose windows, like 2-4 weeks for a feature to move into
> the next branch.
>
> We could also make faster releases, skipping milestones. and going to 6
> week releases.
>
> This means that it would take longer for a feature to make its way from
> idea to release. One release for investigation, one or more for the main
> development, and one for integration and stabilization.
>
> But this also means that releases will be more solid, polished, with
> less bugs, and closer to the user needs.
>
>
> Personally I prefer option D, although it's a bit too much overkill with
> our current limited manpower. We need more contributors and committers!
>
>
> As for a change strategy, we can continue the way we're doing now,
> gradually switching to feature branches and release/pre-release branches
> during the following release.
> --
> Sergiu Dumitriu
> http://purl.org/net/sergiu/