Let's recap a bit:
Several people have raised the argument that build times can be kept in
check via other means (mostly differential builds via some means, be it
custom scripts or switching to gradle). I will start a separate
discussion thread on this topic, since it is a useful discussion in any
case.
I agree with this, and believe it is feasible to update the CI process
to behave as if the repository was split.
The suggestion of a "project split" within a single repository was
brought up.
This approach is a mixed bag; it avoids the downsides to the development
process that multiple repositories would incur, but also only has few
upsides. It seems primarily relevant for local development, where one
might want to skip certain modules when running tests.
There's no benefit from the CI side: since we're still limited to 1
.travis.yml, whatever rules we want to set up (e.g., "do not test core
if only connectors are modified") have to be handled by the CI scripts
regardless of whether the project is split or not.
Overall, I'd like to put this item on ice for the time being; the
subsequent item is related, vastly more impactful and may also render
this item obsolete.
A major topic of discussion is that of the development process. It was
pointed how that having a split repository makes the dev process more
complicated, since certain changes turn into a 2 step process (merge to
core, then merge to connectors). Others have pointed out that this may
actually be an advantage, as it (to some extent) enforces that changes
to core are also tested in core.
I find myself more in the latter camp; it is all to easy for people to
make a change to the core while making whatever adjustments to
connectors to make things fit. A recent change to the ClosureCleaner in
1.8.0 <https://issues.apache.org/jira/browse/FLINK-13586> comes to mind,
which, with a split repo, may have resulted in build failures in the
connectors project. (provided that the time-frame between the 2 merges
is sufficiently large...) As Arvid pointed out, having to feel the pain
that users have to go through may not be such a bad thing.
This is a fundamental discussion as to whether we want to continue with
a centralized development of all components.
Robert also pointed out that such a split could result in us
establishing entirely separate projects. We've had times in the past
(like the first flink-ml library) where such a setup may have simplified
things (back then we had lot's of contributors but no committer to
shepherd the effort; a separate project could be more lenient when it
comes to appointing new committers).
@Robert We should have a SNAPSHOT dependency /somewhere/ in the
connector repo, to detect issues (like the ClosureCleaner one) in a
timely manner and to prepare for new features so that we can have a
timely release after core, but not necessarily on the master branch.
@Bowen I have implemented and deployed your suggestion to cancel Travis
builds if the associated PR has been closed.
On 07/08/2019 13:14, Chesnay Schepler wrote:
Hello everyone,
The Flink project sees an ever-increasing amount of dev activity, both
in terms of reworked and new features.
This is of course an excellent situation to be in, but we are getting
to a point where the associate downsides are becoming increasingly
troublesome.
The ever increasing build times, in addition to unstable tests,
significantly slow down the develoment process.
Additionally, pull requests for smaller features frequently slip
through the crasks as they are being buried under a mountain of other
pull requests.
As a result I'd like to start a discussion on splitting the Flink
repository.
In this mail I will outline the core idea, and what problems I
currently envision.
I'd specifically like to encourage those who were part of similar
initiatives in other projects to share the experiences and ideas.
General Idea
For starters, the idea is to create a new repository for
"flink-connectors".
For the remainder of this mail, the current Flink repository is
referred to as "flink-main".
There are also other candidates that we could discuss in the future,
like flink-libraries (the next top-priority repo to ease flink-ml
development), metric reporters, filesystems and flink-formats.
Moving out flink-connectors provides the most benefits, as we straight
away save at-least an hour of testing time, and not being included in
the binary distribution simplifies a few things.
Problems to solve
To make this a reality there's a number of questions we have to
discuss; some in the short-term, others in the long-term.
1) Git history
We have to decide whether we want to rewrite the history of sub
repositories to only contain diffs/commits related to this part of
Flink, or whether we just fork from some commit in flink-main and
add a commit to the connector repo that "transforms" it from
flink-main to flink-connectors (i.e., remove everything unrelated to
connectors + update module structure etc.).
The latter option would have the advantage that our commit book
keeping in JIRA would still be correct, but it would create a
significant divide between the current and past state of the
repository.
2) Maven
We should look into whether there's a way to share dependency/plugin
configurations and similar, so we don't have to keep them in sync
manually across multiple repositories.
A new parent Flink pom that all repositories define as their parent
could work; this would imply splicing out part of the current room
pom.xml.
3) Documentation
Splitting the repository realistically also implies splitting the
documentation source files (At the beginning we can get by with
having it still in flink-main).
We could just move the relevant files to the respective repository
(while maintaining the directory structure), and merge them when
building the docs.
We also have to look at how we can handle java-/scaladocs; e.g.
whether it is possible to aggregate them across projects.
4) CI (end-to-end tests)
The very basic question we have to answer is whether we want E2E
tests in the sub repositories. If so, we need to find a way to share
e2e-tooling.
5) Releases
We have to discuss how our release process will look like. This may
also have repercussions on how repositories may depend on each other
(SNAPSHOT vs LATEST). Note that this should be discussed for each
repo separately.
The current options I see are the following:
a) Single release
Release all repositories at once as a single product.
The source release would be a collection of repositories, like
flink/
|--flink-main/
|--flink-core/
|--flink-runtime/
...
|--flink-connectors/
...
|--flink-.../
...
This option requires a SNAPSHOT dependency between Flink
repositories, but it is pretty much how things work at the moment.
b) Synced releases
Similar to a), except that each repository gets their own source
release that they may released independent of other repositories.
For a given release cycle each repo would produce exactly one
release.
This option requires a SNAPSHOT dependency between Flink
repositories. Once any repositories has created an RC or
finished it's release, release-branches in other repos can
switch to that version.
This approach is a tad more flexible than a), but requires more
coordination between the repos.
c) Separate releases
Just like we handle flink-shaded; entirely separate release
cycles; some repositories may have more releases in a given time
period than others.
This option implies a LATEST dependency between Flink
repositories.
Note that hybrid approaches would also make sense, like doing b) for
major versions and c) for bugfix releases.
For something like flink-libraries this question may also have
repercussions on how/whether they are bundled in the distribution;
options a)/b) would maintain the status-quo, c) and hybrid
approaches will likely necessitate the exclusion from the
distribution.