<@U02Q3KHC59B>, <@U02KAN6061E>, <@U021C96KUGJ>: so...
# development
w
@sparse-lifeguard-95737, @bitter-ability-32190, @ancient-vegetable-10556: so! i think that you all have the largest (pending?) changes targeted for 2.15.
it sounds like Chris might have his stuff mostly in sometime early next week
b
šŸ«”
w
i donā€™t actually know what is pending for
fmt
/
lint
/
fix
ā€¦ just docs at this pointā€¦?
a
@bitter-ability-32190 I may well have questions for you about the internals/intent behind partitioning, which Iā€™m saving for Monday
šŸ‘ 2
This is the context, with a bunch of relevant PRs linked: https://github.com/pantsbuild/pants/issues/17129
s
the test-batching work took a detour to iterate on the generic partitioning API - the hairiest API change (I hope) merged today, unblocking https://github.com/pantsbuild/pants/pull/17255. once 17255 merges I plan to update https://github.com/pantsbuild/pants/pull/17134 (core
test
goal impl) to use the new generic APIs
a
a generic partitioning API? oooooh
s
and once the core
test
changes are merged, I will follow up with the actual batched
pytest
implementation šŸ˜…
w
cool beansā€¦ so maybe that is the long poleā€¦?
a
we should probably make sure the needs of the environments work is met by the partitioning API
b
So on the docket are my pending PRs. I don't think I'll get to adding new features in 2.15.x past those. 2.16.x might see a pants.toml fixer and a "generic" prettier targetless support šŸ¤“
Then another docs pass for 2.15
(Symlinks aren't a blocker)
w
So on the docket are my pending PRs.
which? maybe worth marking them with the
2.15.x
Milestone
b
Done
I flagged the
external_tools
because I wanna blog about it šŸ˜ˆ
s
@witty-crayon-22786 I expect that I can get the core
test
changes out of ā€œdraftā€ mode on Monday, maybe even merged depending on what comes out in reviews. and then get the batched
pytest
PR up the next day (Iā€™ve been playing with it locally and donā€™t expect it to be too complex)
šŸ‘ 1
b
I think I'm the blocker but I got a busy weekend šŸ˜…
w
yea, donā€™t worry about it: there is no actual time pressure until someone decides they want to land something else on top of all of this and we need to start saying no
b
I think symlinks are too risky to land in 2.15. but the current PR is such that no one should hit that code in the wild
w
yea, that seems like a great thing to wait
b
The following PR that actually wires it up should wait
w
@ancient-vegetable-10556, @sparse-lifeguard-95737, @bitter-ability-32190: how is all of this looking?
would be great to be able to cut in the next few days
b
I need to deprecate uodate-build-files. Will post tomorrow. Then 2.15 is good to go for me
šŸŽ‰ 1
s
My PR for the generic batched test plumbing is up for review now. I have a couple fast-follows for TODOs in that, and then the actual ā€œbatched pytestā€ implementation that Iā€™ll work on tomorrow
(https://github.com/pantsbuild/pants/pull/17134 is the PR generic plumbing PR blocking the rest)
w
sounds goodā€¦ yea, anything that isnā€™t in before the branch cut will need to be cherry-pickedā€¦ but it sounds like maybe some of the fast-follow stuff could be picked (or not, if they can wait for 2.16.x?)
let me know if i can help review anything.
s
The ā€œfast followā€ stuff is already done on branches, I just havenā€™t figured out how to stack PRs across forks šŸ˜…
(They are user-facing API changes I donā€™t want to mix with the plug-in API change)
Iā€™ve wondered about waiting on the pytest-batching - is it possible to ā€œoverrideā€ the definition of a rule by redefining a rule with the same types?
If so we could experiment with the actual batched pytest logic before upstreaming it
(Though, I expect it wonā€™t be much work to make the PR directly tomorrow)
w
is it possible to ā€œoverrideā€ the definition of a rule by redefining a rule with the same types?
not currently, no: would need to be able to ā€œuninstallā€ the existing rule before adding the replacement. weā€™ve discussed that (and i think that it is trivial to do?), but i canā€™t find the ticket for itā€¦
hey gang! today is the day! (?) whatā€™s outstanding for 2.15.x?
b
Before the release we should fix the python run goal, but doesn't can be cherry picked
s
I just opened https://github.com/pantsbuild/pants/pull/17385 as a draft to get review rolling - working on tests now
https://github.com/pantsbuild/pants/pull/17361 needs updated but should be a quick turn-around, will knock it out in a few minutes
w
ah, thanks: sorry, had missed your other thread update
s
pushed some tests to https://github.com/pantsbuild/pants/pull/17385 and took it out of draft mode!
going to file a separate issue for the docs
w
thanks: looking now. cc @hundreds-father-404
h
thanks, pairing and won't be able to take a look at things right now
šŸ‘ 1
w
@sparse-lifeguard-95737: i donā€™t really see any blockers there, so when youā€™re ready, let me know and iā€™ll (set auto-)merge
s
Itā€™s ready as far as I know šŸ˜„
Ah I see your comments came through - let me take a look throughā€¦
@witty-crayon-22786 sorry, I keep stepping on the auto-merge - I pushed one more commit to have pytest partition by environment (I think eventually we donā€™t want plugins to need to think about environments? but for now itā€™ll avoid a class of error). I promise Iā€™m done pushing to the branch now šŸ™‚
so close I can taste it - one of the CI shards hit a flaky failure, can somebody poke it?
w
yep
huzzah! ok folks: it seems like we can cut
2.15.0a0
and
2.15.x
? we can of course continue to cherry-pick, but if there is anything that is already in the queue to land that i should wait for, let me know.
ok @sparse-lifeguard-95737: now you can go on vacation (again)
šŸŒ“ 1
c
I think this one would be good to land in 2.15.x https://github.com/pantsbuild/pants/pull/17330
šŸ‘ 1
cp is of course an option too
w
@curved-television-6568: šŸš¢ with minor fixes: thanks!
šŸ™ 2
c
thanks! šŸ™‚
w
iā€™ll assume that that lands, and include it in the prep.
šŸ‘ 1
c
enabled auto-merge..
feel free to make it go through if thereā€™s any hicups
w
hey folks! please talk with me before merging anything else to
main
until the release prep has landed in https://github.com/pantsbuild/pants/pull/17392
h
Och, I screwed up and merged 17393 right before you merged this prep, sorry! But itā€™s internal, so doesnā€™t change the changelog
So I think you donā€™t need to redo anything?