I'm sure this has been discussed a bunch, but is there any way to "speed up" the PR CI testing, or at least, short-circuit some of it? I'm adding a pretty isolated formatting plugin, but the testing is still around 50 minutes for a test failure which didn't show up on my local machine (for some reason, not sure why it passed locally). Is the pants dependency inference good enough that a simple

./pants --changed-since=... test

would be safe for the PR? If not, maybe run that test before running the whole suite? That way, if any of the modified code is breaking, CI will fail fast? And if not, we just fallback to a regular full test suite (with the modified code test results memoized)?

So where are we at w/ cherry-picking enhancements once rcs have been cut? We got a bit cherry-pick heavy the past few releases, and have reflected that 2.8, 2.10, and 2.11 were too hard to land.

actually, this might be good for @ancient-vegetable-10556

I've just created https://github.com/pantsbuild/pants/pull/15434 which uses the new test sharding functionality to shard our heavy Linux Python test job 3 ways, which should greatly shorten the overall CI wall time.

hi, thinking on adding some initial documentation for the Helm backend, could someone give me some pointers to where the MD files should be placed so they can be picked up when generating the documentation?

relatedly: @hundreds-father-404: i’d like to cut 2.12 docs to add the BSP stuff… unless you have anything queued to land in the next 30 minutes

I'm seeing

Exception: Cannot start Exclusive access on Destination Logging

when using

rule_runner.run_goal_rule(run.Run, args=["app/app.py"], env_inherit={"PATH", "PYENV_ROOT", "HOME"},)

I assume I'm just missing an arg somewhere? Breadcrumb anyone?

I've got release duty today. I'm positive I wont start until after 2 PM US Pacific, but if you want to steer things, pile on.

@fast-nail-55400, @ancient-vegetable-10556, @hundreds-father-404: published the BSP docs that Tom wrote (thanks Tom!): https://www.pantsbuild.org/v2.12/docs/jvm-overview#working-in-an-ide

Wasn't expecting to see intrisic rules being complained about in my test 😕

No installed rules return the type AddPrefix, and it was not provided by potential callers of @rule(<intrinsic>(AddPrefix) -> Digest).

Reviving a discussion from last year based on user feedback: what to do with `name=`: 1) always require it, or 2) usually don't require it for target generators. Feedback welcomed https://docs.google.com/document/d/1jSCJ3pwvIMg5duafsr8KmiLOO_2ARMaYCSrb8K85JQc/edit?usp=sharing

What's the protocol for bumping a tool version (namely

mypy

to

930

)?

weird error when trying to build example-kotlin repo:

ClasspathRootOnlyWasInner: The following targets had dependees, but can only be used as roots in a build graph:
  * src/jvm/org/pantsbuild/example/json/JsonExample.kt	(kotlin_source)

https://github.com/pantsbuild/example-kotlin/runs/6438158675?check_suite_focus=true

In order to fix

name=

(see above thread), it also means fixing

./pants path/to/dir

to run on all targets in the directory rather than the target with

name=

left off 🎉 Could I please get a review of my Specs Redesign Proposal Pt 2, especially the class modeling? https://docs.google.com/document/d/1WWQM-X6kHoSCKwItqf61NiKFWNSlpnTC5QNu3ul9RDk/edit#heading=h.1h4j0d5mazhu

and one more... 🙂 design doc summarizing previous plans on how to handle a

fmt

vs

fix

goal, along with merging

tailor

and

update-build-files

https://docs.google.com/document/d/1lpW2qGmhhT4Um6uGHRcpxwsseA7oxH86FE2L2gfBYWU/edit#

Stripe released “MarkDoc” which is the framework powering their docs site. https://markdoc.io/

ADR (architecture design record) question: there’s quite a few design documents shared as google docs floating around. They serve a great purpose, but I feel it is a rather fractured history if one where to backtrack and try to collect everything that has been decided and implemented (unless you have that and I don’t know about it) My question is merely to poke you into considering of keeping some form of ADR or Master document that links to all relevant design documents that have been accepted.

FYI @hundreds-father-404 this test is definitely flaky:

FAILED src/python/pants/backend/python/goals/tailor_test.py::test_ignore_empty_init[True]

Also FYI folks, just updated

mypy

to 950.

How would people feel about adding

experimental_shell_command

metadat to specify the

type

of asset? I need it to output a

resource

because I can't package a

file

😭

Picking up from the above ^. What would it look like if we unified

files

and

resources

into one target type:

asset

? Are there languages/ecosystems where the distinction is important (from a Build tooling perspective)?

Last year, @hundreds-father-404 gave a Charlas talk in Spanish at Pycon. Now that in-person tech events are gradually resuming, the Pants dev relations team is thinking about how to expand in-person Pants outreach among Hispanophone dev communities. If you'd be interested in getting involved, whether for brainstorming or more, let's chat! 👖

Does anyone use/what do people think of using "cloud" linters like SonarQube in Pants

I may have asked this before... in fact, I was positive I did, but my search-foo turned up nothing. I have a couple of discussions/questions/comments/proposals re: the CC backend compilation and JS/TS backend bundling/transpilation - is a Google Doc the best place for that? Github Discussions don't get any traction, correct?

I should also mention that I'm starting work shortly on a new C++ and Python project for around 6 months, which also seems like a great time to dogfood the CC backend.

I am running Pants from sources (latest checkout as of today) in a local repo as per https://www.pantsbuild.org/v2.12/docs/running-pants-from-sources.

$ ./pants_from_sources version
2.13.0.dev2

The repo is currently on

2.10.0

and has all tests/packaging workflows running fine. When I attempt to run tests using Pants built from sources, I get

$ ./pants_from_sources test project::
16:08:14.71 [ERROR] 1 Exception encountered:

  ResolveError: 'pymysql' was not found in namespace ''. Did you mean one of:
  :my-project

FATAL: exception not rethrown
Aborted (core dumped)

I am confused by this error message.

I wish Python had Rust enums. Any recommendations on the cleanest way to model this type as

None | ExportPythonTool

for a rule return type?

@dataclass(frozen=True)
class ExportPythonTool:
    resolve_name: str
    pex_request: PexRequest

Oof that's a tough one: the main file for Sphinx is

conf.py

.

tailor

is going to think that should be

python_sources

, along with default

sources=

for

python_sources

Has anyone written a Swift backend/plugin? I'm in need of one for a Bluetooth Swift/Python/C++ monorepo - and before I start on it, just wanted to check. I didn't see any tickets about it in the repo, but who knows 🤷

random thought: whether

mypy

needs to check transitive dependees is dependent on whether a file “exports” a type… and the same goes for

javac

, where we actually implemented “exports” (described in the JVM inference post)