It cannot. Instead you use a wrapping dataclass with a field that is optional

Cool, thanks!

🎣 for comments on https://github.com/pantsbuild/pants/pull/15002 It’s a draft PR, but more meant as a bug report with an attached repro 😉

We’re only 1 contributor from 100! https://github.com/pantsbuild/pants/graphs/contributors 😮

We've got 4 post-2.10.0 cherry-picks queued. I'm releaser this week and will be sending up a 2.10.1rc0 but wanted to check in for any strong opinions about when this is done. My preference is ~nowish since at least 1 pick directly addresses a user-reported issue where they are using 2.10.0.

@hundreds-father-404, @bitter-ability-32190: i didn’t do it in https://github.com/pantsbuild/pants/pull/15014, but: it seems like we should probably deprecate

requirement_constraints

in

2.11.x

…?

I'm late to doing this, but I've audited and edited all references to

register_options

in the 2.11.x docs to use the new style options

Should we have some kind of issue for

.pants.d

"hygiene" tracking? Right now I'm seeing a handful of annoying behavior: • pants creates a bunch of empty exception text files in the .pants.d directory ◦ Like a BUNCH • Similarly I see a bunch of dirs for each run in

.pants.d/run-tracker

◦ Not all of them have contents ◦ If they have contents it is a single file ◦ That single file is sometimes empty • If I do any run with

-ldebug

, the pants log quickly becomes useless, as it'd filled with:

14:58:43.45 [DEBUG] Launching 1 roots (poll=true).

◦ This gets written every 0.5 seconds Right now on CI, I'm thinking of wiping that dir before any pants run, because Ideally I zip it up and upload as an artifact

Yeah, I think that's a good idea

What's the best way to track a collection of issues?

What label shoudl we have the above use? It kind of straddles workdir/logging.

logging

?

🧪 Have you ever wanted to split off some part of your

@rule

which has

await Get

in it, but didn't want to go through the hoops of a dedicated

@rule

? ⚡ Behold, the mighty

@rule_helper

. ⚡ Does what it says on the tin! Allows you to split some code off of your

@rule

and

await Get

in it without the overhead of another rule. Example:

@rule_helper
async def _collect_thingamabobs(x, y, z):
    a = await Get(A, B, x)
    if y:
       a += await Get(C, D, z)
    return a

@rule
async def my_rule(...):
    ...
    a = await _collect_thingamabobs(x, y, z)
    ....

Code: https://github.com/pantsbuild/pants/blob/f43b0b92931c63509459f1113f7ca1ee34946225/src/python/pants/engine/rules.py#L282

I'm on release duty today. It looks like 2.11.x has 1 pending fix that Benjy is on top of in addition to a Pex upgrade I'd like to get in that fixes a Pex lock file issue. Let me know, or add to the milestone if you have others that should block release. I'll probably start that this afternoon assuming Benjy's fix is in and cherry-picked. I'll probably spin up the 2.12.0.dev2 release at that same time unless anyone yells.

FYI ASU issue when installing a lock file. I unfortunately do not have time to investigate this right now. Does anyone else reproduce?

Releasers: need help debugging broken releases for 2.11.x...

How can I make a

python_distribution

depending on an

archive

see the built package? When I run

package

for the distro I don't see it packaging the

BuiltPackage

I want to get people's 🧠 on what I'm working on which relates to (this post in general and this post above) I have a C extension module written in C++ which is being used heavily by other Python code. The file itself uses

pybind11

. Ideally, at any point where something depends on this module, it would get inferred. I think that's possible if I code it so the target is a code generator (because it kind of is). This kind of ignores a lot of hermeticity, but oh well. The two problems I face are

pybind11

. (I think I can wire up the codegenerator to invoke

setup.py

much like the

package

rule for

python_distribution

.) It's actually very trivial to add a

HTTPFileTarget(Target)

plus plumbing to Pants (and I want to upstream that). But the problem is that the existing machinery we'd want to re-use for building the sandbox filters the transitive dependencies to be "sources that likely make sense". Since

pybind11

has no

source

, I'll silently get nothing showing up in the sandbox. This smells of a larger problem though. Pants has tons of machinery revolving things with a "source". Codegen, generator targets, etc... But unless your chain of targets/rules/fields starts with a sources field, you don't get to participate (and in this case, my target gets silently filtered). One issue is that a dependency is just a target address, and so you need to know how to get it onto disk. If there was a way to take any target and ask make a request which gets it onto disk then you can have anything depend on anything.

hey folks: reminder to also mark a

Milestone

when you set the

needs-cherrypick

label: usually it’s obvious, but when the conversation needs to happen it’s good to frontload it

fyi you can use this link to get beta access to read me's new editor. https://app.livestorm.co/readme/q2-2022-product-updates doesn't sounds like you need to actually attend the event

hey gang: update on JVM work. the folks working on JVM tickets have upcoming milestones related to polishing the BSP support which @fast-nail-55400 added in the last few weeks: those are tracked in M4 and M5 of https://github.com/pantsbuild/pants/projects/22?fullscreen=true. although it still contained a fair number of issues i’ve closed out M3, and moved everything in there to Triage (roughly prioritized), since we don’t know of any true blockers for folks to use the JVM support, and the issues which remained in M3 will be easiest to tackle based on user feedback about priority

@witty-crayon-22786 should we be applying something like https://github.com/pantsbuild/pants/pull/14788 to command line specs?

hey folks. Would appreciate some help. Say I have collected the sources from

sources=['directory/**/*']

of a target (I work on Debian packages). Now I can use them as I need. However, I also need to have access to the

files

and

packages

items from the target declaration. I’ve looked up how it’s done in the

archive

target: https://github.com/pantsbuild/pants/blob/916f4273d8a6c4368e5fab299ce8e5f9e238336b/src/python/pants/core/target_types.py#L458 and got it working for me. So now I have a directory with files (from

sources

); arbitrary

files

and

.pex

packages from the respective parameters in the target declaration. What I need is to place

files

and

packages

into arbitrary locations inside the

sources

. So let’s say

sources=['directory/**/*']

gave me

directory/path/here/file1.txt

and I want to place

my-project.pex

into

directory/path/here/

(let’s keep it simple and assume I know where to place the items) before passing the

digest

of

sources

into the

Process

. Is it doable and if it is, what would be the best way to achieve this?

Somehow I merged a cherry pick on mobile with the checks failing https://github.com/pantsbuild/pants/pull/15115 I'll revert tonight after the kids go down Should those branches be protected?

@witty-crayon-22786 the implementation for dependencies and parameterization assumes that there is only one match per parameterization. Is that reliable?

Just wrote my first my first

@rule_helper

. Bless you @bitter-ability-32190 😭😍

./pants run build-support/bin/generate_docs.py

is failing for me on

main

😞

Any idea why

flake8

reports could be off by one?

Where is the rule graph seeded with built-in nodes like Process -> FallibleProcessResult?