... i think that the

filter

options existed before we had that list syntax

see the help, but i think it's something more like

--filter-type=+python_library,python_binary

or something =/

(because it's a string option that is thus parsed for a particular usecase)

gotcha thanks

I was wondering if you have the same that python coverage is not including all the sources under src/python even if I set

--test-pytest-coverage=src/python/

some of python libraries are ignored (I’d at least expect to see 0% coverage on them) is there anything I am missing ?

hello! I want use pants in project which i work on but i dont know how to integrate it with code reload. I using flask and

app.run(debug=true)

dont work when i run my build:

./pants run src/python/helloworld:helloworld

. I never use

pants

before so i dont know how to handle code reload in development. I tried use

watchdog

for it but I dont think its way to go:

watchmedo auto-restart --directory=./src ./pants run src/python/helloworld:helloworld

(i know some flags is missing in this example but its only example:D)

Hey there. Is there a way to tell which targets were invalidated during a

pants

run? I keep seeing messages like

Invalidated 2 targets

and I know it has to do with conflicting transitive dependencies but I’m not sure which ones. Knowing which targets were invalidated might help

is there a way to specify additional packages to install for a pants plugin? I'm loading it through backend_packages but need to get it to install some things with pip before it can run the plugin

question: when pants downloads dependencies via ivy, where is that stored? I can find them in both

~/.ivy2

as well as under

.pants.d/ivy

, and I'm not quite sure which one does what. Or is this something we've configured that I just don't know about?

@acoustic-book-58772

~/.ivy2

has the real stuff,

.pants.d/ivy

contains the symlinks/hardlinks to

~/.ivy2

depending on your pants version

ah gotcha thanks!

is there a way to control the bundle directory's name in

dist/

?

instead of

classpath.targetname

?

Hey, guys! trying to setup remote cache for pants, is there a way work with auth 2.0 flow and specify credentials or headers to make urls in `write_to`/`read_from` use it or do you have to create a plugin overriding the default behavior?

Hello, I need help with something. I’m trying to create a static properties file containing information about the product at the time of bundle (version, git revision, jenkins build no. etc.). I implemented a task that creates the file, or a string that can be written as a properties file. How can I have this file included as a resource? Either in its own jar or in one of the jars I’m creating Thanks

Hey, I’m writing a pants plugin and I’m trying to use boto3 in it. Is there some specific way to add 3rdparty dependencies to pants plugins? Pants is failing on the boto3 import when I add the dependency in the BUILD file

Upgraded from some old 1.3 beta to 1.7 recently, and discovered that we had to run

./pants lint

explicitly to run lint checks. but it seems like the checks are taking a long time (like 27 minutes) to run. output has like:

02:20:31 00:07     [python-eval]
                   Invalidated 120 targets.........................................................................................................................
02:47:37 27:13     [pythonstyle]
                   Invalidated 160 targets.

known issue? bizarre? what’s going on?

Hey all, hopefully a qq for folks here. I'm writing a pants plugin to generate code from an OpenAPI spec. I have my basic task (and custom target) defined and it works, running in the 'gen' phase and placing generated Java code into a specified repo directory. However, I have some java_library targets that depend on my custom target and my first build of those targets always fails as if the generated code is not there. Subsequent builds DO work, however, but I'm worried that each build is using the output of the previous build, which is not desired. I've been staring at the existing thrift and antlr codegen plugins for awhile and I'm not seeing what those plugins might be doing to avoid this problem. Anyone recognize what I might be missing here to get my codegen plugin working properly?

Could you say more about this bit?

placing generated Java code into a specified repo directory

By that do you mean that you’re putting the generated code into the

target_workdir

passed to

execute_codegen

?

@fancy-queen-20734 I am not putting the generated code in the workdir. I'm putting it next to where the BUILD file is. Is it important to put the generated code in the target workdir?

yes

the task creates a synthetic target that owns the sources in the passed workdir. If they’re not there, the dependees of the codegen target won’t see them

That must be it then. I'll switch to using the target workdir.

cool

Thanks for the help :)

What docs / examples did you look at while trying to implement the task? I bet they could be clearer about this particular part

i would love an options equivalent to the

//

we have for targets so we can specify globally-scoped (or arbitrarily-scoped) things later on the command line

that's leaning into an implementation detail but e.g. it is impossible to refer to

--binaries-baseurls

on the command line

i'm gonna make this into an issue i think