@average-vr-56795: correct

didn’t see this doc’d anywhere, but a helpful tip for running exactly 1 targeted rust test:

(cd src/rust/engine && cargo test -p <module> <name_of_test>)

e.g.

[omerta pants (kwlzn/5849)]$ (cd src/rust/engine && cargo test -p process_execution ensure_inline_stdio_is_stored)
    Finished dev [unoptimized + debuginfo] target(s) in 0.0 secs
     Running target/debug/deps/process_execution-9ccaddef784530c4

running 1 test
test remote::tests::ensure_inline_stdio_is_stored ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 34 filtered out

   Doc-tests process_execution

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

woot. thanks

oh that is clutch

@witty-crayon-22786 btw, the error message from returning an

Err

in

expand()

is great -- i obviously need to fix the message but plenty of context i think

Exception message: Build graph construction failed: ExecutionError Received unexpected Throw state(s):
Computing Select(Specs(dependencies=(DescendantAddresses(directory=u''),)), =TransitiveHydratedTargets)
  Computing Task(transitive_hydrated_targets, Specs(dependencies=(DescendantAddresses(directory=u''),)), =TransitiveHydratedTargets)
    Computing Task(transitive_hydrated_target, tests/python/pants_test/backend/jvm/tasks/coverage:coverage, =TransitiveHydratedTarget)
      Computing Task(hydrate_target, tests/python/pants_test/backend/jvm/tasks/coverage:coverage, =HydratedTarget)
        Computing Task(hydrate_sources, SourcesField(address=BuildFileAddress(tests/python/pants_test/backend/jvm/tasks/coverage/BUILD, coverage), arg=sources, filespecs={u'exclude': [], u'globs': [u'test_*.py', u'*_test.py']}), =HydratedField)
          Computing Snapshot(Key { id: 3324, type_id: TypeId(94630918927376) })
            Throw(PathGlobs expansion failed: Throw(???/globs with exclude patterns Gitignore { set: GlobSet { len: 0, strats: [] }, root: "", globs: [], num_ignores: 0, num_whitelists: 0, matches: ThreadLocal { local_data: None } } did not match: ["\"tests/python/pants_test/backend/jvm/tasks/coverage/*_test.py\""], "<pants native internals>"))
              Traceback (no traceback):
                <pants native internals>
              Exception: PathGlobs expansion failed: Throw(???/globs with exclude patterns Gitignore { set: GlobSet { len: 0, strats: [] }, root: "", globs: [], num_ignores: 0, num_whitelists: 0, matches: ThreadLocal { local_data: None } } did not match: ["\"tests/python/pants_test/backend/jvm/tasks/coverage/*_test.py\""], "<pants native internals>")

nice

i added some comments to the review.

i think that solving the

(input glob string -> PathGlob -> PathStat)

problem would also allow us to render the input glob string in all cases (rather than the compiled version)

that's exactly what i was about to comment lol

as in, the reason why i added a

RecursiveWildcard

well, that's a many to many relationship

i think regardless of RecursiveWildcard. think of something like

blah/*/thing/**/file.txt

right, that now turns into a single DirWildcard

...even though it has recursion?

the recursion is done when we do the

parse_globs()

call in

expand_single()

we were already doing doing that and if we moved the recursion on

**

into the

expand_single()

method (and part of that logic was in there anyway) we get 1:1 string -> PathGlob by adding the RecursiveWildcard -- maybe there's a better name for that

we weren't recursing in expand_single...

happily we are still able to use the same

parse_globs()

method in the expand_single() method and in PathGlobs::create

we were calling parse_globs() in the loop_fn

sure, and then the recursion happened in the loop_fn

we weren't recursing as in we weren't doing the double star logic in there, that's what i moved into expand_single()

so... i'm thinking of the

(input glob string -> PathGlob -> PathStat)

mapping as being something that is just maintained inside the loop.

when you expand a glob, you get a few pathstats and few pathglobs out

you add them to the mapping

ok, but then how do you want to represent the input glob strings -- as another element of a PathGlobs?

sure

oh

then sure

yeah then i can revert basically all of that