pantsbuild #development

Channels

welcome

github-notifications

development

aloof-angle-91616

09/10/2020, 5:24 PM

well we already do default the name to the containing directory and advise people to use that for 111

hundreds-father-404

09/10/2020, 5:25 PM

Yes, but that is one single consistent rule regardless of target type. Now, we’d be saying that each target type might have its own logic, and you have to figure that out to be able to reference your target

aloof-angle-91616

09/10/2020, 5:25 PM

i think the explicitness argument is reasonable though. if the default is well chosen (?) it could be overcome

aloof-angle-91616

09/10/2020, 5:26 PM

i guess also because we have file args working great now too the need for this is reduced

👍 1

witty-crayon-22786

09/10/2020, 7:01 PM

https://twitter.com/rustlang/status/1304080840402632704?s=09

💯 1

hundreds-breakfast-49010

09/10/2020, 9:53 PM

is it just me or did we break ctrl-c in the no-pantsd case at some point?

aloof-angle-91616

09/10/2020, 9:54 PM

i think i remember now that i reproed that maybe a month ago and i’m quite sorry for not filing an issue — i remember trying to fix it and just having 3 other things to do that day

aloof-angle-91616

09/10/2020, 9:55 PM

if you post an issue i can try to repro

aloof-angle-91616

09/10/2020, 9:55 PM

but i think it came up very easily

hundreds-breakfast-49010

09/10/2020, 9:55 PM

I'm investigating ctrl-c in the --no-pantsd case at the moment, it looks like

./pants --no-pantsd test <test target>

seems to have broken ctrl-c handling

👀 1

hundreds-breakfast-49010

09/10/2020, 9:56 PM

but other goals seem not to. e.g.

./pants --no-pantsd fmt src/python/pants::

hundreds-father-404

09/11/2020, 4:35 AM

Hi, I can’t figure out a Rust error about referring a local variable. I understand the principle, but can’t figure out how I’m still referencing the local var. Would anyone mind taking a look, please? It’s this file: https://github.com/pantsbuild/pants/pull/10760/files#diff-28aa60818675b10d3b7e84f5e6c78be5

hundreds-father-404

09/11/2020, 4:35 AM

Copy code

error[E0515]: cannot return value referencing local variable `file_contents_and_directories`
   --> src/intrinsics.rs:364:3
    |
331 |     let digests: Vec<_> = file_contents_and_directories
    |                           ----------------------------- `file_contents_and_directories` is borrowed here
...
364 | /   async move {
365 | |     let digests = future::try_join_all(digests).await.map_err(|e| throw(&e))?;
366 | |     let digest = store
367 | |       .merge(digests)
...   |
371 | |   }
372 | |   .boxed()
    | |__________^ returns a value referencing data owned by the current function

error: aborting due to previous error

For more information about this error, try `rustc --explain E0515`.

fast-nail-55400

09/11/2020, 5:09 AM

you have two closures involved, do you need to move into the

map

closure?

fast-nail-55400

09/11/2020, 5:09 AM

(which surrounds the

async move

closure)

fast-nail-55400

09/11/2020, 5:12 AM

never mind, in the map closure, you have implicitly borrowed

file_contents_and_directories

👍 1

fast-nail-55400

09/11/2020, 5:13 AM

then you

future::try_join_all(digests).await.map_err(|e| throw(&e))?;

fast-nail-55400

09/11/2020, 5:13 AM

digests

as a parameter refers to the binding to which the map is assigned, right?

fast-nail-55400

09/11/2020, 5:14 AM

just a guess

hundreds-father-404

09/11/2020, 5:14 AM

What I’m a little confused by is why it worked before, but what I changed to break it. This code is the same general shape. I renamed

file_values

file_contents_and_directories

, but same type. And then the

async move {}

is more complex in that it has an

if else

fast-nail-55400

09/11/2020, 5:14 AM

🤷

hundreds-father-404

09/11/2020, 5:15 AM

digests as a parameter refers to the binding to which the map is assigned, right?

Iiuc,

digests

takes ownership of the created

Vec<Future<Result<Digest, String>>>

fast-nail-55400

09/11/2020, 5:16 AM

the indentation is confusing in the diff

fast-nail-55400

09/11/2020, 5:16 AM

so that later

async move

after

digests

is bound earlier then, got it

hundreds-father-404

09/11/2020, 5:16 AM

Agreed. I wish we did 4 space indentation in Rust

fast-nail-55400

09/11/2020, 5:17 AM

https://doc.rust-lang.org/stable/error-index.html#E0515

👍 1

fast-nail-55400

09/11/2020, 5:19 AM

what is the exact type of

digests

hundreds-father-404

09/11/2020, 5:19 AM

That error explanation makes a lot of sense. I can’t grok what is dangling here, though. I don’t know why

file_contents_and_directories

still ends up being referenced. I would think that after binding

digests

, that

file_contents_and_directories

is dropped Maybe the issue is borrowing

file_contents_and_directories

, and it shouldn’t borrow?

fast-nail-55400

09/11/2020, 5:19 AM

is it a

Vec<&Digest>

hundreds-father-404

09/11/2020, 5:20 AM

what is the exact type of digests?

According to Rustc:

Copy code

std::vec::Vec<impl futures::Future>