-
Notifications
You must be signed in to change notification settings - Fork 476
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: allow groups on submodules #2263
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice! I think I would avoid putting storing all attributes on a Justfile
, for a couple reasons. One abstract reason is that I try not to make changes in anticipation future needs. There's a good chance those future needs may manifest, or they might manifest differently than expected. And one concrete reasons is that attributes serialize in a somewhat odd way, so they will be difficult to parse using jq
, if you use --dump --dump-format json
to dump a justfile, whereas a groups: Vec<String>
field will be easy to parse, so I think I would separate out groups and just store the group names on Justfile
.
@casey Thanks for the feedback! Great points, I'll get this updated to just pass in the group names instead. |
09ab72e
to
42f6c7f
Compare
@casey I switched to passing around |
42f6c7f
to
69cca4a
Compare
Fixed linter errors and re-pushed. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice, left some initial comments! To check if the PR will pass CI, run just ci
, which runs clippy and errors on warnings.
@@ -771,6 +771,88 @@ fn doc_attribute_on_module() { | |||
.run(); | |||
} | |||
|
|||
#[test] | |||
fn group_attribute_on_module() { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We should have a test for the ordering of submodules in groups if --unsorted
is used.
My first thought was that submodules should be ordered based on their position in the justfile
. However, I'm now not sure if that's correct. I think it's natural to have submodules all declared together at the top or bottom of a justfile
, and natural to list submodules at the end of each group.
So maybe my comment above is off, and we should, instead of using submodule.name.unwrap().offset
in the ordering key, use usize::MAX
instead.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hm, I tested setting submodule.name.unwrap().offset
vs. usize::MAX
, and both have the same behavior. Both implementations output the submodules at the end of a group because of how we output recipes and then submodules in subcommand.rs
:
for (i, group) in ordered.into_iter().enumerate() {
for recipe in recipes {...}
for (i, submodule) in submodules.iter().enumerate() {...}
}
If we wanted the ability to order based on the submodule's position in the justfile
, then we would need to interleave the recipes and submodules. But since it sounds like we want them at the end of a group, I'll go with usize::MAX
. Let me know if you disagree.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I added tests for --unsorted
with and without --list-submodules
. I ended up going with submodule.name.unwrap().offset
because I think that better guarantees the submodules are ordered how they appear in the justfile
at the bottom of a group.
69cca4a
to
c1eeda3
Compare
@casey Thanks for the all the great feedback and the |
c1eeda3
to
20a852c
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice, looks good to me!
Hello! I've been working on a fix for #2243.
I added the whole
attributes
object instead of just passing around the group like how thedoc
is. My rationale for this was to make it easier to have access to anything stored in theattributes
in the future, let me know if that is the wrong path though.This is my first time contributing to
just
, I'm open to any feedback!