fix(elixir): separate function from parameter matching #383
Conversation
Previous queries were mixing the parameter matches and the function matches into the same query. This would make it so that functions that didn't have parameters wouldn't match @function.inner or @function.outer queries. Separating them makes it easier to focus on the specifics of fulfilling each query. This should resolve issue nvim-treesitter#249.
skbolton
force-pushed
the
fix-elixir-function-bodies
branch
from
b2f62b2 to
5822de7
Compare
|
Sorry @kiyoon for the barage of PR's I was trying to break up my work into smaller pieces. I wasn't sure all who is involved in this project and how many eyes would see it. |
|
When I test it, it seems to be not working with functions with parameters in #249 examples? |
|
Thanks a lot for your work, and no worries, it makes it easier to focus on one problem at a time. |
There was a problem hiding this comment.
Changing the following appears to address the remaining issues mentioned in #249.
The binary_operator case captures the guard scenario.
The identifier case addresses another similar issue when the braces are omitted for parameterless functions:
def foo, do: 12
def foo do
12
end
| (do_block "do" . (_) @_do (_) @_end . "end") | ||
| (do_block "do" . ((_) @_do) @_end . "end") | ||
| ] | ||
| (arguments (call)) |
There was a problem hiding this comment.
| (arguments (call)) | |
| (arguments ([ | |
| (call) | |
| (identifier) | |
| (binary_operator) | |
| ])) |
| "defnp" | ||
| "defp" | ||
| )) | ||
| (arguments (call)) |
There was a problem hiding this comment.
| (arguments (call)) | |
| (arguments ([ | |
| (call) | |
| (identifier) | |
| (binary_operator) | |
| ])) |
| (arguments ((_) @parameter.inner) @_delimiter .) | ||
| ] (#make-range! "parameter.outer" @parameter.inner @_delimiter)) | ||
| (arguments | ||
| (call) |
There was a problem hiding this comment.
| (call) | |
| ([ | |
| (call) | |
| (identifier) | |
| (binary_operator) | |
| ]) |
|
Thanks for the reminder on this PR @yihuikhuu. I'll take another stab at this incorporating your feedback this week. |
|
I should have checked existing PRs before opening #605 😅 Mine seems to be a lot simpler, but maybe it's too generic? I've tested it a bit and it seems to match all call parameters in every situation (and it helps simplify other queries as well) without breaking other stuff, but maybe I'm missing something. |
Previous queries were mixing the parameter matches and the function
matches into the same query. This would make it so that functions that
didn't have parameters wouldn't match @function.inner or @function.outer
queries.
Separating them makes it easier to focus on the specifics of fulfilling
each query. This should resolve issue #249.