feat: allow fixup and amend in subject #48
Conversation
gbprod
force-pushed
the
feat-fixup-amend
branch
7 times, most recently
from
9782144
to
b84b3dc
Compare
gbprod
force-pushed
the
feat-fixup-amend
branch
from
b84b3dc
to
4965426
Compare
| @@ -33,6 +33,9 @@ module.exports = grammar({ | |||
|
|
|||
| subject: ($) => | |||
| seq( | |||
| optional( | |||
| seq(alias(choice('fixup!', 'amend!'), $.subject_prefix), WHITESPACE) | |||
There was a problem hiding this comment.
node_type subject_prefix is undefined. nvim-treesitter/nvim-treesitter#5774 (comment).
We'll also need:
subject_prefix: ($) => ...
There was a problem hiding this comment.
It's an alias, you don't need any rule associated to the name (Ref: https://tree-sitter.github.io/tree-sitter/creating-parsers#the-grammar-dsl)
I'm not sure that it's the origin of your issue, have you updated your parser ? :TSUpdate gitcommit ?
There was a problem hiding this comment.
If you are referring to anonymous nodes, the name part might need to be 'subject_prefix' not $.subject_prefix. See
nvim-treesitter/nvim-treesitter#5774 (comment) --- I'm pretty sure the parser is up-to-date. Maybe it's fine on the treesitter side (because the test parse), but somehow nvim's treesitter parse does not recognize this rule (or node_type). There is clearly an error on the neovim's query parsing side.
There was a problem hiding this comment.
It's definitely fine. For example,
https://github.com/tree-sitter/tree-sitter-c/blob/25371f9448b97c55b853a6ee8bb0bfb1bca6da9f/grammar.js#L161-L163
There was a problem hiding this comment.
If you look at the generated parser.c, neither ts_symbol_metadata nor ts_symbol_names, etc. contains sym_subject_prefix. I think the neovim's bundled libtree-sitter (I tested with 0.9.4 and 0.10-devel on macOS) is implemented in such a way that if that is not recognized as a valid node type, a TSQueryErrorNodeType error will be thrown. Maybe we have a different tree-sitter lib/version linked. Anyway, the generated parser's node definition doesn't look correct to me.
There was a problem hiding this comment.
Sounds like a tree-sitter-grammars/tree-sitter-markdown@110b89d
There was a problem hiding this comment.
It's definitely fine. For example,
tree-sitter/tree-sitter-c@25371f9/grammar.js#L161-L163
alias($.preproc_unary_expression, $.unary_expression),
alias($.preproc_binary_expression, $.binary_expression),
alias($.preproc_parenthesized_expression, $.parenthesized_expression),
I meant, it's fine to use $.unary_expression but the difference here is that there exist rules for all these three node types. https://github.com/tree-sitter/tree-sitter-c/blob/25371f9448b97c55b853a6ee8bb0bfb1bca6da9f/grammar.js#L929. $.subject_prefix (gitcommit) does not.
And I didn't re-generate parser.c, the one that was built in commit 7e3ad5f was used so I'm not sure if this is x86 vs arm64 issue.
There was a problem hiding this comment.
Thanks for investigations @wookayin but I'm retty sure that's it's not a grammar issue.
@lucario387 I'm building on x86 but I've switched to node 20 and node-gyp 10, maybe I should retry using node 18 and node-gyp 9 ?
Thanks for your help :)
There was a problem hiding this comment.
I don't know, I got the same generated parser.c as you, running on x86 :)
@amaanq will need to jump in to help here I feel
There was a problem hiding this comment.
I've tried to build using docker, node 18, node 20, node-gyp 9, node-gyp 10, python 3.11, python 3.12 but I still get the same parser.c...
I've created an issue here if you prefer : #51
This will parse git commit subject that are prefixed with
fuxup!oramend!.Ref: #34