GNU bug report logs - #74963
Ambiguous treesit named and anonymous nodes in ruby-ts-mode

Previous Next

Full log

Message #47 received at 74963 <at> debbugs.gnu.org (full text, mbox):

From: Yuan Fu <casouri <at> gmail.com>
To: Juri Linkov <juri <at> linkov.net>
Cc: Dmitry Gutov <dmitry <at> gutov.dev>, 74963 <at> debbugs.gnu.org
Subject: Re: bug#74963: Ambiguous treesit named and anonymous nodes in
 ruby-ts-mode
Date: Sun, 12 Jan 2025 23:47:33 -0800

> On Jan 12, 2025, at 11:31 PM, Juri Linkov <juri <at> linkov.net> wrote:
> 
>> I see that all ts-modes solve this common problem each in its own way
>> (here 'list' indicates a list of strings that should match node names):
>> 
>>  c-ts-mode:    (regexp-opt list 'symbols)
>>  js-ts-mode:   (concat "\\_<" (regexp-opt list t) "\\_>")
>>  java-ts-mode: (rx (or list))
>>  ruby-ts-mode: (rx bol (or list) eol)
>> 
>> Currently there is no uniform way to handle this frequent need.
>> 'concat' like above looks too ugly, but 'regexp-opt' with the
>> 'symbols' arg produces a strange regexp for matching symbols.
> 
> I was thinking about adding two functions treesit-regexp-strict
> and treesit-regexp-lax.  But then discovered that some things
> require specifying both strict and lax matches for the same thing.
> For example, take treesit-thing-settings from c-ts-mode:
> 
>    (sentence
>     ,(regexp-opt '("preproc"
>                    "declaration"
>                    "specifier"
>                    "attributed_statement"
>                    "labeled_statement"
>                    "expression_statement"
>                    "if_statement"
>                    "switch_statement"
>                    "do_statement"
>                    "while_statement"
>                    "for_statement"
>                    "return_statement"
>                    "break_statement"
>                    "continue_statement"
>                    "goto_statement"
>                    "case_statement")))
> 
> "preproc" can be lax, this is fine to match all preprocessor directives.
> But "declaration" should be strict and should not match "parameter_declaration".
> Also "specifier" should not match "attribute_specifier" and "storage_class_specifier",
> but only "enum_specifier" and "union_specifier" that end with the semicolon.
> Also no need to specify all statements separately, it should be sufficient
> to use lax match with "statement".
> 
> The most expressive language to specify all these requirements is the rx macro,
> so let's use it in ts-modes.  Here is how the 'sentence' thing will look like:
> 
>    (sentence
>     ,(rx (or (and bos (or "declaration"
>                           "enum_specifier"
>                           "union_specifier")
>                   eos)
>              (or "preproc"
>                  "statement"))))

Looks good. I’ve always used rx, it has the additional benefit of being macro expanded at compile time.

Also, I finally added support for ‘and’, ‘named’ and ‘anonymous’. I haven’t test it yet (sorry).

Yuan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.