Re: Tree sitter support for C-like languages

[Theodor Thornhill]

Eli Zaretskii <eliz@gnu.org> writes:

>> Cc: emacs-devel@gnu.org
>> Date: Fri, 11 Nov 2022 07:01:02 +0100
>> From:  Theodor Thornhill via "Emacs development discussions." 
>> <emacs-devel@gnu.org>
>> 
>> See below patch.  The struct issue was an ordering problem when I
>> extracted the styles.  Should be good now
>
> Thanks, a few minor comments below.
>

Thanks!

>> diff --git a/etc/NEWS b/etc/NEWS
>> index 9ed78bc6b3..3ce9810ece 100644
>> --- a/etc/NEWS
>> +++ b/etc/NEWS
>> @@ -2786,8 +2786,32 @@ when visiting JSON files.
>>  ** New mode ts-mode'.
>>  Support is added for TypeScript, based on the new integration with
>>  Tree-Sitter. There's support for font-locking, indentation and
>> -navigation.  Tree-Sitter is required for this mode to function, but if
>> -it is not available, we will default to use 'js-mode'.
>> +navigation.  Tree-Sitter is required for this mode to function.
>> +
>> +** New mode c-ts-mode'.
>> +Support is added for C, based on the new integration with
>> +Tree-Sitter. There's support for font-locking, indentation and
>> +navigation.  Tree-Sitter is required for this mode to function.
>
> First, please use 2 spaces between sentences.
>

Yeah - my slipup.

> Next, I suggest a slightly different boilerplate for these items:
>
>   ** New mode FOO-ts-mode.
>   A major mode based on the Tree-sitter library for editing programs
>   in the FOO language.  It includes support for font-locking,
>   indentation, and navigation.
>

Sure - adopted.

>> +(defvar c++-ts-mode--preproc-keywords
>> +  '("#define" "#if" "#ifdef" "#ifndef" "#else" "#elif" "#endif" "#include")
>> +  "C++ keywords for tree-sitter font-locking.")
>
> This list seems to be incomplete: what about the following?
>
>   #error
>   #warning
>   #include_next
>   #line
>   #pragma
>
> Doesn't Tree-sitter's C++ grammar support those?
>

Doesn't seem like it does, no.


>> +(defvar c-ts-mode--keywords
>> +  '("const" "default" "enum" "extern" "inline" "static"
>> +    "struct" "typedef" "union" "volatile" "goto" "register"
>> +    "sizeof" "return"
>> +    "while" "for" "do" "continue" "break"
>> +    "if" "else" "case" "switch")
>> +  "C keywords for tree-sitter font-locking.")
>
> Are these all the keywords in C?  Or is it just what Tree-sitter's C
> parser support for now?
>

I added a few more.  Some were missing, that's true.

>> +(defvar c-ts-mode--preproc-keywords
>> +  '("#define" "#if" "#ifdef" "#ifndef" "#else" "#elif" "#endif"
>> +    "#include")
>> +  "C keywords for tree-sitter font-locking.")
>
> Same comment here about additional preprocessor directives.

Same answer here :)


See below patch.  Please note that I now have merged C and C++ into one
file, and both inherit from a new parent mode: 'c-ts-mode--base-mode'.
I simplified some of the indentation rules as well.

Unless some of you have any strong objections I think applying them to
feature/tree-sitter now is a good time, to allow other people to more
easily giving them a testrun.

What do you think?

Theo

0001-Add-Tree-Sitter-modes-for-C-like-languages.patch
Description: Text Data