Re: Tree sitter support for C-like languages

[Yuan Fu]

> On Nov 14, 2022, at 12:10 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> 
>> From: Yuan Fu <casouri@gmail.com>
>> Date: Mon, 14 Nov 2022 11:51:59 -0800
>> Cc: monnier@iro.umontreal.ca,
>> theo@thornhill.no,
>> emacs-devel@gnu.org
>> 
>>>> Well we update the parse tree and re-parse, but we currently don’t update 
>>>> the nodes created from the old tree. Keeping all nodes updated requires us 
>>>> to track all live nodes and update them whenever the buffer is edited.
>>> 
>>> I guess I still don't understand what exactly do you mean by "update
>>> the node".  Can you explain that in more detail?
>> 
>> My bad. So when buffer changes (insert in X, delete from X to Y), we inform 
>> tree-sitter of this change by “updating” the tree:
>> 
>>  const TSInputEdit edit =
>>    treesit_prepare_input_edit (start_byte, old_end_byte, new_end_byte);
>> 
>>  ts_tree_edit (tree, &edit);
>> 
>> Then when we re-parse, tree-sitter knows which part of the buffer has 
>> changed and needs to be re-parsed, and only parses those, hence “incremental 
>> parsing”. 
>> 
>> Tree-sitter nodes needs similar updates, so that it is in sync with the 
>> buffer text.
> 
> Doesn't the call to ts_tree_edit update those nodes?  

No.

> That is, aren't
> those nodes a part of the tree that gets updated by the ts_tree_edit
> call?

The node stores some information in itself (start_byte, end_byte, inlined data, 
etc), and references the tree for the rest. The information it stores needs to 
be updated separately.

Yuan