Re: How to add pseudo vector types

[Yuan Fu]

> On Jul 27, 2021, at 11:40 PM, Stephen Leake <stephen_leake@stephe-leake.org> 
> wrote:
> 
> Yuan Fu <casouri@gmail.com> writes:
> 
>>>> 
>>>> We need to “delete” the hidden text and “re-insert” when we widen the
>>>> buffer. I’ll try to make it a no-op as long as we remember to widen
>>>> before calling tree-sitter to parse anything.
>>> 
>>> First, the only thing TS deletes is tree nodes, not text; it does not
>>> have a copy of the buffer.
>>> 
>>> Why do you think we need to delete the tree nodes corresponding to the
>>> hidden text? They provide exactly the context needed to parse the
>>> visible text properly.
>> 
>> I don’t think we need to, but I assume that tree-sitter will delete
>> the corresponding nodes if we hide the text from it. 
> 
> No, tree-sitter only deletes nodes that cover changes.
> 
> So don't send a change that deletes the hidden text; just send changes
> in the visible part of the text (that's the only place the user can make
> changes). tree-sitter will only run the scanner on the change regions,
> so it will only request text from the visible part of the buffer;
> all the requests will succeed.

Then we are not hiding the hidden text from tree-sitter. The implementation you 
described, IIUC, is essentially do nothing special when the buffer is narrowed. 

> 
>> For us, the text is there, just hidden; for tree-sitter, the text is
>> deleted.
> 
> No, it simply won't notice that it can't access that part of the buffer,
> because it will never try.
> 
> What, exactly, will the buffer-text fetch code do if tree-sitter
> violates the narrowing (by some error in tree-sitter or user code)?
> throw an exception? return a null string?

In my current implementation, if tree-sitter access buffer content outside of 
narrowed region, it reads whitespaces, if it access buffer content outside of 
the buffer, it reads null string. Neither case throws an error.

Yuan