bug#60953: The :match predicate with large regexp in tree-sitter font-lock seems inefficient

[Dmitry Gutov]

On 30/01/2023 20:42, Eli Zaretskii wrote:
> > Date: Mon, 30 Jan 2023 20:20:46 +0200
> > Cc: casouri@gmail.com, 60953@debbugs.gnu.org
> > From: Dmitry Gutov <dgutov@yandex.ru>
> >
> > On 30/01/2023 19:49, Eli Zaretskii wrote:
> > > > Date: Mon, 30 Jan 2023 19:15:07 +0200
> > > > Cc: casouri@gmail.com, 60953@debbugs.gnu.org
> > > > From: Dmitry Gutov <dgutov@yandex.ru>
> > > >
> > > > > fast_looking_at already does an anchored match, so I'm not sure I
> > > > > follow.  I don't even understand why you need th \` part, when the
> > > > > match will either always start from the first position or fail.
> > > >
> > > > The regexp might include the anchors, or it might not.
> > > >
> > > > It might also use a different anchor like ^ or $ or \b.
> > >
> > > OK, but it always goes only forward, so narrowing to the beginning
> > > shouldn't be necessary.  Right?
> >
> > Are you saying that fast_looking_at ("\\`", ...) will always succeed?
> >
> > And fast_looking_at ("^", ...), etc.
>
> For example, for "^", if you hint that it must look back to make sure
> there's a newline there, then your narrowing will also prevent it from
> doing that, right?

fast_looking_at ("^", ...) succeeds inside a narrowing because it always 
succeeds at BOB. Even though there are no physical newlines before BOB.

> > > > One possible alternative, I suppose, would be to create a raw pointer to
> > > > a part of the buffer text and call re_search directly specifying the
> > > > known length of the node in bytes. If buffer text is one contiguous
> > > > region in memory, that is.
> > >
> > > It isn't, though: there's the gap.  Which is why doing this is not
> > > recommended; instead, use something like search_buffer_re, which
> > > already handles this complication for you.  (Except that
> > > search_buffer_re is a static function, so only code in search.c can
> > > use it.  So you'd need to make it non-static.)
> >
> > Interesting. Does search_buffer_re match the \` anchor at POS and \' at
> > LIM? IOW, does in treat the rest of the buffer as non-existing? Or could it?
>
> That is the low-level subroutine called by re-search-forward, so you
> know the answers already, I think?  IOW, that function behaves exactly
> like re-search-forward in those situations.

So, I suppose not?

But that doesn't answer the question "Could it?".