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

[Eli Zaretskii <eliz@gnu.org>]

> Date: Mon, 30 Jan 2023 02:49:47 +0200
> From: Dmitry Gutov <dgutov@yandex.ru>
> Cc: casouri@gmail.com, 60953@debbugs.gnu.org
> 
> Code review welcome.

See some below.

> Is applying (and undoing) the narrowing this way legal enough? Or should 
> I go through some error handlers, or ensure blocks, etc?

Mmm... no.  You should use Fnarrow_to_region, I think.

But why do you need to narrow there?  fast_looking_at will not go
beyond end_pos/end_byte anyway, there's no need to restrict it.

Or are you thinking about widening a buffer that is already narrowed?
But if so, can we have parser data beyond the restriction?

> +      Lisp_Object predicates = AREF(predicates_table, match.pattern_index);
> +      if (EQ (predicates, Qt))
> +	{
> +	  predicates = treesit_predicates_for_pattern (treesit_query, 0);
> +	  ASET(predicates_table, match.pattern_index, predicates);

Our style is to leave a blank between ASET and the left parenthesis.

> +  set_buffer_internal (buffer);
> +
> +  TSNode treesit_node = XTS_NODE (node)->node;
> +  ptrdiff_t visible_beg = XTS_PARSER (XTS_NODE (node)->parser)->visible_beg;
> +  uint32_t start_byte_offset = ts_node_start_byte (treesit_node);
> +  uint32_t end_byte_offset = ts_node_end_byte (treesit_node);
> +  ptrdiff_t start_byte = visible_beg + start_byte_offset;
> +  ptrdiff_t end_byte = visible_beg + end_byte_offset;
> +  ptrdiff_t start_pos = buf_bytepos_to_charpos (buffer, start_byte);
> +  ptrdiff_t end_pos = buf_bytepos_to_charpos (buffer, end_byte);
> +  ptrdiff_t old_begv = BEGV;
> +  ptrdiff_t old_zv = ZV;

Since you switch to BUFFER, you can use BYTE_TO_CHAR, no need for
buf_bytepos_to_charpos.

> +  SET_BUF_BEGV(buffer, start_pos);
> +  SET_BUF_ZV(buffer, end_pos);

And here I suggest an additional optimization, since you already know
the byte positions:

  BEGV = start_pos;
  BEGV_BYTE = start_byte;
  ZV = end_pos;
  ZV_BYTE = end_byte;