Hello Dimitrios,
I cannot approve or deny your patch, but I have one question.
[...]
> 2012-06-04 Dimitrios Apostolou <ji...@gmx.net>
>
> * line-map.c (linemap_enter_macro): Don't zero max_column_hint in
> every macro. This improves performance by reducing the number of
> reallocations when track-macro-expansion is on.
[...]
> === modified file 'libcpp/line-map.c'
[...]
> @@ -331,7 +331,6 @@ linemap_enter_macro (struct line_maps *s
> num_tokens * sizeof (source_location));
>
> LINEMAPS_MACRO_CACHE (set) = LINEMAPS_MACRO_USED (set) - 1;
> - set->max_column_hint = 0;
I am wondering why this change implies better performance.
Is this because when we later want to encode a new line/column, and
hit the spot below, (in linemap_line_start via
linemap_position_for_column), we call less linemap_add (and thus
allocate less maps):
if (line_delta < 0
|| last_line != ORDINARY_MAP_STARTING_LINE_NUMBER (map)
|| SOURCE_COLUMN (map, highest) >= (1U << column_bits))
map = (struct line_map *) linemap_add (set, LC_RENAME,
ORDINARY_MAP_IN_SYSTEM_HEADER_P
(map),
ORDINARY_MAP_FILE_NAME (map),
to_line);
The reason why the "if" would fail more, is that the condition
"SOURCE_COLUMN (map, highest) >= (1U << column_bits)
would condition would fail more often, because the column_bits would
be smaller more often.
Does that match your measurements?
Thanks.
--
Dodji
