Hi!
On Jul 10, Sergey Vojtovich wrote:
> ChangeSet@stripped, 2007-07-10 15:37:39+05:00, svoj@stripped +3 -0
> BUG#29445 - match ... against () never returns
>
> Part 1:
> Do not perform index search for a plus-word with truncation operator
> in case there are other plus-words in expression.
>
> Specifically this optimization was enforced for expression like
> "+word1 +word2*", but wasn't enforced for expression like
> "+word1 word2*".
>
> diff -Nrup a/storage/myisam/ft_boolean_search.c b/storage/myisam/ft_boolean_search.c
> --- a/storage/myisam/ft_boolean_search.c 2007-06-01 00:28:00 +05:00
> +++ b/storage/myisam/ft_boolean_search.c 2007-07-10 15:37:37 +05:00
> @@ -474,7 +474,8 @@ static void _ftb_init_index_search(FT_IN
> ftbe->up->flags|= FTB_FLAG_TRUNC, ftbe=ftbe->up)
> {
> if (ftbe->flags & FTB_FLAG_NO || /* 2 */
> - ftbe->up->ythresh - ftbe->up->yweaks >1) /* 1 */
> + ftbe->up->ythresh - ftbe->up->yweaks >
> + test(ftbe->flags & FTB_FLAG_YES)) /* 1 */
> {
> FTB_EXPR *top_ftbe=ftbe->up;
> ftbw->docid[0]=HA_OFFSET_ERROR;
Okay, good.
But notice, this is inside a for() loop going up the expression tree,
and the comment says "in 1 and 2, +/- need not be on the same expr.
level, but can be on any upper level, as in +word +(trunc1* trunc2*)".
Please add a test case for this (you can use fulltext.test for
inspiration :).
ok to push with a test case.
Regards / Mit vielen Grüssen,
Sergei
--
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Sergei Golubchik <serg@stripped>
/ /|_/ / // /\ \/ /_/ / /__ Principal Software Developer
/_/ /_/\_, /___/\___\_\___/ MySQL GmbH, Radlkoferstr. 2, D-81373 München
<___/ Geschäftsführer: Kaj Arnö - HRB
München 162140
