Hi!
On Nov 29, Tatjana Azundris Nurnberg wrote:
> Sergei Golubchik <serg@stripped> writes:
>
> > good. just one question:
> > you can have signed variables (GET_INT, GET_LONG, GET_LL) in
> > getopt_ull_limit_value() ?
>
> Yes, isn't it fun. ;-/ We're fine as far as command-line options are
> concerned (now that I've corrected the GET_* for innobase to their
> actual types), but you'll notice that we declare ull globally, but
> not ll. Which is because we only use ull in set_var. Nevertheless,
> we have variables like "concurrent-insert" which are eligible for
> SET, are signed, but go through "unsigned" (or rather, "we only care
> for length, not signedness") in set-var. Thanks to the extra code,
> that can no longer give rise to bugs, but I'm the first to agree
> that it isn't as pretty as it could be. May be worthwhile to write
> a ticket "check whether signed server-variables actually need to be
> signed, and if so, create a new signed sys-var type for them."
Can you fix it in this patch ? It's rather illogical that we check
signed types in getopt_ull_limit_value() and not in
getopt_ll_limit_value().
Regards / Mit vielen Grüssen,
Sergei
--
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Sergei Golubchik <serg@stripped>
/ /|_/ / // /\ \/ /_/ / /__ Principal Software Developer
/_/ /_/\_, /___/\___\_\___/ MySQL GmbH, Dachauer Str. 37, D-80335 München
<___/ Geschäftsführer: Kaj Arnö - HRB
München 162140
