Hi Juha,
I just finished merging your patch with the latest usrloc/registrar
changes.
one of the biggest changes was an optimization in case of a register
with multiple contacts - common info (callid, cseq, received, socket,
etc) are calculate only once.
Also I switched TABLE_VERSION to UL_TABLE_VERSION - makes more sense..
Regarding the 0xFFFFFFFF value, I will take a look into db/mysql stuff
to see what about.
regards,
bogdan
Juha Heinanen wrote:
Bogdan-Andrei Iancu writes:
> thanks for the patch - I will committed ASAP- I will have to merge it
> with some changes on usrloc I working on right now....it's just a matter
> of couple of days.
fine, i'm not in a hurry with it. if you make changes to usrloc module,
one other thing i noticed is that it contains location table version
twice. in ul_mod.h:
#define TABLE_VERSION 1001
and in udomain.c:
#define UL_TABLE_VERSION 1001
in my patch i changed the latter to
#define UL_TABLE_VERSION TABLE_VERSION
but it would be cleaner to rename in ul_mod.h TABLE_VERSION to
UL_TABLE_VERSION and remove the define from udomain.c.
i also noticed that i was not able to write all ones (0xFFFFFFFF) to
methods column of location table. i first tried to copy how flags was
written:
vals[6].type = DB_INT;
vals[6].nul = 0;
vals[6].val.bitmap_val = _c->flags;
but got 0 to the table with value 0xFFFFFFFF. i then tried to change
DB_INT to DB_BITMAP, but that didn't help either. i suspect that there
is a bug in db or mysql code. finally i just tested if value is
0xFFFFFFFF and if so, wrote NULL.
-- juha
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
