Hotboot, duplicate corpse items

Posted by Zeno on Sat 02 Oct 2004 05:52 AM — 11 posts, 34,786 views.

Zeno USA Sat 02 Oct 2004 05:52 AM #0

Seems that everytime I hotboot, it duplicates items in a corpse?

Quote:
Tarnoth chats 'umm i kinda need a imms help all the changes tripled the stuff on my corpse ;( and now i have two exta dragon arms that i cant drop'

Its true, had to check it myself. No idea what to look for. Anyone know?

Greven Canada Sat 02 Oct 2004 03:25 PM #1

Are you using Samson's hotboot? If so, you may have missed the step that stops writing corpses, and then you get multiple ones.

Zeno USA Sat 02 Oct 2004 04:15 PM #2

Yeah I am. Don't think I missed any steps. One second.

[EDIT] Checked the snippet intructions, I didn't miss anything to do with corpses in save.c, and I don't think I missed anything at all. As I said, corpses are fine, but items in the corpse are not.


    if ( obj->prev_content && os_type != OS_CORPSE )
        if( os_type == OS_CARRY )
           fwrite_obj( ch, obj->prev_content, fp, iNest, OS_CARRY, hotboot );

Samson USA Sat 02 Oct 2004 05:52 PM #3

YOu should have had a step which told you to do this:

Quote:

Then locate this segment:

/* Corpse saving. -- Altrag */
fprintf( fp, (os_type == OS_CORPSE ? "#CORPSE\n" : "#OBJECT\n") );

Directly ABOVE that, add this:

/* DO NOT save corpses lying on the ground as a hotboot item, they already saved elsewhere! - Samson */
if( hotboot && obj->item_type == ITEM_CORPSE_PC )
return;

That is what stops a hotboot from saving a corpse since corpses are already saved in another section.

Zeno USA Sat 02 Oct 2004 06:25 PM #4

Yeah. Already in.

    /*    Munch magic flagged containers for now - bandaid */
    if ( obj->item_type == ITEM_CONTAINER
    &&   IS_OBJ_STAT( obj, ITEM_MAGIC ) )
        xTOGGLE_BIT( obj->extra_flags, ITEM_MAGIC );

    /* DO NOT save corpses lying on the ground as a hotboot item, they already saved elsewhere! $
    if( hotboot && obj->item_type == ITEM_CORPSE_PC )
        return;

    /* Corpse saving. -- Altrag */
    fprintf( fp, (os_type == OS_CORPSE ? "#CORPSE\n" : "#OBJECT\n") );

    if ( iNest )

Greven Canada Sun 03 Oct 2004 05:09 AM #5

Can you confirm that this didn't happen before you put hotboot? It may be a problem with the corpse code, unlikely, but possible.

Zeno USA Sun 03 Oct 2004 05:33 AM #6

Hmmmm. Yes, I believe it was fine pre-hotboot. I didn't put in hotboot until a while after I started the MUD.

Greven Canada Sun 03 Oct 2004 05:59 AM #7

Hmmm, well from the little code thats here it looks like everything should be good. You might want to try booting in gdb, putting a break on something that will specifically target the items in the corpse so you can get the memory adress of the corpse. You can then remove that break, and put a conditional break in fwrite_obj, checking that the object equals the memory adress you got before. You could also set up a conditional check to check the name of the items. That may give you some more info, you would be able to find out where its being written from, and narrow your search down a little.

Zeno USA Sat 05 Mar 2005 03:38 PM #8

Hmm it's still happening of course. There was one shard in the corpse:

[ 2681] Derius: Sat Mar  5 07:23:50 2005 :: Derius(0) got a purified shard from in corpse Derius(11) in room 2681.
(hotboot)
[ 2681] Derius: Sat Mar  5 08:11:11 2005 :: Derius(0) got a purified shard from in corpse Derius(11) in room 2681.

I compared write_corpses to current SmaugFUSS, no difference. Calls to fwrite_obj also have no difference.
[EDIT] Just checked the function fwrite_obj, no difference.
(besides what object fields I have added)

Also I cannot reproduce this every time.

Asean Novari USA Sat 05 Mar 2005 06:40 PM #9

I had this problem a bit ago.. it fixed itself and went on to a different section.. for me its now happening to lockers.

Zeno USA Sat 05 Mar 2005 09:50 PM #10

How did you fix it?