RE: Huge Medit Bug.

From: Darklord (
Date: 12/16/96

      Ok, one of my coders installed improved_edit.c, the patch.  He
had to do it manually because our Circle code has been modified so
greatly, that it's almost unrecognizable to be circle.  After that he
also installed tfiles_editor.c, also the patch.  It seems to be working
except for medit.  Since then, we have been crashing when using medit
without fail.  I have tried to track this down myself, but I am looking
for a better coder out there to help me with this problem.  Here's what's
up:  You edit a new mob, ok.. looking good.  It lets you edit it etc.
(BTW: Using Oasis).  When you get prompted, as you're leaving medit, to
save the mobile, you type: yes you crash.  Gdb produces the following:

AB> Core was generated by `bin/circle 1234'.
AB> Program terminated with signal 11, Segmentation fault.
AB> Reading symbols from /lib/
AB> Reading symbols from /lib/
AB> #0  0x40079e1d in free ()
AB> (gdb) bt
AB> #0  0x40079e1d in free ()
AB> #1  0x4017baf4 in ?? ()
AB> #2  0x1 in ?? ()
AB> #3  0x104ec in ?? ()
AB> Cannot access memory at address 0x81e58955.

Ok, next bug is when you edit an already existing mob.  It let's you edit
it, AND save it.  But when you go to edit another mob, existing or new
mob, it crashes right away.  Doesn't even prompt you with the medit
menu.  Gdb produces the following:

AB> Core was generated by `bin/circle 1234'.
AB> Program terminated with signal 11, Segmentation fault.
AB> Reading symbols from /lib/
AB> Reading symbols from /lib/
AB> #0  0x400798f8 in malloc ()
AB> (gdb) bt
AB> #0  0x400798f8 in malloc ()
AB> #1  0x40089418 in _libc_cat ()
AB> #2  0x2d1d in ?? ()
AB> Cannot access memory at address 0x2e12.
I have spent the last week trying to track this down, so if anyone has
any suggestions, I would love some assistance on this.


Fate, Implementor of the Ruins.
Best Kickass Mud Around. 8000

 Ok... first of all.. i have had this program for what seems like 2 years..
i have also had no problems with it.. execpt for medit in exactly the same
manner.. however.. here is where your problems lie.. i dont use gdb since
i am good enough to track stuff down :P

  Anyway.. for medit.. first bug you mentioned.. lies in the medit_save_to_disk
this is where it is.. i know for a fact.. because i can edit the mobs save
them to memory.. no problem.. but its the write to the disk thats where the
screwup is.

  The second bug is also in there on the create existing mob proto... i believe
or when it tries to allocate memory for an existing mob.. etc..

  So far i have only had it crash on me when trieg to edit an existing mob
for zone 12  only.. the immortal zone.. i duno whats whith that.. i made
new mobs for it.. imported other mobs into that zone.. i tried it all.

  Also.. which is really wierd.. i have a heavily modified mud..and i have
the mobprogs and all.. so i stuck that save to disk thing in the medit.
(this was when it somehow worked the first time) so then after a year .. i was
just fooling around and took that part out.. (mobprog stuff out of medit) then
it worked again.. no problems.. well 

  One week later.. i didnt do nothin to it .. just ading some ACMD and stuff
like that..  the medit stoped working again.. and acts up just like you mentioned.

  Now.. i dont run all my areas in the mud.,. because for one.. theres not
enough people to open a second port for testing purposes only.. and second
because it takes to long to load when i have to compile about every second.

  Well.. i have also tried adding just the medit on a stock base.. and i 
get the same errors.. in fact i still have it ready to go.  I am convienced
it is just a memory leak somewhere.. and the error message is "Buss Error"
thats all i know.. 

  The only recomendation i can see.. is to take out all your areas.. start
new.. and then trouble shoot from there.. one time i took out a few zones
and it worked again.. so somethign in the memory might have got conflicted
on boot up..

| Ensure that you have read the CircleMUD Mailing List FAQ: |
|   |

This archive was generated by hypermail 2b30 : 12/18/00 PST