Re: [CODE] Zedit (yea, again...)

From: Wayde Milas (thebard@RARLOA-4.PR.MCS.NET)
Date: 07/23/97

>Ok, before you all tear my head off about looking in the mailing list
>archives, I have. Now, let me explain, and maybe someone has a clue...
>In zedit_new_zone commented out the following:
>        new_table[i] = zone_table[i];
>    else
>      new_table[i+1] = zone_table[i];
>  //free(zone_table); Why? Looks valid?
>  zone_table = new_table;
>  top_of_zone_table++;
>If I remember right, this is the prescribed fix for the zedit new crash.
>Ok, first off, after looking at zone_table with gdb, it looks valid. Its
>not a null pointer. Not only that, but the logic seems correct. Why does
>this free cause problems?
>Second, it seems commenting out this line does not fix everything. After
>compiling, and running, if I do a zedit new 151 say, and then do a show
>zones, the last couple zones are all fuxored. total garbage. This leads me
>to believe that something is wrong with the creation/offset?
>Third, I remember reading something about someone taking purify to this
>chunk of code and figuring out what was going on a couple months back.
>Looks like the mailing archives are current only up to febuary. Can anyone
>point me in the right direction here. I'm setting up a large mud with
>scores of builders, and olc is of the utmost importance to keep things
>BTW, about purify, I believe he deleted a the table with an offset or
>something? Anyone remember?

Just an update.. upgraded to glibc 5.4.33 from 5.3.xx. Zedit new still
chokes. Takes a bit longer, the crasg is not immediate, but still a crash.

system - Redhat 4.2
Dual PPro. pretty much stock redhat. gcc

Any ideas anyone?

Thanks, Wayde

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

This archive was generated by hypermail 2b30 : 12/08/00 PST