On Fri, 13 Oct 1995, Brad Johnson wrote: > I've got circlebpl8 set up on a Linux box with mobprograms thrown in, and > I just made some changes to the code to add in two more equipment > locations. The equipment locations work fine, but now, whenever the > (thankfully) single mob that I have set up to talk says something, the > mud will promptly give me a seg. fault. I'm new at Linux (or Unix for > that matter), so I have no idea how to make it give me a core dump to > look through. If anyone can tell me how to set Linux up to give me a > dump when it crashes, or better yet, knows how I might fix the problem > with the mobprograms, I'd be very grateful! Be sure to compile with the flags to create debug info. I /THINK/ it's -g, but since the mud's down for a little while, I can't be sure. try 'man gcc' or 'man cc' depending on which compiler you use (see your Makefile) then: find the core. It could be that there isn't one, and then you have to unlimit the coresize. how? depends on your shell etc. csh: unlimit coredumpsize The rest, I don't know. It could be that even that won't help, and then you'd have to talk to your sysadmin, but I doubt that such will be the case. then: cross your fingers, go to the dir where core is, and type 'gdb' If that command wasn't found, sysadmin again. In gdb, type core core, and it will load and examine the core. then, type backtrace. It will tell you in what function it crashed, by what functions it was called, etc. use print to check out vars, e.g. print *ch or print ch->name and so on. Check out help, too :) Gekke Eekhoorn of BUGmud.
This archive was generated by hypermail 2b30 : 12/07/00 PST