Return to BSD News archive
Path: sserve!manuel.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!elroy.jpl.nasa.gov!usc!wupost!darwin.sura.net!Sirius.dfn.de!news.dfn.de!Germany.EU.net!mcsun!sunic!isgate!krafla!adamd From: adamd@rhi.hi.is (Adam David) Newsgroups: comp.os.386bsd.questions Subject: Re: 386BSD & MUDs Keywords: Muds, argh!!! Message-ID: <6502@krafla.rhi.hi.is> Date: 7 Mar 93 15:34:05 GMT References: <1993Mar7.055922.7770@wam.umd.edu> Sender: usenet@rhi.hi.is Lines: 40 Nntp-Posting-Host: hengill.rhi.hi.is publius@wam.umd.edu (Inigo Montoya) writes: >Has anyone successfully compiled a MUD under 386BSD? First I tried MudOS 0.9 >No luck...then I went to LPMud 3.1.2 No luck either. With MudOS, it got a >little of the way through the compilation and just stopped... seems to have >crashed the system or something, as I can no longer ^Z the process. With >LPMud, cc seems to crash altogether, aborting with a signal 10 or something. >I then find a core dump in the current directory. This probably has nothing whatsoever to do with the MUD source. There are bugs in 386bsd that make compiling a test of mental stamina. >What is the problem here? Are these all bugs? Do I need the patchkits? Or do I >just not have enough RAM (4 megs) to do such compilations? Yes, you do need the patchkits (make sure you get the latest). There is a major bootstrap problem because the unpatched kernel is simply not stable enough to compile the kernel and other necessary support files in order to upgrade. (it can be done but don't expect to retain any vestige of sanity). Therefore it is probably best to get hold of a kernel which is already compiled and use that while you are compiling your kernel. Make sure patch 83 is excluded from whatever kernel you are using, it can always be added later if necessary. With patchkit 0.2.1 the system is stable enough to compile on, but random crashes do still occur - it does not make any difference whether the code being compiled is simple or complicated, large or small, few files or many. It is truly random, and even happens immediately after a "normal" (?) shutdown-startup sequence (or after half an hour of intensive compiling). Expect this to stabilise somewhat when the FS buffer cache gets fixed. p.s. my apologies for sounding so negative, my machine got caught in a crash-reboot-crash loop 3 times yesterday (needed on average 2 hours spoonfeeding each time), and today I decided to recompile everything (to make sure) and crashed about 10 times in a row for each of several trivial and small files in libc.a (rebooting twice each time). I've got to go now, I'll be busy again (its the compiler next). -- Adam David (adam@veda.is)