On Sun, 16 Jun 1996, circle wrote: > I have found this elf /a.out descussion wierd since I am using elf and I > compiled circle patch 11 with no problems and no warnnings first shot. I > will warn you though when I tryed first I had a bad header file in the > includes on my system and I got tunes of errors and when I did get it to > work I got tuns of crashes. After much work I found that I had some old > headers and when replaced it compiled flawless. I am not sure if this is > what you are running into but I truthfully don't think it is elf. I use ELF on my Linux system, and it works beautifully. I haven't encountered a single error in compilation of the CircleMUD code (except for the warnings in the OLC stuff) and have run into no abnormal errors. If your header files are a problem, well, that's not ELF, that's your kernel. Probably you're using one of the advanced beta kernels (1.3.xx) that hasn't got (m)any of the bugs squashed yet. Going back to 1.2.13 should work okay, though it may damage some of your kernel's functionality. I assume that's what you mean when you say that you used an older version of the headers. > On Sun, 16 Jun 1996, Hades wrote: > > > Is the diffrences between ELF and .aout worth bothering to deal with all the > > hassles and bugs of the elf code? I am this close to going back to a.out. > > WHat are the actual diffrences? I'm fairly sure that there are no bugs in the "elf code." From what I understand, one of the major differences is that the elf programs make extensive use of the load-on-demand libraries. I thought you could do this with a.out, too, but I don't know. Anyway, the point is: ELF binaries work fine with Circle on (at least) kernel 1.2.13. If you're using this kernel and elf binaries and you're sure you have all the correct elf utilities that you need, the problem isn't elf at all, but something else. If you're using a beta kernel, well, you have to expect errors on those. I *strongly* recommend moving back to the last stable Linux version, which I'm pretty sure is 1.2.13. This will save you hours of hair-pulling and recompiling no matter what you use Linux for. The only legitimate reason for using a beta kernel for a system that's supposed to be stable (that I can see) is that you need some of the new drivers that aren't available in other versions. Of course, you're welcome to use a beta if you need the bleeding-edge features. But you also have to expect bugs. Sorry about all of this. Rant off, etc. So's I don't contradict an earlier message of mine entirely, here's a CircleMUD question. Does anyone have any Circle statistics? i.e., how many users a CircleMUD running Linux on a P75 w/16MB of RAM can handle before choking (I consider "choking" a 50% cut in responsiveness)? If someone can illustrate a sort of bell curve or anything this way, I'd be incredibly pleased. I guess what I want is more of a survey. How many users can your MUD get, all active, before performance drops below what you consider to be an acceptable level? Failing that, before it drops below the level you want it at. Do you run out of memory or processor cycles first? Have you found that running on one operating system over another makes for speed increases? That sort of thing. A performance evaluation of CircleMUD. I'd assume that Mr Elson has some sort of information like this someplace, though it's probably fairly old, too. If anyone else is interested in this, I'd be willing to write back with the results, assuming I get any. Please reply to me, personally, by the way; this doesn't really qualify as general-interest knowledge. ------------------------------------------------------------------ Barid Bel Medar icarus@berkshire.net Knights of the Cosmos Shayol Ghul Resort and Health Spa ------------------------------------------------------------------ "I am returning this otherwise good typing paper to you because someone has printed gibberish all over it and put your name at the top." - English Professor, Ohio University ------------------------------------------------------------------
This archive was generated by hypermail 2b30 : 12/18/00 PST