On 02-09-11 07.42, "Kras Kresh" <kras_kresh@HOTMAIL.COM> wrote: >> From: George Greer <greerga@CIRCLEMUD.ORG> >> Reply-To: Circle Discussion List <CIRCLE@post.queensu.ca> >> To: CIRCLE@post.queensu.ca >> Subject: Re: [CIRCLE] write_to_descriptor bug >> Date: Tue, 10 Sep 2002 14:12:06 -0700 >> >> On Tue, 10 Sep 2002, Kras Kresh wrote: >> >>> WARNING: write_to_descriptor: socket write would block, about to close >>> >>> I'm getting that, then the mud crashes. Any ideas??? >> >> It means some user had so much data that the kernel's socket buffer is >> full, CircleMUD's internal buffer is full, and they were sent yet more >> information. At that point there's nowhere to put the data so they get >> punted from the MUD. It shouldn't crash you though. > > Could it have to do with my updating only part of the cvs? I'm pretty sure > comm.c file is completely updated with my adjustments inside which shouldn't > cause any problems from what you're saying about what it does. Does any > functions that may relate to this be in another CircleMUD file that I have > not completely updated? Do you know when it happens? Perhaps when somebody is in OLC, or an editor, or could it perhaps be that your settings are so high for the buffers that Circle just protests this way? Would be easier to discover the nature and cause of the problem if you did more extensive testing and sent your results to the list. Regards, /Torgny -- +---------------------------------------------------------------+ | FAQ: http://qsilver.queensu.ca/~fletchra/Circle/list-faq.html | | Archives: http://post.queensu.ca/listserv/wwwarch/circle.html | | Newbie List: http://groups.yahoo.com/group/circle-newbies/ | +---------------------------------------------------------------+
This archive was generated by hypermail 2b30 : 06/25/03 PDT