From: David G. <dav...@po...> - 2006-10-06 01:20:01
|
I'm getting these entries in my error log file: [04/Oct/2006:14:57:13] iocp772 error writing "iocp772": connection reset by peer /baby/birth/firstbath1.png [04/Oct/2006:14:57:17] iocp868 error writing "iocp868": connection reset by peer /baby/birth/grandma.png [04/Oct/2006:14:58:13] iocp772 read error during state start: error reading "iocp772": connection reset by peer [04/Oct/2006:14:58:18] iocp868 read error during state start: error reading "iocp868": connection reset by peer The first 2 are good, but the last 2 are not. I know this is 95% my fault as I'm running some alpha code with IOCPSOCK doing the tcp channel driver, but I'm having a bit of a difficult time reading the tclhttpd script files. I know I should cheat and run all this in a tcl debugger and trace through script execution that way to come to an understanding of it all, but maybe I'm a bit lazy.. Anyways.. I see in Httpd_ReturnFile it calls Httpd_Suspend which turns off the file events and lets fcopy proceed with a background transfer. If the fcopy fails with a transfer error (like ECONNRESET above) see HttpdCopyDone, shouldn't the socket be closed permanently and not just a 1.1 reset? It looks to me like the socket is just being reset and restarted to HttpdRead instead of dumped. /me off to trace this deeper.. most likely this is my fault with my channel driver, but can't seem to isolate where I'm getting tclhttpd confused. -- David Gravereaux <dav...@po...> $ make war make: *** No rule to make target `war'. Stop. Try `love' instead |