Bogdan Pintea wrote: > Great, looking forward to it. > > Speaking of "<script>", since the patch is not yet applied, I'd sneak in > a feat. req.: I've always though that it would be helpful to also have > the route name (/number) in the log messages generated from script; now, > I add it manually, but I guess that adding it automatically shouldn't be > that difficult.
Sure, I considered this feature (and the possibility to automatically log routes whenever they are entered/left) too but as far as I remember I was told it's not possible to obtain such info now. Anyway, I plan to look at it ...someday. --ondra > > > Bogdan. > > Ondrej Martinek wrote: >> there's an example of the "stderr" log output: >> $ ser -ddDDEf test.cfg >> >> ============== >> SER compiled w/ -DNO_DEBUG: >> >> Listening on >> udp: localhost [127.0.0.1]:5070 >> 0(18127) BUG: <core>: get_sock_info_list: invalid proto 4 >> Aliases: >> >> 0(18127) WARNING: <core>: tcp_options: tcp_connect_wait depends on >> tcp_buf_write, disabling... >> 0(18127) INFO: <core>: init_tcp: using epoll_lt as the io watch >> method (auto detected) >> 0(18127) BUG: <core>: core limits increased only to 4194304 >> 0(18127) WARNING: xlog: xl_mod_init: DIFFERENT hostname 'ds2' and >> gethostbyname 'ds2.sip-server.net' >> 0(18127) INFO: <core>: udp_init: SO_RCVBUF is initially 110592 >> 0(18127) INFO: <core>: udp_init: SO_RCVBUF is finally 262142 >> 4(18131) INFO: <core>: signal 2 received >> 2(18129) INFO: <core>: signal 2 received >> 3(18130) INFO: <core>: signal 2 received >> 3(18130) INFO: <core>: signal 15 received >> 0(18127) NOTICE: <core>: Thank you for flying ser >> >> ============== >> SER compiled w/o -DNO_DEBUG: >> >> Listening on >> udp: localhost [127.0.0.1]:5070 >> 0(21524) BUG: <core> [socket_info.c:201]: get_sock_info_list: invalid >> proto 4 >> Aliases: >> >> 0(21524) WARNING: <core> [tcp_options.c:102]: tcp_options: >> tcp_connect_wait depends on tcp_buf_write, disabling... >> 0(21524) INFO: <core> [tcp_main.c:3744]: init_tcp: using epoll_lt as >> the io watch method (auto detected) >> 0(21524) BUG: <core> [daemonize.c:344]: core limits increased only to >> 4194304 >> 0(21524) WARNING: xlog [xl_lib.c:1691]: xl_mod_init: DIFFERENT >> hostname 'ds2' and gethostbyname 'ds2.sip-server.net' >> 0(21524) INFO: <core> [udp_server.c:166]: udp_init: SO_RCVBUF is >> initially 110592 >> 0(21524) INFO: <core> [udp_server.c:217]: udp_init: SO_RCVBUF is >> finally 262142 >> 2(21526) BUG: <script>: Slow timer executed >> 3(21527) BUG: <script>: Fast timer executed >> 2(21526) BUG: <script>: Slow timer executed >> 3(21527) INFO: <core> [main.c:729]: signal 2 received >> 3(21527) INFO: <core> [main.c:729]: signal 15 received >> 1(21525) INFO: <core> [main.c:729]: signal 15 received >> 5(21529) INFO: <core> [main.c:729]: signal 15 received >> 0(21524) NOTICE: <core> [main.c:640]: Thank you for flying ser >> >> ============== >> >> Ondrej Martinek wrote: >>> Hi guys, >>> >>> I've already made some changes to the logging in SER. The log >>> messages will automatically include the severity, the position >>> (filename and line) and the module name (or "<core>"/"<script>" >>> string). Currently, it's being reviewed by Jan Janak. >>> >>> Moreover, I wrote a script that replaces LOG(static_severity, ...) by >>> corresponding dedicated macros and removes the module name and the >>> severity string from the log messages in SER sources. >>> >>> Regards >>> --ondra >>> >>> Bogdan Pintea wrote: >>>> I was actually wondering: shouldn't all logging be, finally, leveled? >>>> >>>> Right now, there are 3 logging types: >>>> - using dedicated macros for severity (ERR, WARN etc); >>>> - LOG(<severity>, ...), manually beautified -- severity and maybe >>>> function name present (the case of current commit); >>>> - as previous, but either no severity or no location info, or both >>>> missing; >>>> >>>> This is "hard" to deal with, both for a human and an automated log >>>> parser (besides the unprofessional, chaotic looks :-) ) . >>>> >>>> Are there clear disadvantages against sticking to only logging with >>>> dedicated severity macros and (progressively) migrate all log >>>> statements to that model? >>>> >>>> >>>> Bogdan. >>>> >>>> PS. I know not all compilers in all their versions have location >>>> helpers and that manual beautifying helps with those, but this >>>> support should nowadays be marginally (if at all) needed; besides, >>>> current status is mixed enough to be uncomfortable in those >>>> situations, as well. >>>> >>>> Nils Ohlmeier wrote: >>>>> calrissian 2008/08/24 00:15:43 CEST >>>>> >>>>> SER CVS Repository >>>>> >>>>> Modified files: >>>>> . forward.c Log: >>>>> fixed wrong function name in error message >>>>> Revision Changes Path >>>>> 1.118 +2 -2 sip_router/forward.c >>>>> http://cvs.berlios.de/cgi-bin/viewcvs.cgi/ser/sip_router/forward.c.diff?r1=1.117&r2=1.118 >>>>> >>>>> >>>>> _______________________________________________ >>>>> Serdev mailing list >>>>> [email protected] >>>>> http://lists.iptel.org/mailman/listinfo/serdev >>>>> >>>>> >>>> >>> _______________________________________________ >>> Serdev mailing list >>> [email protected] >>> http://lists.iptel.org/mailman/listinfo/serdev > _______________________________________________ Serdev mailing list [email protected] http://lists.iptel.org/mailman/listinfo/serdev
