head 1.4; access; symbols; locks alexis:1.4; strict; comment @# @; 1.4 date 99.04.28.11.02.28; author alexis; state Exp; branches; next 1.3; 1.3 date 99.02.20.18.07.07; author alexis; state Exp; branches; next 1.2; 1.2 date 99.01.15.13.15.25; author alexis; state Exp; branches; next 1.1; 1.1 date 99.01.07.13.03.21; author alexis; state Exp; branches; next ; desc @TODO list for PPPLC @ 1.4 log @*** empty log message *** @ text @------------------------------------------------------------------------------- Title: TODO LIST FOR PPPLC ------------------------------------------------------------------------------- To be Checked ------------------------------------------------------------------------------- dhuy: the 'pending disconnects cancelled' and 'disconnect sheduled' message need to be separated into two messages, then 'dhuy -c' can make use of the second one. ppplcd: abstract the msg queue engine for using in other programs ppplcd: abstract all state information into one file? ------------------------------------------------------------------------------- For Version: 1.5 ------------------------------------------------------------------------------- ppplc: reorder smf_submit parameters (msgid first), then split svrmsginfo_data into ppplc_msginfo_data and sm_msginfo_data, then add the DEL message to this, and have sm.c's funcs always check the sm_msginfo_data table first (maybe second, so dflt behaviour can be overruled), then move the QUIT message in there too, and then define a preprocessor-function version of smf_quit() (which will call smf_submit() (with NULL for message information array, since smb_submit() will find the message in it's own internal table). There are callers to the current smf_setquit() which use the return code; they'll just have to stop - it's probably on the info retrieveing function, which is a bit daft since if the SM is running then state_quit is obviously not set. ppplc: DONEWS, AUTORECONN should be QUIT should all be one parameter messages. ppplc: is MSGID_DEL still implemented as a conversion to a _NOACTION? it certainly looks like it from running 'dhuy' lib: there are some dummy scripts under ppplc/lib, remove them ppplcd: sort out fgets() overflows - without using #define's ppplcd: nsswitcher must be 'settled' by the time pppd establishes connection ppplcd: ppp_interface_no can sometimes be unset - *always* check it ppplcd: kill/restart still fails to set state properly - it is perhaps impossible for ppplcd to work out which pppd it is in charge of - it could accept a command 'reattach ', or it could use it's own lock files to work out which ppp interface is up but not bound to a ppplcd process ppplcd: is second connect attempt causing concurrent pppd invocation still? example: netmask should be 255.255.255.255 ? ppplcd: connect action should be config-defined command config: pppd command should be config-defined with a *lot* of substitutions! ppplcd: at the time that the state_wanted is toggled on, the chat file should be generated. utils.c: my_lock() is being used to deliberately fail creating a lock as a way to tell that the server is running. Unfortunately it's failing because it can't write to /var/run, not because there is a lock file there already! The idea is right I think, but it means I need another lock function my_testlock(), or the order of templock creation and real lock reading needs to be jiggled a bit. ppplcd: is -nf option documented in man page? ------------------------------------------------------------------------------- For Version: 1.6 ------------------------------------------------------------------------------- INSTALL: update regarding ip-up interfacing ipc: add support for text messages ppplcd: allow more config values to be defaulted clients: add option to specify server (or use default) - 1st stop, lock file! man: add manual pages for all example scripts scripts: news and mail scripts should guess where a lot of stuff is, so that the Makefile.defs definitions are optional (there is another ~/dev/XXX that already does this) scripts: should take verbosity level from ppplcd environment variable, a la I_AM_A_DAEMON. This would also require debug levels to be bitmapped for scripts - as it is for C programs. ------------------------------------------------------------------------------- $Header: /diskb/home/alexis/dev/supported/ppplc/doc/RCS/TODO,v 1.3 1999/02/20 18:07:07 alexis Exp alexis $ ------------------------------------------------------------------------------- @ 1.3 log @*** empty log message *** @ text @d4 8 d14 18 a32 1 ppplcc: manual page says 'C..............ommands' d34 5 a38 2 ppplcd: kill/restart still fails to set state properly ppplcd: abstract the msg queue engine for using in other programs a39 9 ppplcd: abstract all state information into one file? ipc: add support for text messages ppplcd: allow more config values to be defaulted clients: add option to specify server (or use default) - 1st stop, lock file! man: add manual pages for all example scripts scripts: news and mail scripts should guess where a lot of stuff is, so that the Makefile.defs definitions are optional (there is another ~/dev/XXX that already does this) ppplcd: problem with disabled mail - use sd 135? d52 11 a62 4 ppplcd: if quality level is set to 0 from cfg file, but quality interval is 900, and idle check interval is 600, get 'warning low interval' - even though interval won't be used. ppplcd: chat script should be generated at state_want change time. a65 4 dhuy: the 'pending disconnects cancelled' and 'disconnect sheduled' message need to be separated into two messages, then 'dhuy -c' can make use of the second one. INSTALL: update regarding ip-up interfacing d67 1 a67 1 $Header: /diskb/home/alexis/dev/ppplc/doc/RCS/TODO,v 1.2 1999/01/15 13:15:25 alexis Exp alexis $ @ 1.2 log @assigned *next* release number @ text @d6 2 d43 1 a43 1 INSTALL: update regardind ip-up interfacing d45 1 a45 1 $Header: /diskb/home/alexis/dev/ppplc/doc/RCS/TODO,v 1.1 1999/01/07 13:03:21 alexis Exp alexis $ @ 1.1 log @Initial revision @ text @d4 1 a4 1 For Version: 1.4 a6 3 ip-up: write a new version that uses the enviroment vars to determin interface number news-suck-multiple: tidy up, check newsserver definition a17 1 man: didn't get MARKER_PATCHLEVEL substituted at install time d41 1 a41 2 ip-up: a second versions exists now. This should be merged and the INSTALL document updated accordingly. The same goes for the ip-down script. d43 1 a43 1 $Header$ @