100 lines
4.8 KiB
Plaintext
100 lines
4.8 KiB
Plaintext
#Fcron documentation Fcron: how and why? Release notes: fcron 3.2.0
|
|
Thanks
|
|
|
|
Copyright © 2000-2014 Thibault Godouet Fcron 3.2.0 Web page :
|
|
http://fcron.free.fr
|
|
_______________________________________________________________________
|
|
|
|
Fcron documentation
|
|
Prev Chapter 1. Fcron: how and why? Next
|
|
_______________________________________________________________________
|
|
|
|
1.5. Todo
|
|
|
|
Here are some things that should be done ... Note that some of these
|
|
entries are just some ideas, which may be done in the long run.
|
|
|
|
If you see something you'd really like to see implemented here, feel
|
|
free to send me a mail: it may make me move it to the top of my to-do
|
|
list ;) .
|
|
|
|
If you plan to help me, please send a mail at <fcron@free.fr> to
|
|
prevent two people from doing the same thing. You can send me some
|
|
propositions as well, or ask me to move something to the top of the
|
|
heap ;) .
|
|
|
|
1.5.1. High priority
|
|
|
|
* Option to compile and install from git sources without generating
|
|
the doc
|
|
* register in OS suspend/hibernate mechanism to stop fcron when going
|
|
to sleep and start it again when resuming from sleep (see FAQ
|
|
entry).
|
|
*
|
|
|
|
1.5.2. Low priority
|
|
|
|
* Add unit tests using some unit-test framework (turn the tests/*
|
|
code into unit tests)
|
|
* For environment settings, make a var substitution.
|
|
* add a mailsubject option, for custom mail subjects (for instance,
|
|
in case of a job containing something secret -- password, etc -- in
|
|
the command line).
|
|
* Use directory notifications (FAM) / inotify, and support a fcrontab
|
|
and a cron.d as Vixie cron does (directly included in fcron itself,
|
|
not thanks to a script as now). However the parsing work is done by
|
|
fcrontab, and should probably keep on being done by fcrontab for
|
|
security and stability reasons: have fcron call fcrontab to do that
|
|
job?
|
|
* could be worth checking fcron for memory leaks using specialized
|
|
library (just in case...)
|
|
* option to put a maximum limit on the execution time of a task +
|
|
terminate it if not finished yet + send email to let the user know
|
|
* setting to limit the number of jobs of a single user in the
|
|
serialq/lavgq to X jobs + make sure root always has Y slots that it
|
|
can use in those queues (i.e. number of slots used by root + number
|
|
of free slots >= Y)
|
|
* Test (and use ?) docbook2x-man -- xlstproc ? cf
|
|
http://antoine.ginies.free.fr/docbook/ch09.html
|
|
* find way to have the non translated pages of the French translation
|
|
be updated automatically (changes, todo, etc -> copied from the
|
|
English doc)
|
|
* add a return code for jobs which would mean that they should not be
|
|
run again (in case of an error, etc...).
|
|
* PAM support (included in fcron, but needs more tests by people
|
|
using PAM - not implemented in fcrondyn: is it needed in fcrondyn
|
|
anyway?)
|
|
* support for per user serial queue (especially for root)
|
|
|
|
1.5.3. Ideas
|
|
|
|
* add a system of modules: at startup, fcron loads some modules (.so
|
|
files). Then, when fcron should run the job, it calls a function of
|
|
the module, and let it manage the job (threads?). (do linux
|
|
dlopen(), etc exist on other systems? - yes: thanks Harring ;)
|
|
dlopen exists on all POSIX system - even windoze - abait it does
|
|
not use the same insterface of function calls. But it can be easily
|
|
implemented to port to another system.).
|
|
* (related to the system of modules? create a kind of dependencies?)
|
|
Add a way to run a job only if a command returns true. It would
|
|
permit, for instance, to run some jobs on a laptop only if it is on
|
|
AC power.
|
|
and depending on the return value of the command, cancel the
|
|
execution, or wait x seconds before trying again, etc. In this
|
|
case, needs change the way the jobs are queued?
|
|
* add an option/module to prevent fcron from running a job if the
|
|
system is running on battery (anacron does it?)
|
|
* Add a way to stop a job (SIGSTOP) if the lavg gets to high, and
|
|
restart it (SIGCONT) when the lavg has decreased.
|
|
* Add a timeout for task completion: if a task takes more than a
|
|
specified timeout, then start another script with the initial task
|
|
PID as argument. That would allow a kill with different signals
|
|
(-HUP, -TERM -STOP, .... or any other action like sending a mail to
|
|
sysadmin to alert him/her that a task is taking too much time and
|
|
that he has to have a look at it.) (add an option to specify the
|
|
cmd, and combine w/ option until)
|
|
_______________________________________________________________________
|
|
|
|
Prev Home Next
|
|
Release notes: fcron 3.2.0 Up Thanks
|