Note: this code was written as a school homework. It is not going to be maintained and is not meant to be used in production. The only reason for publishing this is that its one of requirements. (the commit history is missing since I use an implicit versioning system on projects I don't intend to publish)
Cron daemon is used for task scheduling. The daemon wakes up every minute, goes through all records and for every one decides whether it is to be ran. If so, the supplemented command is executed in a separate shell and its output then outputted into the daemon's output.
Command is executed as a separate process in /bin/sh. All processes get executed paralelly, thus order is not defined. When process finishes, it is checked and its output written to standard output. This is done every minute.
Tasks are accepted in the following format: , where the first 5 fields are comma separated lists of numbers and ranges. Note that this format uses spaces as field separators, thus spaces are forbidden inside of a field (e.g. '1, 2' as is invalid).
Standard shell-like variables can be used in the file. Line in format of = can be used to set the variable varname to the value 'value'. Later the variable may be used as $varname. Double dollar is resolved to a single dollar. Variables are not evaluated recursively.
There are four variables set by default: LOGNAME, HOME, SHELL and PATH. LOGNAME and HOME are set from env, SHELL is set to /bin/sh and path is set to /bin:/usr/bin by default. All of these values may be overriden.
To compile this project, use the make command. The output file is mycrond.
Some tests are provided with expected output in folders test1, test2, test3.
The binary was checked for memory leaks by valgrind.
following options are available: -f (for now obligatory) -d daemonize -s silent -o once (for memtest purposes)
cronfile: path=/media/data 30 8-20 * * 2 backup $path
command: mycrond -f cronfile
This executes 'backup /media/data' 12 times on every wensday (at 8:30,9:30 ...)