While looking into #258 (use of argparse instead of optparse*), I noticed that:
some of the Cmd subclasses had separate fields for usage and descr_text, but others didn't. So I refactored all Cmd subclasses to have them.
having the subcommand options in an iterable form would be very convenient for use with argparse's subcommand support. Hence, I pulled out the options to a new options field in these classes.
This left much of the get_parser() and description() methods repetitive. By moving ignore_unknown_options to a field as well, I could eliminate all definitions of get_parser and description in favour of methods in the parent Cmd class.
By renaming the cluster_cmds() and node_cmds() functions, I could also do some DRY refactoring in the main ccm script, so there's that as well.
Finally, I added some empty lines here and there to stop flake8 complaining about PEP8 guidelines (2 blank lines after class/function definitions).
While looking into #258 (use of
argparseinstead ofoptparse*), I noticed that:Cmdsubclasses had separate fields forusageanddescr_text, but others didn't. So I refactored allCmdsubclasses to have them.argparse's subcommand support. Hence, I pulled out the options to a newoptionsfield in these classes.This left much of the
get_parser()anddescription()methods repetitive. By movingignore_unknown_optionsto a field as well, I could eliminate all definitions ofget_parseranddescriptionin favour of methods in the parentCmdclass.By renaming the
cluster_cmds()andnode_cmds()functions, I could also do some DRY refactoring in the mainccmscript, so there's that as well.Finally, I added some empty lines here and there to stop
flake8complaining about PEP8 guidelines (2 blank lines after class/function definitions).Thoughts?
* You can see the WIP state of that here.