Do something when a root command's flag is set about bashly

[DannyBen]

but I think it would be nice to have a way to do this by convention.

I thought about this, but decided against.
In Rails, here is a concept of before and after filter, which is similar to what you are describing - since this is a "before" filter.

I found this convention to be more trouble than its worth, and it "gets in your way" more often than not.

The straight forward solution would be to add a library function (bashly add lib), and call it from any command file that needs it.

That said, your use case of using --debug to set -x indeed shows the weak spot of the decision to avoid some sort of "before" filter.

Let me sleep on it.

[DannyBen]

I found the issue where this was considered originally: #116

I wonder if adding optional before.sh and after.sh (similar to how initialize.sh works) should be reconsidered.
Meaning - if such files are found, they will be executed in the appropriate place, then you can simply do

# src/before.sh
if [[ ${args[--debug]} ]]; then
  set -x
fi

[DannyBen]

Thanks for raising this issue Wes.

This feature is now released in 1.0.2

Docs:

• Writing Your Scripts > Before/after hooks
• Advanced Features > Before/After Hooks
• Examples > Before/After Hooks Example

[DannyBen]

Thank you.

1. Well, when you have subcommands, the root_command will never be called, this is why it is not there.
2. As you noticed, the initialize is not the place - it is done before anything else.

Defining flags on the root, just makes these flags available in all subcommands. Note however, that input order matters:

$ <command> <root flags> <subcommand> <subcommand flags and args>

With this bashly.yml

name: cli
help: Sample application
version: 0.1.0

flags:
- long: --debug
  short: -d
  help: Enable debug mode

commands:
- name: download
  alias: d
  help: Download a file

  flags:
  - long: --force
    short: -f
    help: Overwrite existing files

you get:

$ ./cli --debug download --force
args:
- ${args[--debug]} = 1
- ${args[--force]} = 1

Does this help?
Was this not intuitive/clear enough in the docs? Let me know where I can improve the behavior or documentation.

[westhomas]

Yes, that's clear from the docs, which are very good, btw.

I guess what I'm asking for is some way to run code after the arguments are initialized but before the command/sub-command is invoked.

For example, if I have the --debug flag on the root command, could there be some way to define a function that gets called?

      # :flag.case
      --debug)

        # :flag.case_no_arg
        args['--debug']=1
        shift
        # <========= need a function call right here
        ;;

My only alternative, is to check for ${args[--debug]} in every single sub-command's *_command.sh file. It works, but I think it would be nice to have a way to do this by convention.

[DannyBen]

This is now implemented in the edge version.

You can run bashly add hooks to add the two before/after hook files, and edit them as you see fit.

[DannyBen]

@westhomas were you able to test it?

[westhomas]

Just tested it. It's working very nicely! Thank you!