Fixed /jira subcommand parsing consistency

[levb]

Cleaned up /jira sub-command parsing

Ticket: https://mattermost.atlassian.net/browse/MM-15335

[crspeller]

@levb Considered using https://github.com/spf13/cobra ? That's what MM server uses.

[levb]

@crspeller sure thing. IMO we should consider extracting some of this functionality into p.Helpers, and can then choose whatever implementation we want, right? Separate ticket?

[crspeller]

@levb Certainly a separate ticket for moving stuff out to p.Helpers. But you might consider doing the cobra migration for this ticket. Your call.

[cpoile]

@crspeller isn't Cobra for CLI, and this is for slash command parsing?

[crspeller]

@cpoile Aren't slash commands like a CLI?

It have used cobra for this before. An example is matterbuild here: https://github.com/mattermost/matterbuild/blob/master/server/server.go

[cpoile]

@crspeller that's awesome, didn't realize.. :)

[cpoile]

Looks great to me, one minor non-blocking nit-pick.

[cpoile]

nit: more idiomatic way:

if h, ok := ch.handlers[strings.Join(args[:n], "/")]; ok {
   return h(p, c, args[n:]...)
}

But 0/5 personally (because there's no chance you'll have a nil value instead of a missing entry).

[levb]

I disagree. In this case there is no semantic difference between "no entry" and a nil entry, so checking the value is the right way.

[cpoile]

Sounds good.

[hanzei]

I didn't know a SetArgs method exists in Cobra. I was looking for a smarter way to parse slash commands for a long time. Locking forward to seeing this as a helper.

[levb]

Will consider COBRA later, as a separate PR