Add HelpFlag as part of VisibleFlags.

[harshavardhana]

Additionally also allow help command to be hidden

[harshavardhana]

Any feedback on this @jszwedko ?

[jszwedko]

Hi @harshavardhana,

Apologies for the delay in review. Do you mind adding some additional details about the what this change is enabling?

It also appears to alter the existing interface by changing the behavior of the HideHelp field. If that was intentional, we may need to retarget this to the v2 branch to avoid breaking compatibility.

[wking]

Part of what this PR is doing is splitting HideHelp into separate booleans for the command and option. That would let us address #222 and avoid things like listing:

USAGE:
    [global options] command [command options] [arguments...]

for an app whose only command is help where the meat of the app is in the app-level Run. With this PR, setting HideHelpCommand but not HideHelp will give you:

USAGE:
    [global options] [arguments...]

I agree that there are backwards-compat concerns, but think we can address them with a slight tweak. Instead of re-purposing HideHelp, I'd deprecate it (to be removed in v2) and add both HideHelpCommand and HideHelpFlag. Then you could hide if the specific property was set or the deprecated generic property was set. That covers all cases except:

1. Alice's new code sets HideHelpFlag true, and hands the App off to Bob's code.
2. Bob's old code sets HideHelp false, because it wants to ensure that the --help flag is visible.

Bob's usecase would work now (because the only way Alice could have disabled --help is via HideHelp), but Bob's usecase would break with my proposed approach. Still, I think that passing App and Command instances around with different parties having different opinions on whether --help should be displayed is unlikely. In most cases, these are only used locally within a single repository.

[jszwedko]

@wking aha, I see, thank you for providing additional context. I like your idea of splitting out the flags and deprecating HideHelp.

[devmattrick]

Sorry for the ping, but are still plans to implement this?

[AkihiroSuda]

Any chance to get this reopened?

[coilysiren]

@AkihiroSuda @devmattrick either of you are free to open a new PR! Please create an associated issue first though.

[AkihiroSuda]

Will try, #523 seems the existing issue ticket that covers this.

[AkihiroSuda]

PR: #1083