Subparsers of subparsers : missing parsing exception ?

[Altyrost]

While working with the example of subparsers in the documentation, with git:

some_exe submodule is considered valid by the parser, even if the subparser cannot have any other argument than the subparser update.

Did I miss something here ?

[skrobinson]

If I understand correctly, you expect an error if update is missing after submodule in the parsed string.

Unless a missing argument (or subparser) is required, the parser will not report an error. Since subparsers do not have a required setting, they can never be missing.

But you can add the same logic to the git example with is_subcommand_used.

  if (program.is_subcommand_used(submodule_command) &&
      !submodule_command.is_subcommand_used(submodule_update_command)) {
    std::cout << "submodule command called without update subcommand" << std::endl;
  }

[Altyrost]

Right. Shouldn't it be considered an error during the parse_arg ?

[skrobinson]

I don't believe it is always an error. From a UI view, some authors may prefer that some_exe submodule -h shows that the update subcommand exists. An example from git is the stash command which has list, show, drop, etc. subcommands.

But, others may want to say submodule without update is an error and they can express that condition with is_subcommand_used.