Re: [PATCH v2 19/24] Introduce complete_command

[Pedro Alves <palves at redhat dot com>]

On 6/3/19 8:28 PM, Tom Tromey wrote:
>>>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:
> 
> Pedro> This adds a completion helper routine that makes it possible for a
> Pedro> command that takes another command as argument, such as "frame apply
> Pedro> all COMMAND" as "thread apply all COMMAND", to complete on COMMAND,
> Pedro> and have the completion machinery recurse and complete COMMAND as if
> Pedro> you tried to complete "(gdb) COMMAND".  I.e., we'll be able to
> Pedro> complete like this, for example:
> 
> Pedro> 	* completer.c (complete_command): New.
> 
> I think that the function should have a different name.  Traditionally,
> *_command is the name of the command "*" in gdb, and there's already a
> function called complete_command.  Overloading makes this not matter too
> much, but I also think it's best to reserve overloading for functions
> that are semantically similar.

Hmmm.  Somehow I forgot the "complete" command.  Bizarre, since the
approach here is actually stolen from the same trick that I had
applied to make the "complete" command complete command arguments
correctly a couple years ago.  Eh.

The name I had chosen as following the complete_xxx pattern in the
completer.h functions -- complete_expression, complete_symbol,
complete_source_filenames, complete_files_symbols, etc.

Not sure what to call this.  It ultimately call complete_line, and it's
used to complete nested commands so, maybe

  complete_nested_command_line 

?

Thanks,
Pedro Alves