On 5/25/18 1:42 PM, Luca Boccassi wrote:
>> There is already logic that determines whether the shell is trying to
>> complete a command word (in_command_position). The existing code does
>> not attempt programmable completion if in_command_position == 1. The
>> additional functionality would:
>>
>> 1. Add an option to the complete builtin to specify how to complete
>> command names, and store it in a specially-named compspec, like
>> completion for empty lines does.
>>
>> 2. Add code to invoke that completion, if it exists and programmable
>> completion is active, before attempting bash's default completion,
>> if in_command_position == 1.
>>
>> Chet
>
> Hi,
>
> Thanks for the guidance! Unfortunately I'm still not quite there yet -
> inlined are the changes based on my understanding of the above. It
> implements a "usercmd/-U" option similar to the existing -E for empty
> line.
>
> But at the moment all it does is to allow (via "complete -U -F foo") to
> complete when nothing is typed in, which seemed to be possible already
> with -E. If some characters are already typed in, it will still do the
> default completion to commands in the PATH.
>
> What have I missed?
You should make sure you don't add your code in the section with the rest
of the programmable completions, since that block is not entered if
in_command_position != 0.
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU c...@case.edu http://tiswww.cwru.edu/~chet/
