Re: company 0.2.1 -- extensible inline text completion mechanism

[Nikolaj Schumacher]

hugo <address@hidden> wrote:

> At first sight, his version for completing elisp (perhaps his version
> uses only the current buffer and not the internal lists?) is much
> faster,

If I see correctly, both offer symbols from the global list.  But while
his also offers all words from the buffer, mine attempts to correctly
determine the local variables.  It seemed more than fast enough.

If you have the time, could you move point to a place where it seems
slow and do:

M-: (benchmark-run 100 (company-elisp 'candidates ""))

> OTOH, none of those seems to provide a function like completing-read
> that I could use in my libraries...

Company can't work in a blocking call.  Are you interested in the
result, or do you just want to "start completion with these candidates"?


regards,
Nikolaj Schumacher