Closed jesuspc closed 7 years ago
Yes, it probably can be added.
We should take a look at the table structure of those docsets and see where to fetch that data.
I have a working solution for this (adding "%f" for filename), but I'm afraid there can be a performance deterioration. I'll push it in a branch and we should try it for some time.
I'm probably hairsplitting, but want to be sure that for normal cases there is no noticeable delay (ill-formed strings would probably show the delay, but f*ckit)
there you go. test it in that branch for now. https://github.com/areina/helm-dash/pull/161
That's great! Thanks!
It really makes a difference.
I'm merging it for now. If it becomes a performance issue, we can optimize it by evaluating %f lazily.
Currently there are three format tokens available:
%d - docset name %n - name of the token %t - type of the token
Even when using all of those there are still circumstances in which a search pattern outputs multiple results that can not be distinguished. For instance:
That's the same search pattern result in the Dash UI:
Note that in the second case there is additional information available (the module where the function is defined).
Would it be possible to provide additional candidate-format tokens so the additional information that is shown in the Dash.app also appears in the helm buffer?