"Nicolas Richard" <theonewiththeevill...@yahoo.fr> writes: > [fu2 gmane.emacs.bbdb.user] > > Eric Abrahamsen <e...@ericabrahamsen.net> writes: >> Incidentally, I'm using `ido-everywhere', and would like to write this >> function so that it didn't _rely_ on ido, but made use of ido when >> `ido-everywhere' was true. Ie, I'd like to replace the >> `ido-completing-read' calls with something more generic that still made >> use of ido when it was turned on. Any suggestions? > > ido-everywhere is only meant for buffer/file (see its docstring). People > who really want ido everywhere should probably use ido-ubiquitous which > sets completing-read-function to ido-completing-read (in fact, a wrapper > around it because it cannot handle all cases that completing-read should > handle). > > Hence, my suggestion would be to use completing-read instead of > ido-completing-read in your code, and configure ido-ubiquitous > (available from marmalade and melpa) for controlling what to use. > > Related package : ido-hacks (which also enables ido in more places) > > I *guess* it'll then also work automagically with icomplete-mode > (shipped with recent emacs) if that is what the user enable instead of > ido-ubiquitous.
Gosh, where did you guys dig this thread up from? :) Thanks for the refinements on the original function, and for the tips on ido related packages in elpa repos -- I hadn't been aware of that. I've been a tiny bit annoyed with this function ever since I wrote it (obviously not annoyed enough, since that was years ago), as progressive narrowing didn't work the way I wanted it to. Kevin, I'll give your version a whack, and maybe come back with some more refinements in a bit. Thanks! E
