Andre Poenitz wrote: >> | Think plugin. >> >> inset-insert-pluginfoo still works... > > but LFUN_INSERT_PLUGIN_FOO will not as this is set at compile time. > >> | duplicate the code'. The code that's already there seems to >> | support the 'single lfun' approach, so this recommendation is >> | actually supporting your approach... >> >> And it is not hard to combine the handling of very similar lfuns >> further back in the processing. > > Yes, but it is not possible to create new lfuns at run tim right > now..
Guys, all this is very well and, indeed, it's nice to have a vision of the future. However, at the moment it is just a distraction. The current code works with lfun-inset-foo so use that and leave plugin lfuns discussion to a later day... Incidentally, Lars, the minibuffer tab completion stuff is currently broken for things like "dialog-open preferences" for exactly the reasons you describe. (Yes, my fault.) I think that we should fix that before 1.4.x is released, but let's leave the disussion about how best to proceed to a later date. (Ie, I have no time at the moment...) -- Angus
