Richard Heck wrote: >>> The attached patch is a start on this. It does work, in the sense that >>> you can have both completion-accept and cell-forward bound to Tab, and >>> this works in tables. >>> >> >> "completion-accept" is added from me and works in a different way then >> "complete" >> which was bound to tab before. i'm not sure, but most people probably want >> "complete" one. >> >> > Yes, complete suggests a completion, whereas completion-accept accepts it. > I think you probably do NOT want Tab bound to complete. I did that first, > and then Tab never works if you are at the end of a word in a table > (because LFUN_COMPLETE will be enabled). Of course, we can bind > LFUN_COMPLETE to something else....
wouldn't binding tab to the table movement _on_the_first_place_in_lfuns_list solve this? completion-accept was pushed mainly because i use completion in a bit different fashion and don't want to force anybody on it. >> what is the advantage of having "\addbind" instead of using old "bind"? >> >> > See the answer to JMarc. yes that makes sense. >>> The problem is that it is entirely unclear how to integrate this with the >>> shortcut editing UI. Ideas are welcome. >>> >> >> maybe to allow multiple bindings in shortcut pane and only show warning >> about >> already used key? >> >> > yes, then we could do that. But note that there'd be no way to manage the > order of the bindings then, and no indication of what they are. So I'm > inclined not to allow "adding" a binding via the shortcut panel unless > those problems are solved. And maybe we could also disallow rebinding a key > that has multiple bindings. I don't know. This seems to call for listing > the bindings by key, rather than by LFUN. i dont have strong opinion about that. but as is shown above order maybe important. > By the way---this isn't necessarily for you, Pavel---I've noticed some > oddities in the shortcut dialog while working on this. First, it seems > bindings from site.bind don't show up in the dialog. They should, at least > so you know what they are. Attempts to overwrite these bindings seem to > fail, as well, though they "look" as if they work. For example, if you try > to bind Shift-Tab to accent-umlaut, it shows up in the panel, but it > doesn't work. Second, but presumably easier to solve: Hitting "Remove" > doesn't active "Apply". Bo maybe has the idea what going on? pavel
