I suppose you wanted a reply-all ... Mark Triggs <[EMAIL PROTECTED]> writes:
> Matthieu Moy <[EMAIL PROTECTED]> writes: > >> Mark Triggs <[EMAIL PROTECTED]> writes: >> >>> Any thoughts on the best way to fix this? I suppose >>> tla--get-buffer-create could be modified to accept one or more paths, >> >> It already accepts nil. I think this is what you need. Note that with >> nil as an argument, you can only have one *tla-bookmarks-missing* >> buffer at a time. Perhaps we could provide a way to have several, but >> I'm not sure how usefull this would be (maybe more confusing than >> usefull). > > Yeah, I thought about passing nil but because it'd only allow for a > single *tla-bookmarks-missing* buffer, I thought I'd check what others > thought. You're right that it could be confusing too, so I'll go with > this approach for now. The other option would be to pass a string which is not a directory path to tla--get-buffer-create, but this will most probably not work with the current implementation which has some hacks to unify "/path/to/file", "/path/to/file/", or even "/path//to/file" ... >> By the way, the tla-bookmarks-missing function could be modified not >> to use tla-bookmarks-missing-do-todolist. I had introduced this >> "todolist" to make the processes asynchronous with the old process >> sentinel scheme (I had no other way to tell the process sentinel what >> was still to be done) but this is now rather useless. > > Yeah. When I was working on the tla missing stuff, I originally > replaced everything but ran into problems and decided to start small and > just replace some of it. Now that that's working it might be easier to > remove the unnecessary parts, so I might take a shot at this when I get > a chance (but don't let me stop anyone else who would like to do it :o) > > Thanks! > > Mark -- Matthieu
