rgheck <[EMAIL PROTECTED]> writes: > Here's why I'm unhappy with the patch. It's simple in a way: We have a > vector of commands to execute rather than just one---that was the > first problem. So now you could do something like: > lyx -x "file-open /tmp/branchTest.lyx" -x "buffer-end" -x > "break-paragraph" -x "self-insert This is so cool!" > with the expected results.
This is indeed cool. > BUT this will not work when we are exporting. The reason is that when > you use the export feature, LyX calls Bufer::dispatch() rather than > (say) lyx::dispatch(), and the only thing that previously got handled > in Buffer::dispatch() is LFUN_BUFFER_EXPORT. So to get branch > activation and de-activation to work when we're exporting, I had to > put those in Buffer::dispatch() It is a very good idea to put all the things that are buffer-specific in Buffer::dispatch(). We should arrange to have this method called from LyXFunc::dispatch() (look at the default branch of the big switch). I think it is even possible to make LyXFunc::dispatch work properly when there is no view, but this is probably more work. > ---and then call them from Text::dispatch() so they'll work > normally. Huh? In the current situation, the best would be BufferView::dispatch, actually. This is relevant for all actions that are global to the current document being edited. Text::dispatch() is for actions that are relevant when the cursor is in a paragraph. > This does not look pretty---though maybe there are so few things > you'd actually want to do when exporting that it's OK. I don't know. > But maybe you might want to alter some of the other buffer params > (say, line spacing) before you export. Yes. JMarc