On Monday 30 September 2002 6:47 pm, Andre Poenitz wrote: > On Mon, Sep 30, 2002 at 06:33:03PM +0100, Angus Leeming wrote: > > As for labels within mathed --- well personally, I think > > that André should move his CommandInset into the outside > > world and we should use that for the labels in 1.3. This > > code has been working for months although few people have > > used it. I do not see a cleaner/smaller solution to the > > problem otherwise. > > Hm... this is really just the tip of the iceberg "Inset > Unification". I'd say leave it as it is for 1.3 as I don't > think we can "just move it". The whole frontend stuff > currently depends on insets being derived from > inset/InsetCommand and not mathed/commandinset. As long as you > store backreferences there, there is no easy way to take care > of both hierarchies without duplicating some code. Even if > this is only _feature_ freeze and I believe we could > completely switch over without too much of a problem, this is > not "just a few minutes".
I'm suggesting "Inset Unification" for the Label Inset only. Reasons for: * if the mathed label inset derived from CommandInset then it would look like labels outside which is a good thing in and of itself. * If CommandInset were moved outside of mathed and insetlabel derived from it then code duplication would be minimal. We'd have to duplicate the ControlCommand class in the controllers directory. That's it. * Doing this would provide a really good test case for your unification idea with minimal impact on the rest of LyX. Reasons against: * The current code works and Qt users can change their mathed labels from the command line. What do the rest of you think? Lars? > [OTOH if "Christmas" is still the target time for 1.3pre I > think it should be possible to "just do it" and stabilize > within a _few_ weeks] Angus
