On Wed, 2 May 2001, Andre Poenitz wrote:
> > so this whole embedding stuff is to be delayed, it will be pretty hard
> > to maintain such a thing without having it in CVS.
> >
> > Lars, will you make a concession to accept the embedding code into a
> > branch? It might be better anyhow to work this out in a branch rather
> > than on the trunk.
>
> Once started it should not affect "trunk code" much. So it should not be
> too hard to do it on the main trunk once everything else is done *cough*,
> 1.2 is out, and so on.
Doing various things like "character-delete" will probably touch some
trunk code. but what do I know, I havent played with the core so far.
> Starting a branch right now would probably mean making the language
> decision _now_ and that is something I oppose.
Not necessarily, we agreed(?) that the core now is the lyxfunc-alike
support methods, and the "Official embedded language" should be named
before release (and be complete for the release). Theoretically after
everything is in place, adding a language shouldn't be more than a day or
two of work. The hard part is finding how to do the configure script and
how to link everything to LyX, once this is solved the rest is simple.
> Preliminary necessary stuff like finding some naming scheme for the
> required low level LyXFunc stuff do not need a branch either, and I'd
> doubt that the corresponding implementation is worth the branching either.
>
> So my prefered road map would be to get current CVS stable (mathed is
> certainly not yet ripe for 1.2), get 1.2 out, take a breath, put on the
> asbestos suits, and than start with language specific work...
I'm no CVS guru, it's your call on what to do and what not to do on a
branch. Though my idea is that we work the implementation on a branch and
when everything in it is stable enough we merge it back. This is probably
what should have been done with table and mathed, but then Lars is the
overlord and what he says goes.
I have no problem delaying everything until 1.2.0, I should probably
return to InsetGraphics to complete it or at least make it presentable in
1.2.0
Baruch
