On Tue, Mar 15, 2005 at 09:59:37PM +0100, Georg Baum wrote: ...
> After rereading the old posts I think I know now what the problem is:
Good research. I missed all that.
> Creating a new mbox via the GUI creates a MathMBoxInset. The MathMBoxInset
> seems to be broken and asserts. André sent a fix for this one but forgot
> to attach the patch, and since then he seems to be silent.
> But this is not the worst problem: When reading from a file, an mbox
> creates a MathBoxInset("mbox").
> We need to settle either on MathMBoxInset or MathBoxInset("mbox") for all
> cases. The former is not possible without a lot of other changes. I tried
> to do that and created a halfworking experimental patch (mbox-test.diff),
> but I don't like the result. It will not apply cleanly, don't try it
> unless you are really curious.
> The latter is easy and is basically the same that is done in 1.3. André
> proposed to choose this solution, and I agree. The disadvantage is that
> on-screen display is not as nice, but is has another big advantage: A
> MathMBoxInset is stored in the .lyx file in textinset format, a
> MathBoxInset("mbox") is stored in .tex format.
I assume the on-screen appearance can be fixed in LaTeXFeatures. See how
it is done with mathcircumflex.
> We never had a lyx2lyx
> conversion for this.
>
> Conclusion: I propose to apply the attached patch 1823.diff. OK?
>
>
> Georg
Hmmm... sounds good.
BTW I don't much like the practice of leaving old stuff around, commented out.
It will be forgotten about, nobody will remember why it wasn't deleted
straight-out, and nobody will dare to delete it. And we already have an
official versioning system in use ;-)
- Martin
pgp3YN8PmmCvH.pgp
Description: PGP signature
