Andre Poenitz <[EMAIL PROTECTED]> writes:
>
> On Mon, Oct 25, 2004 at 12:23:17AM +0200, Andreas Vox wrote:
> > During my tests I found strange <+> ... </+> tokens in the produced
> > MathML code.
> > I tracked this down to math_extern.C, where the normal code is replaced
> > by MathExFuncInsets.
>
> That's funny, this part of math_extern.C should not be involved at
> all. In theory. What peice of code calls these transformations?
Well, I'm not the author ;-), but I would guess in math_mathmlstream.C:
MathMLStream & operator<<(MathMLStream & ms, MathArray const & ar)
{
mathmlize(ar, ms);
return ms;
}
I greped for all "MTag(...)" calls in src/mathed and ExFuncInset was the only
candidate to produce <+>. Then I greped for "new ExFuncInset" and settled
for math_extern.C.
I feel like Sherlock Holmes now ;-)
>
> > The attached patch produces correct MathML for these four operators
> > again, but I don't know if it breaks the intended
> > external-functionality.
>
> I think all that's missing is the implementation of the function
> 'void mathmlize(MathMLStream &) const' for MathCharInset.
That might as well help for bug #1729.
For bug #1728 8the <+> symptom) I'm not so sure ...
You probably guessed that well-formed if not valid Docbook XML is
on my list for 1.4.0, no? :-)
Cheers
/Andreas
