Hello Hans,
I figured out where spurious loading of lm-ec and lm-texnansi in XeTeX
comes from.
One is comming from symb-glm.tex. But, when looking at it, it seems to
me that it's not even needed any more, since \leftguillemot,
\rightguillemot, \guilsingleleft and \guilsingleright are defined in
encodings.
Esp if I'm reading the comments: "but was also one of the reasons for
triggering the lm project"
The only possible side effect could be the ability to use
\symbol[xleftguillemot]
or even more obscure usage: calling font "Guil" directly.
But I doubt that anyone uses it that way, esp. since it depends on LM,
and people could use any other font (it looks to me it was used the
other way round, when no guilemots existed in cmr).
Is it possible to supress loading of that file, at least for XeTeX
where it's of no use, or at least rewriting it in some sensible way?
lm-texnansi comes from meta-tex.tex. Any chance to change that piece
of code, so that that font won't be needed any more (I have always
wondered what that definition in meta-tex is used for)?
\truefontname inside NEWTEX in font-ini seems to deliver names
suitable for luaTeX, but not for XeTeX.
Mojca