[lltx] [tex-live] Updating TeXLive 2010 using tlgmr

[Reinhard Kotucha]

On 2011-09-13 at 17:03:23 +0200, SKL wrote:

 > Checking some discussion boards I believe that this might be a bug
 > in XeTex - I am not using any of the LM fonts AFAIK as I have
 > substituted everything with Apple fonts, but nevertheless XeTeX
 > requires those fonts to be present for some reason (of course it
 > might also be that I am still using LM in some of the less obvious
 > places because I failed to specify a correct font, but I am not
 > sure how to test for this

No, it has nothing to do with XeTeX.  It's rather a LaTeX issue, or,
in particular, a fontspec issue. 

It happens with LuaTeX too:

---------------------------------
\documentclass{minimal}
\usepackage{fontspec}
\fontspec{CharisSIL}
\stop
---------------------------------

$ strace -eopen,stat lualatex -interaction=batchmode testfont.tex 2>&1 | grep lmroman
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf", {st_mode=S_IFREG|0644, st_size=111536, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf", {st_mode=S_IFREG|0644, st_size=111536, ...}) = 0
open("/home/reinhard/.texlive2011/texmf-var/luatex-cache/generic/fonts/otf/temp-lmroman10-regular.lua", O_RDONLY) = 7
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf", {st_mode=S_IFREG|0644, st_size=111536, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmromanslant10-regular.otf", {st_mode=S_IFREG|0644, st_size=118508, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmromanslant10-regular.otf", {st_mode=S_IFREG|0644, st_size=118508, ...}) = 0
open("/home/reinhard/.texlive2011/texmf-var/luatex-cache/generic/fonts/otf/temp-lmromanslant10-regular.lua", O_RDONLY) = 8
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmromanslant10-regular.otf", {st_mode=S_IFREG|0644, st_size=118508, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-italic.otf", {st_mode=S_IFREG|0644, st_size=118828, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-italic.otf", {st_mode=S_IFREG|0644, st_size=118828, ...}) = 0
open("/home/reinhard/.texlive2011/texmf-var/luatex-cache/generic/fonts/otf/temp-lmroman10-italic.lua", O_RDONLY) = 8
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-italic.otf", {st_mode=S_IFREG|0644, st_size=118828, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-bold.otf", {st_mode=S_IFREG|0644, st_size=111240, ...}) = 0
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-bold.otf", {st_mode=S_IFREG|0644, st_size=111240, ...}) = 0
open("/home/reinhard/.texlive2011/texmf-var/luatex-cache/generic/fonts/otf/temp-lmroman10-bold.lua", O_RDONLY) = 8
stat("/usr/local/texlive/2011/texmf-dist/fonts/opentype/public/lm/lmroman10-bold.otf", {st_mode=S_IFREG|0644, st_size=111240, ...}) = 0


I suppose that this happens when fontenc is loaded before the font is
selected.  The default font is Latin Modern.

The problem is not new, I encountered it with pdfTeX too:

---------------------------------
\documentclass{minimal}
\usepackage[T1]{fontenc}
\stop
---------------------------------

$ strace -eopen,stat pdflatex -interaction=batchmode testpdftex.tex 2>&1 | grep ecr
stat("/usr/local/texlive/2011/texmf-dist/fonts/tfm/jknappen/ec/ecrm1000.tfm", {st_mode=S_IFREG|0644, st_size=3584, ...}) = 0
open("/usr/local/texlive/2011/texmf-dist/fonts/tfm/jknappen/ec/ecrm1000.tfm", O_RDONLY) = 6

It doesn't happen with
---------------------------------
\documentclass{minimal}
\usepackage{mathpazo}
\stop
---------------------------------
, presumably because mathpazo correctly switches to Palatino before
switching to T1.  It would be nice if fontspec can do the same.

Regards,
  Reinhard

-- 
----------------------------------------------------------------------------
Reinhard Kotucha                                      Phone: +49-511-3373112
Marschnerstr. 25
D-30167 Hannover                              mailto:reinhard.kotucha at web.de
----------------------------------------------------------------------------
Microsoft isn't the answer. Microsoft is the question, and the answer is NO.
----------------------------------------------------------------------------