fontspec: Unknown feature `' in font `TeX Gyre Heros'.

Martin,

;;
causes an empty feature request.

Cheers,

Ross

Sent from my iPad

Post by Martin SchrÃ¶der
Hi,
-----------------
\setmainfont[Ligatures=TeX]{TeX Gyre Heros}
\setsansfont[Ligatures=TeX]{TeX Gyre Heros}
-----------------
I get the warning
-----------------
Requested font "TeX Gyre Heros/ICU:script=latn;language=DFLT;mapping=tex-text;;
" at 10.0pt
-> C:/Program Files (x86)/MiKTeX 2.9/fonts/opentype/public/tex-gyre/texgyreher
os-regular.otf
Unknown feature `' in font `TeX Gyre Heros'.
-----------------
Why? The ligs seem to work. :-)
System is MikTeX 2.9 (XeTeX is 0.9997.4 (MiKTeX 2.9), fontspec is
2010/11/17 v2.1e).
Best
Martin
--------------------------------------------------
http://tug.org/mailman/listinfo/xetex

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman

Martin Schröder

2010-12-22 19:36:20 UTC

Post by Ross Moore
;;
causes an empty feature request.

So it's an error in fontspec?

Best
Martin

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Ross Moore

2010-12-22 19:55:52 UTC

Hi Martin,

Post by Ross Moore
;;
causes an empty feature request.

So it's an error in fontspec?

I don't pass such judgements.

Should an empty element in an array always be ignored silently?
Should an extra code block search for, or catch, such a thing
before passing on to achieve the main purpose of the coding?
Presumably that might depend upon how much damage is being done.

Writing a message to warn the user of their sloppiness is quite
a reasonable action, in my opinion. This is done a lot in TeX
and LaTeX, so why not here?

Post by Martin SchrÃ¶der
Best
Martin

Cheers,

Ross

------------------------------------------------------------------------
Ross Moore ross.moore-EKjDg2YBFtyHXe+***@public.gmane.org
Mathematics Department office: E7A-419
Macquarie University tel: +61 (0)2 9850 8955
Sydney, Australia 2109 fax: +61 (0)2 9850 8114
------------------------------------------------------------------------

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Martin Schröder

2010-12-22 20:01:10 UTC

Post by Ross Moore
Writing a message to warn the user of their sloppiness is quite
a reasonable action, in my opinion. This is done a lot in TeX

What's wrong with my input?

Best
Martin

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Ross Moore

2010-12-22 20:30:17 UTC

Hi Martin,

Post by Ross Moore
Writing a message to warn the user of their sloppiness is quite
a reasonable action, in my opinion. This is done a lot in TeX

What's wrong with my input?

Sorry.

Post by Martin SchrÃ¶der
"TeX Gyre Heros/ICU:script=latn;language=DFLT;mapping=tex-text;;
" at 10.0pt

to see the ;; .

The possible <return> after it, before the closing " is a possible
worry too --- though that may be just the number of characters in
the output message.

I get no message, with:

This is XeTeX, Version 3.1415926-2.2-0.9997.4 (TeX Live 2010)

fontspec.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX
fontspec-patches.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX

Post by Ross Moore
System is MikTeX 2.9 (XeTeX is 0.9997.4 (MiKTeX 2.9),
fontspec is 2010/11/17 v2.1e).

So maybe a little buglet was introduced in the latest update,
and MikTeX just serves it up regardless.
Does it happen with other fonts too?

Post by Martin SchrÃ¶der
Best
Martin

Herbert Schulz

2010-12-22 21:04:33 UTC

Post by Ross Moore
Hi Martin,

Post by Ross Moore
Writing a message to warn the user of their sloppiness is quite
a reasonable action, in my opinion. This is done a lot in TeX

What's wrong with my input?

Sorry.

Post by Martin SchrÃ¶der
"TeX Gyre Heros/ICU:script=latn;language=DFLT;mapping=tex-text;;
" at 10.0pt

to see the ;; .
The possible <return> after it, before the closing " is a possible
worry too --- though that may be just the number of characters in
the output message.
This is XeTeX, Version 3.1415926-2.2-0.9997.4 (TeX Live 2010)
fontspec.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX
fontspec-patches.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX

Post by Ross Moore
System is MikTeX 2.9 (XeTeX is 0.9997.4 (MiKTeX 2.9),
fontspec is 2010/11/17 v2.1e).

So maybe a little buglet was introduced in the latest update,
and MikTeX just serves it up regardless.
Does it happen with other fonts too?

Post by Martin SchrÃ¶der
Best
Martin

Cheers,
Ross

Howdy,

No problem here with fontspec 2.1e in TeX Live 2010 with the file

%%!TEX TS-program = xelatex
%%!TEX encoding = UTF-8 Unicode
\listfiles
\documentclass{article}

\usepackage{fontspec}
\setmainfont[Ligatures=TeX]{TeX Gyre Heros}
\setsansfont[Ligatures=TeX]{TeX Gyre Heros}
\setmonofont[Scale=MatchLowercase]{Monaco}

\begin{document}
Some text and some \textsf{Sans Text}.
\end{document}

and the list of files is:

article.cls 2007/10/19 v1.4h Standard LaTeX document class
size10.clo 2007/10/19 v1.4h Standard LaTeX file (size option)
fontspec.sty 2010/11/17 v2.1e Advanced font selection for XeLaTeX/LuaLaTeX
expl3.sty 2010/11/13 v2083 L3 Experimental code bundle wrapper
l3names.sty 2010/10/03 v2064 L3 Experimental Naming Scheme for TeX Primitiv
es
etex.sty 1998/03/26 v2.0 eTeX basic definition package (PEB)
l3basics.sty 2010/10/03 v2063 L3 Experimental basic definitions
l3expan.sty 2010/09/20 v2037 L3 Experimental Argument Expansion module
l3tl.sty 2010/12/18 v2103 L3 Experimental Token Lists
l3int.sty 2010/10/17 v2077 L3 Experimental Integer module
l3quark.sty 2010/09/20 v2036 L3 Experimental Quark Commands
l3seq.sty 2010/03/29 v1879 L3 Experimental sequences and stacks
l3toks.sty 2010/09/20 v2037 L3 Experimental Token Registers
l3prg.sty 2010/10/03 v2063 L3 Experimental control structures
l3clist.sty 2010/10/09 v2071 L3 Experimental comma separated lists
l3token.sty 2010/10/03 v2063 L3 Experimental token investigation and manipu
lation
l3prop.sty 2010/11/23 v2087 L3 Experimental Property Lists
l3msg.sty 2010/10/02 v2052 L3 Experimental LaTeX Messages module
l3io.sty 2010/10/03 v2063 L3 Experimental i/o module
l3skip.sty 2010/11/23 v2088 L3 Experimental skip registers
l3box.sty 2010/09/26 v2048 L3 Experimental Box module
l3keyval.sty 2010/04/11 v1890 L3 Experimental keyval processing
l3keys.sty 2010/11/11 v2082 L3 Experimental key-value support
l3precom.sty 2010/02/09 v1793 L3 Experimental precompilation module
l3xref.sty 2010/02/09 v1786 L3 Experimental cross referencing
l3file.sty 2010/03/21 v1853 L3 Experimental file loading
l3fp.sty 2010/12/18 v2104 L3 Experimental floating-point operations
l3luatex.sty 2010/07/18 v1985 L3 Experimental LuaTeX functions
calc.sty 2007/08/22 v4.3 Infix arithmetic (KKT,FJ)
xparse.sty 2010/10/13 v2073 Generic document command parser
xkeyval.sty 2008/08/13 v2.6a package option processing (HA)
xkeyval.tex 2008/08/13 v2.6a key=value parser (HA)
fontspec-patches.sty 2010/11/17 v2.1e Advanced font selection for XeLaTeX/Lu
aLaTeX
fixltx2e.sty 2006/09/13 v1.1m fixes to LaTeX
fontenc.sty
eu1enc.def 2010/05/27 v0.1h Experimental Unicode font encodings
eu1lmr.fd 2009/10/30 v1.6 Font defs for Latin Modern
xunicode.sty 2010/11/06 v0.96 provides access to latin accents and many othe
r characters in Unicode lower plane
fontspec.cfg

Good Luck,

Herb Schulz
(herbs at wideopenwest dot com)

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Ross Moore

2010-12-22 21:48:14 UTC

Post by Ross Moore
This is XeTeX, Version 3.1415926-2.2-0.9997.4 (TeX Live 2010)
fontspec.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX
fontspec-patches.sty 2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX

Post by Martin SchrÃ¶der
System is MikTeX 2.9 (XeTeX is 0.9997.4 (MiKTeX 2.9),
fontspec is 2010/11/17 v2.1e).

So maybe a little buglet was introduced in the latest update,
and MikTeX just serves it up regardless.
Does it happen with other fonts too?

Howdy,
No problem here with fontspec 2.1e in TeX Live 2010 with the file

Yep the difference between 2.1d and 2.1e is not relevant to XeTeX.
viz.

[GlenMorangie:~/Downloads/wspr-fontspec-21a44ef/hide] rossmoor% diff -C1 ../fontspec.sty `kpsewhich fontspec.sty`
*** ../fontspec.sty 2010-12-23 07:54:23.000000000 +1100
--- /Users/rossmoor/Library/texmf/tex/latex/fontspec/fontspec.sty 2010-11-09 10:04:45.000000000 +1100
***************
*** 16,18 ****
\ProvidesPackage{fontspec}
! [2010/11/17 v2.1e Advanced font selection for XeLaTeX/LuaLaTeX]

--- 16,18 ----
\ProvidesPackage{fontspec}
! [2010/11/07 v2.1d Advanced font selection for XeLaTeX/LuaLaTeX]

***************
*** 28,31 ****
\luatex_if_engine:TF {
! \RequirePackage{luaotfload}
! \RequireLuaModule{fontspec}
}{
--- 28,31 ----
\luatex_if_engine:TF {
! \RequirePackage{luatextra}[2010/05/10]
! \luatexRequireModule{fontspec}
}{

Post by Herbert Schulz
Good Luck,
Herb Schulz
(herbs at wideopenwest dot com)

Herb is on MacOS X, as am I.
So there must be a difference due to platform, or to support
files used by XeTeX but not showing in the .log file.
That is potentially disturbing, though the effect on typesetting
seems to be imperceptible.

All the best,

Ross

------------------------------------------------------------------------
Ross Moore ross.moore-EKjDg2YBFtyHXe+***@public.gmane.org
Mathematics Department office: E7A-419
Macquarie University tel: +61 (0)2 9850 8955
Sydney, Australia 2109 fax: +61 (0)2 9850 8114
------------------------------------------------------------------------

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Martin Schröder

2010-12-23 09:39:35 UTC

Here's my minimal input:

\listfiles
\documentclass{article}
\usepackage[ngerman,english]{babel}
\usepackage{fontspec}
\setmainfont[Ligatures=TeX]{TeX Gyre Heros}
\setsansfont[Ligatures=TeX]{TeX Gyre Heros}
\begin{document}
A
\end{document}

*File List*
article.cls 2007/10/19 v1.4h Standard LaTeX document class
size10.clo 2007/10/19 v1.4h Standard LaTeX file (size option)
babel.sty 2008/07/06 v3.8l The Babel package
bblopts.cfg 2006/07/31 v1.0 MiKTeX 'babel' configuration
ngermanb.ldf 2008/07/06 v2.6n new German support from the babel system
english.ldf 2005/03/30 v3.3o English support from the babel system
fontspec.sty 2010/11/17 v2.1e Advanced font selection for XeLaTeX/LuaLaTeX
expl3.sty 2010/11/13 v2083 L3 Experimental code bundle wrapper
l3names.sty 2010/10/03 v2064 L3 Experimental Naming Scheme for TeX
Primitives
etex.sty 1998/03/26 v2.0 eTeX basic definition package (PEB)
l3basics.sty 2010/10/03 v2063 L3 Experimental basic definitions
l3expan.sty 2010/09/20 v2037 L3 Experimental Argument Expansion module
l3tl.sty 2010/11/17 v2084 L3 Experimental Token Lists
l3int.sty 2010/10/17 v2077 L3 Experimental Integer module
l3quark.sty 2010/09/20 v2036 L3 Experimental Quark Commands
l3seq.sty 2010/03/29 v1879 L3 Experimental sequences and stacks
l3toks.sty 2010/09/20 v2037 L3 Experimental Token Registers
l3prg.sty 2010/10/03 v2063 L3 Experimental control structures
l3clist.sty 2010/10/09 v2071 L3 Experimental comma separated lists
l3token.sty 2010/10/03 v2063 L3 Experimental token investigation
and manipulation
l3prop.sty 2010/11/23 v2087 L3 Experimental Property Lists
l3msg.sty 2010/10/02 v2052 L3 Experimental LaTeX Messages module
l3io.sty 2010/10/03 v2063 L3 Experimental i/o module
l3skip.sty 2010/11/23 v2088 L3 Experimental skip registers
l3box.sty 2010/09/26 v2048 L3 Experimental Box module
l3keyval.sty 2010/04/11 v1890 L3 Experimental keyval processing
l3keys.sty 2010/11/11 v2082 L3 Experimental key-value support
l3precom.sty 2010/02/09 v1793 L3 Experimental precompilation module
l3xref.sty 2010/02/09 v1786 L3 Experimental cross referencing
l3file.sty 2010/03/21 v1853 L3 Experimental file loading
l3fp.sty 2010/11/25 v2092 L3 Experimental floating-point operations
l3luatex.sty 2010/07/18 v1985 L3 Experimental LuaTeX functions
calc.sty 2007/08/22 v4.3 Infix arithmetic (KKT,FJ)
xparse.sty 2010/10/13 v2073 Generic document command parser
xkeyval.sty 2008/08/13 v2.6a package option processing (HA)
xkeyval.tex 2008/08/13 v2.6a key=value parser (HA)
fontspec-patches.sty 2010/11/17 v2.1e Advanced font selection for
XeLaTeX/LuaLaTeX
fixltx2e.sty 2006/09/13 v1.1m fixes to LaTeX
fontenc.sty
eu1enc.def 2010/05/27 v0.1h Experimental Unicode font encodings
eu1lmr.fd 2009/10/30 v1.6 Font defs for Latin Modern
xunicode.sty 2010/11/06 v0.96 provides access to latin accents and
many other characters in Unicode lower plane
fontspec.cfg

Best
Martin

--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
http://tug.org/mailman/listinfo/xetex

Ross Moore

2010-12-23 23:46:25 UTC

Hi Martin,

Post by Herbert Schulz
\listfiles
\documentclass{article}
\usepackage[ngerman,english]{babel}

On seeing this, my first inclination was that babel is
redundant with XeTeX, and maybe was changing a catcode,
or something like that.
But testing with and without showed no difference.

Then I remembered that some messages are written to the .log file,
but *not* to the console. Checking for this, I found that yes I *do*
see the messages that you reported.

viz.

\g_fontspec_family_TeX Gyre Heros_int=\count171
fontspec info: Defining font family 'TeXGyreHeros(0)' for font 'TeX Gyre Heros
' with options [Ligatures=TeX].
\g_fontspec_TeXGyreHeros(0)_prop=\toks39
(fontspec) <->"TeX Gyre Heros/ICU:script=latn;language=DFLT;mapping=tex-text;;
"
(fontspec) <->"TeX Gyre Heros/ICU:script=latn;language=DFLT;mapping=tex-text;;
+smcp;"
(fontspec) <->"TeX Gyre Heros/B/ICU:script=latn;language=DFLT;mapping=tex-text
;;"
(fontspec) <->"TeX Gyre Heros/B/ICU:script=latn;language=DFLT;mapping=tex-text
;;+smcp;"
(fontspec) <->"TeX Gyre Heros/I/ICU:script=latn;language=DFLT;mapping=tex-text
;;"
(fontspec) <->"TeX Gyre Heros/I/ICU:script=latn;language=DFLT;mapping=tex-text
;;+smcp;"
(fontspec) <->"TeX Gyre Heros/BI/ICU:script=latn;language=DFLT;mapping=tex-tex
t;;"
(fontspec) <->"TeX Gyre Heros/BI/ICU:script=latn;language=DFLT;mapping=tex-tex
t;;+smcp;"
Unknown feature `' in font `TeX Gyre Heros'.
fontspec info: Font Lucida Grande does not contain any OpenType `Script' infor
mation.
\g_fontspec_family_Lucida Grande_int=\count172
fontspec info: Defining font family 'LucidaGrande(0)' for font 'Lucida Grande'
with options [Ligatures=TeX].
\g_fontspec_LucidaGrande(0)_prop=\toks40
(fontspec) <->"Lucida Grande/AAT:mapping=tex-text;;"
(fontspec) <->"Lucida Grande/B/AAT:mapping=tex-text;;"
fontspec info: Could not resolve font Lucida Grande/I (it probably doesn't exi
st).
(fontspec) <->"Lucida Grande/BI/AAT:mapping=tex-text;;"
Unknown feature `' in font `Lucida Grande'.

I have an extra \setmainfont[Ligatures=TeX]{Lucida Grande}
to see whether there is dependence upon the font itself.
Apparently not.

However, if there are no options
\setmainfont{TeX Gyre Heros}
then there is no extra empty entry as ;; .
Hence also no "Unknown feature `' " message.

So yes, this is surely due to the coding in fontspec.sty
creating that extra blank entry when processing options.
Or it comes from other (LaTeX 3 ?) coding that fontspec just
uses, without later eliminating the redundant entry.

Thus it is slightly inefficient coding that does no real harm.

\cs_new:Nn \fontspec_update_featstr:n {
\bool_if:NF \l_fontspec_firsttime_bool {
\sclist_gput_right:Nx \l_fontspec_rawfeatures_sclist {#1;}
}
}

appending a ';' character following the option string.