Re: “Missing Glyphs!” / LinLibertine
>> Am Dienstag, den 30.01.2018, 13:09 -0500 schrieb Scott Kostyshak: >> Thank you for your detailed explanations. >> Scott > +1. > Jürgen. + 2. It's much appreciated. Seb.
Re: “Missing Glyphs!” / LinLibertine
On 2018-01-30, Scott Kostyshak wrote: > On Tue, Jan 30, 2018 at 09:42:14AM +, Guenter Milde wrote: >> On 2018-01-29, Jürgen Spitzmüller wrote: >> > Am Montag, den 29.01.2018, 10:55 +0100 schrieb S. Noble: >> >> Thanks, Scott and Günter, for your quick, helpful replies. >> >> I had followed Scott’s advice and found the offending character. It >> >> was indeed the lower-case Greek letter “sigma”, not in a word-final >> >> position (i.e., “σ”). (There was just one of them in the document). >> >> I was then puzzling over LyX’s error message — “There is no v in font >> >> LinLibertineT-osf-lgr”. >> > That's not LyX's error message, but LaTeX's (or polyglossia's for that >> > matter). >> While Polyglossia is known for a lot of false positives when checking >> character support, it is not to blame here. Polyglossia tests for script >> support in the used fonts, relying on often missing information in the >> font metadata and gives errors (not just warnings). >> The "There is no … in font …" warnings are from TeX. > Do you think it is worthwhile to report bugs regarding missing > information in the font metadata, or in your experience is it not likely > to lead to fixes? I suppose I would need to find the upstream font > definitions and report them there, for each individual issue? This is about Polyglossia and system fonts, so one may try this but it may be any font on your computer... More important would be a fix for the test (making it a warning only) in Polyglossia. Günter
Re: “Missing Glyphs!” / LinLibertine
Am Dienstag, den 30.01.2018, 13:09 -0500 schrieb Scott Kostyshak: > Thank you for your detailed explanations. +1. Jürgen > > Scott signature.asc Description: This is a digitally signed message part
Re: “Missing Glyphs!” / LinLibertine
On Tue, Jan 30, 2018 at 09:42:14AM +, Guenter Milde wrote: > On 2018-01-29, Jürgen Spitzmüller wrote: > > Am Montag, den 29.01.2018, 10:55 +0100 schrieb S. Noble: > > >> Thanks, Scott and Günter, for your quick, helpful replies. > >> I had followed Scott’s advice and found the offending character. It > >> was indeed the lower-case Greek letter “sigma”, not in a word-final > >> position (i.e., “σ”). (There was just one of them in the document). > >> I was then puzzling over LyX’s error message — “There is no v in font > >> LinLibertineT-osf-lgr”. > > > That's not LyX's error message, but LaTeX's (or polyglossia's for that > > matter). > > While Polyglossia is known for a lot of false positives when checking > character support, it is not to blame here. Polyglossia tests for script > support in the used fonts, relying on often missing information in the > font metadata and gives errors (not just warnings). > The "There is no … in font …" warnings are from TeX. Do you think it is worthwhile to report bugs regarding missing information in the font metadata, or in your experience is it not likely to lead to fixes? I suppose I would need to find the upstream font definitions and report them there, for each individual issue? > ... Thank you for your detailed explanations. Scott
Re: “Missing Glyphs!” / LinLibertine
On 2018-01-29, Jürgen Spitzmüller wrote:
> Am Montag, den 29.01.2018, 10:55 +0100 schrieb S. Noble:
>> Thanks, Scott and Günter, for your quick, helpful replies.
>> I had followed Scott’s advice and found the offending character. It
>> was indeed the lower-case Greek letter “sigma”, not in a word-final
>> position (i.e., “σ”). (There was just one of them in the document).
>> I was then puzzling over LyX’s error message — “There is no v in font
>> LinLibertineT-osf-lgr”.
> That's not LyX's error message, but LaTeX's (or polyglossia's for that
> matter).
While Polyglossia is known for a lot of false positives when checking
character support, it is not to blame here. Polyglossia tests for script
support in the used fonts, relying on often missing information in the
font metadata and gives errors (not just warnings).
The "There is no … in font …" warnings are from TeX.
>> It didn’t seem to make sense, for two reasons: (1) the problem
>> was certainly the letter “sigma”, but in principle it has nothing to
>> do with a “v”; and (2) in the PDF print-out, the letter “sigma” is
>> printed correctly. So, all in all, nothing is missing. I couldn’t
>> figure it out or find anything about it, and then came Günter’s
>> explanation.
The problem is that currently, in LyX's Unicode translation file we use the
Latin transcription for Greek:
In the Latin transcription, "s" stands for "\textautosigma", which
depending on the next characters is printed as "sigma" or "finalsigma".
This is fine if you type Greek text using Latin characters but bad if in
a document you write a single literal "σ" and in the PDF you get a single
"ς". Therefore, LyX uses the "never-to-become-finalsigma" transcript "sv"
with the "LGR ligature break character" "v" (cf.
https://www.lyx.org/trac/ticket/4966):
0x03c3 "\\textgreek{sv}" "textgreek"
"force=iso8859-7;cp437;cp437de;cp865,mathalpha" "\\sigma" "" # GREEK SMALL
LETTER SIGMA
That "v" is missing in Libertine LGR is a bug in this font. May be they call
it a feature, because the effect of a missing character is almost the same
as a "ligature break" character: it prevents ligating (fine) but
* it also triggers a warning
* there is a/no ligature break in the text if you copy-paste from the PDF.
LyX turns warnings about missing characters into errors for good reason.
Unfortunately this can give rise to false positives in rare cases.
We had a similar problem with Unicode fonts lacking the ligature break
character (u200C, ZWNJ). However, we cannot tell LyX to ignore missing
"v" characters (which stand for ZWNJ in the LGR font encoding but are
printable characters normally).
The proper solution is to use the LaTeX internal character representation
(LICR) in the Unicode translation file:
0x03c3 "\\textgreek{\\textsigma}""textgreek"
"force=iso8859-7,mathalpha" "\\sigma" "" # GREEK SMALL LETTER SIGMA
A patch is ready since some years but hold back because Greek LICR macros
were a new feature at this time. I believe it is time to apply the patch
now. It works stable on my local installation and fixes the problem + some
others.
>> So at least I’ve understood (sort of), but this is strange behaviour
>> from LyX. LaTeX, on the other hand, doesn’t complain and prints the
>> “sigma” identically.
> You get the message as well, LaTeX is just not aborted (since it's a
> warning, actually).
This depends: If you use literal Greek characters in the LaTeX source, there
is no warning.
A simple workaround in LyX is to set the LaTeX input encoding to UTF-8:
Document>Settings>Language>Encoding>Other: Unicode (utf8)
("Language default" does not work, because we force-convert sigma to
override inputenc's "iso8859-7.def" replacing it with "autosigma".
https://www.lyx.org/trac/ticket/8509)
>> I'm wondering if the problem might be solved by LyX following LaTeX's
>> advice. If, as LaTeX complains, "The libertine-type1 package is
>> obsolete", then perhaps LyX should be using, as
>> LaTeX's suggests, "\usepackage[type1]{libertineRoman}" in the
>> Preamble instead of "\usepackage{libertine-type1}".
This will not solve the warning issue (the font itself is the same).
Nevertheless, it may be a good idea to update the font call once we can
safely expect users to have the new interface or after a test in the LaTeX
configuration.
...
Günter
Test document:
#LyX 2.2 created this file. For more info see http://www.lyx.org/
\lyxformat 508
\begin_document
\begin_header
\save_transient_properties true
\origin unavailable
\textclass scrartcl
\begin_preamble
\usepackage[type1]{libertineRoman}
\end_preamble
\use_default_options false
\maintain_unincluded_children false
\language ngerman
\language_package default
\inputencoding utf8
\fontencoding global
\font_roman "default" "default"
\font_sans "default" "default"
\font_typewriter "default" "default"
\font_math "auto" "auto"
\font_default_family default
\use_non_tex_fonts false
\font_sc false
\font_osf false
\font_sf_scale 100 100
\font_tt_scale 100 1
Re: “Missing Glyphs!” / LinLibertine
2018-01-30 9:42 GMT+01:00 S. Noble : > It’s interesting to hear all that. As nice as the Linux Libertine font is, > it’s the one that gives me the most issues, such as the one described > above. I’m sure that the hassles I experience are different from the ones > you face, but when you say that it’s getting annoying, I fully agree. > Never mind, this is just some ranting. It's not too hard to deal with this. > Again, LyX is fantastic software—keep up the great work! > Thank you. Jürgen > Seb. >
Re: “Missing Glyphs!” / LinLibertine
2018-01-30 9:42 GMT+01:00 S. Noble : > > Yes, we can (and will) do that. But it's getting annoying, to be honest. > This is the fourth time a new package for libertine is released > and the old one declared obsolete over the last 5 years. We > had libertine-type1, which was then declared obsolete by libertine > (or libertine-legacy), which was then declared obsolete by a new > libertine-type1. Now libertineRoman declares libertine-type1 obsolete. This > is just crazy. > > Jürgen > > > It’s interesting to hear all that. As nice as the Linux Libertine font is, > it’s the one that gives me the most issues, such as the one described > above. I’m sure that the hassles I experience are different from the ones > you face, but when you say that it’s getting annoying, I fully agree. > I simply use the OpenType Libertinus fonts (an actively developed fork of Linux Libertine and Biolinum) with XeTeX https://github.com/khaledhosny/libertinus No problem at all. [If you don't mind the self promotion, I wrote for my personal blog a quick introduction to the use of OpenType on XeTeX/LyX. This is the first article of the "miniseries" https://frommindtotype.wordpress.com/2018/01/15/using-opentype-on-xetexlyx-short-texts/ The fourth and last one will be published next Monday.] Regards, Ricardo > Again, LyX is fantastic software—keep up the great work! > > Seb. >
Re: “Missing Glyphs!” / LinLibertine
> Yes, we can (and will) do that. But it's getting annoying, to be honest. This > is the fourth time a new package for libertine is released > and the old one declared obsolete over the last 5 years. We had > libertine-type1, which was then declared obsolete by libertine (or > libertine-legacy), which was then declared obsolete by a new libertine-type1. > Now libertineRoman declares libertine-type1 obsolete. This is just crazy. > > Jürgen It's interesting to hear all that. As nice as the Linux Libertine font is, it's the one that gives me the most issues, such as the one described above. I'm sure that the hassles I experience are different from the ones you face, but when you say that it's getting annoying, I fully agree. Again, LyX is fantastic software--keep up the great work! Seb.
Re: “Missing Glyphs!” / LinLibertine
Am Montag, den 29.01.2018, 10:55 +0100 schrieb S. Noble:
> Thanks, Scott and Günter, for your quick, helpful replies.
> I had followed Scott’s advice and found the offending character. It
> was indeed the lower-case Greek letter “sigma”, not in a word-final
> position (i.e., “σ”). (There was just one of them in the document).
> I was then puzzling over LyX’s error message — “There is no v in font
> LinLibertineT-osf-lgr”.
That's not LyX's error message, but LaTeX's (or polyglossia's for that
matter).
> It didn’t seem to make sense, for two reasons: (1) the problem
> was certainly the letter “sigma”, but in principle it has nothing to
> do with a “v”; and (2) in the PDF print-out, the letter “sigma” is
> printed correctly. So, all in all, nothing is missing. I couldn’t
> figure it out or find anything about it, and then came Günter’s
> explanation.
> So at least I’ve understood (sort of), but this is strange behaviour
> from LyX. LaTeX, on the other hand, doesn’t complain and prints the
> “sigma” identically.
You get the message as well, LaTeX is just not aborted (since it's a
warning, actually).
> I'm wondering if the problem might be solved by LyX following LaTeX's
> advice. If, as LaTeX complains, "The libertine-type1 package is
> obsolete", then perhaps LyX should be using, as
> LaTeX's suggests, "\usepackage[type1]{libertineRoman}" in the
> Preamble instead of "\usepackage{libertine-type1}".
Yes, we can (and will) do that. But it's getting annoying, to be
honest. This is the fourth time a new package for libertine is released
and the old one declared obsolete over the last 5 years. We had
libertine-type1, which was then declared obsolete by libertine (or
libertine-legacy), which was then declared obsolete by a new libertine-
type1. Now libertineRoman declares libertine-type1 obsolete. This is
just crazy.
Jürgen
signature.asc
Description: This is a digitally signed message part
Re: “Missing Glyphs!” / LinLibertine
> Using the above approach you can narrow your document down and find
> which character is causing the issue.
>
> Best,
>
> Scott
>
> The v is defined as a "ligature break" (\textcompwordmark in LaTeX)
> in the LGR font encoding. It is, e.g. used to prevent sigma (σ) becoming a
> final sigma (ς) at the end of a word.
>
> It is missing in Libertine LGR.
> As the ligature break is an invisible character, it may be there is no
> problem with you file. Just check Greek parts -- maybe combined with the
> method proposed by Scott.
>
> Günter
Thanks, Scott and Günter, for your quick, helpful replies.
I had followed Scott's advice and found the offending character. It was
indeed the lower-case Greek letter "sigma", not in a word-final position
(i.e., "σ"). (There was just one of them in the document).
I was then puzzling over LyX's error message -- "There is no v in font
LinLibertineT-osf-lgr". It didn't seem to make sense, for two reasons:
(1) the problem was certainly the letter "sigma", but in principle it
has nothing to do with a "v"; and (2) in the PDF print-out, the letter
"sigma" is printed correctly. So, all in all, nothing is missing. I
couldn't figure it out or find anything about it, and then came Günter's
explanation.
So at least I've understood (sort of), but this is strange behaviour
from LyX. LaTeX, on the other hand, doesn't complain and prints the
"sigma" identically.
I'm wondering if the problem might be solved by LyX following LaTeX's
advice. If, as LaTeX complains, "The libertine-type1 package is
obsolete", then perhaps LyX should be using, as LaTeX's suggests,
"\usepackage[type1]{libertineRoman}" in the Preamble instead of
"\usepackage{libertine-type1}". For my part, I have no idea.
Anyway, thanks for your help. LyX is fantastic software--keep up the
development!
Sebastien
Re: “Missing Glyphs!” / LinLibertine
> On Sun, Jan 28, 2018 at 04:03:24PM +, S. Noble wrote: >> Hello, List Members, >> I'm experiencing strange issues with LyX's "Libertine" font when I >> attempt to compile a document. I'm now using LyX 2.2.2 on Mac OSX >> 10.11.6 (El Cap.). I didn't have these issues in the past, when I first >> created the documents. (I can't remember on which version of LyX they >> were created, but it was't that long ago). >> One of the documents, for example, is in French and German, with French >> being the main language. When I hit "View" or "Update", LyX gives me the >> following error: >> - "Missing Glyphs!" >> with the following description: >> - "Missing character: There is no v in font LinLibertineT-osf-lgr!" I could reproduce: LinLibertineT-osf-lgr is a font in LGR (logal GReek) encoding, so the problem seems to be in a Greek text part. The v is defined as a "ligature break" (\textcompwordmark in LaTeX) in the LGR font encoding. It is, e.g. used to prevent sigma (σ) becoming a final sigma (ς) at the end of a word. It is missing in Libertine LGR. As the ligature break is an invisible character, it may be there is no problem with you file. Just check Greek parts -- maybe combined with the method proposed by Scott. Günter
Re: “Missing Glyphs!” / LinLibertine
On Sun, Jan 28, 2018 at 04:03:24PM +, S. Noble wrote: > Hello, List Members, > > I'm experiencing strange issues with LyX's "Libertine" font when I > attempt to compile a document. I'm now using LyX 2.2.2 on Mac OSX > 10.11.6 (El Cap.). I didn't have these issues in the past, when I first > created the documents. (I can't remember on which version of LyX they > were created, but it was't that long ago). > > One of the documents, for example, is in French and German, with French > being the main language. When I hit "View" or "Update", LyX gives me the > following error: > > - "Missing Glyphs!" > > with the following description: > > - "Missing character: There is no v in font LinLibertineT-osf-lgr!" > > I've tried to scan the source code, but can't find the offending > character(s). > > If I chose "Show Output Anyway", the document appears to compile > correctly, although the document is long and complex and I can't run the > risk of a missing character in a footnote somewhere. Dear S. Noble, thanks for all of the details you give. I do not how to fix the issue. However, it is possible that you had the same issue as you did before upgrading LyX, it's just that LyX did not report the error. The reason is that starting as of 2.2.0, LyX started reporting this missing characters as errors, partly because of the reason you mention in the paragraph above (if you are missing a character e.g. in a footnote somewhere you likely won't realize on your own). This change in LyX's behavior was documented in the release notes of LyX 2.2.0: https://wiki.lyx.org/LyX/ReleaseNotes#toc2 Although I don't know how to fix the problem, I can give you a better idea of how to find it: 1. backup your document 2. delete half of your document 3. if you still get an error, go to step 2. 4. if you do not get an error after deleting half of your document, undo deleting. Go to step 2 and delete the *other* half of your document. Using the above approach you can narrow your document down and find which character is causing the issue. Best, Scott signature.asc Description: PGP signature
“Missing Glyphs!” / LinLibertine
Hello, List Members,
I'm experiencing strange issues with LyX's "Libertine" font when I
attempt to compile a document. I'm now using LyX 2.2.2 on Mac OSX
10.11.6 (El Cap.). I didn't have these issues in the past, when I first
created the documents. (I can't remember on which version of LyX they
were created, but it was't that long ago).
One of the documents, for example, is in French and German, with French
being the main language. When I hit "View" or "Update", LyX gives me the
following error:
- "Missing Glyphs!"
with the following description:
- "Missing character: There is no v in font LinLibertineT-osf-lgr!"
I've tried to scan the source code, but can't find the offending
character(s).
If I chose "Show Output Anyway", the document appears to compile
correctly, although the document is long and complex and I can't run the
risk of a missing character in a footnote somewhere.
Even more strangely, if I export to LaTeX and compile using a TeX
editor, the document compiles correctly with no issues at all.
However, the LaTeX compile log file does indicate a warning:
- "The libertine-type1 package is obsolete. For new documents, please
use \usepackage[type1]{libertineRoman}"
I note that LyX (2.2.2) still puts "\usepackage{libertine-type1}" in the
Preamble.
This looks like a font problem, and I know that these things are a
nuisance. I use the "Libertine" font a lot, and I'd like it just to
work, as it used to. I can't run the risk of losing characters in long,
complex documents.
I've searched for possible similar problems on the forum and elsewhere,
but have found none. If anyone knows of the problem, and how to fix it,
I'd greatly appreciate any pointers anyone might be able to provide.
Sebastien
