alerque / libertinus Goto Github PK

Dear Khaled,

the package amsmath allows to have braces of different sizes using \bigl\{ ... \bigr\}. The spacing seems to be wrong and the style of the larger braces (curved) does not match that of the smaller ones (straight). The long vertical bar \big| is larger than the corresponding braces. See for example the following document:

`\documentclass{article}
\usepackage{fontspec}
\usepackage{unicode-math}
\usepackage{amsmath}

\begin{document}
[
\bigl{ { s\in S,|, f(s) = f(t)},\big|,t\in S\bigr}
]

\setmainfont[Mapping=tex-text]{Libertinus Serif}
\setsansfont[Mapping=tex-text]{Libertinus Sans}
\setmathfont{Libertinus Math}

[
\bigl{ { s\in S,|, f(s) = f(t)},\big|,t\in S\bigr}
]
\end{document}`

which yields (using xelatex):

Let me thank you for the great work you did on the beautful Libertinus font!

With best regards
Ralf.

Hi Khaled,
Following up my email, I report here the bug I noticed in MS Word (2010) - I am not sure if happens also with the LaTeX.

When I write a fraction within round brackets, the brackets appear truncated and smaller than the fraction. They are definetely different from the appearance in the file "libertinus-testmath.pdf"

I also noticed other aesthetic problems with the square brackets, the square root symbol, the integral symbol, the double integral symbol, slash with fractions, summation symbol, etc. Please find below the attached files with all the examples I am referring to.

Libertinus math bugs.pdf

Libertinus math bugs.docx

All the best,
Claudio

The official Hanyu Pinyin specification has some different shapes from many Latin based scripts.

1. a looks similar to handwritten style
2. tone 2 looks like reversed acute

You can seen from following official sources:
http://www.moe.gov.cn/s78/A19/yxs_left/moe_810/s230/195802/t19580201_186000.html
http://www.china-language.gov.cn/126/2008_2_1/3_126_3278_0_1201864229916.html
http://www.moe.gov.cn/publicfiles/business/htmlfiles/moe/s230/201001/75607.html

Additionaly, seen from Xiandai Hanyu Cidian (现代汉语词典, 現代漢語詞典) and Xinhua Zidian (新华字典, 新華字典), these letters have different mark positing from Vietnamese: ế, ề

The combining acute accent U+0301 is misplaced on Cyrillic uppercase vowels and on wide lowercase ones.

The bug is also on legacy Linux Libertine fonts.

Example file

\documentclass{article}

\usepackage{fontspec}
\setmainfont{Libertinus Serif}
\setsansfont{Libertinus Sans}

\newcommand\CyrillicAccents{А а А́ а́ \\ Е е Е́ е́ \\ И и И́ и́ \\ О о О́ о́ \\
  У у У́ у́ \\ Ы ы Ы́ ы́ \\ Э э Э́ э́ \\ Ю ю Ю́ ю́ \\ Я я Я́ я́}

\begin{document}

\parbox[t]{3cm}{\CyrillicAccents}
\parbox[t]{3cm}{\sffamily\CyrillicAccents}

\end{document}

Reference: http://tex.stackexchange.com/questions/199080/misplaced-accents-in-cyrillic-text

Output

I offer these suggestions

• GPL: Well, pedantic I am :) I suggest to drop it, its rather useless for fonts because it may require documents to be GPL compatible also, and should be discouraged
• Changelog.libertine: rename to FONTLOG.linuxlibertine.txt
• LICENSE.txt: drop it
• README.linuxlibertine: append .txt
• remove binaries from the repo root, only distribute them via release zips
• OFL.txt put the copyright notices all on 1 line, so they match the copyright string inside the sources

I have bit of italicized text inside curly quote marks that come from the roman set. The resulting kerning has the ! clobbered by the ”.

I'm not even really sure whether to blame the font or the typesetter here as I don't know how kerning works across styles, but it seems like this is something that should work somehow. Since it fails in both SILE and LibreOffice (while also working reasonable well for other fonts) it seems like it might be a font thing. Here's a contrived example in LibreWriter:

Thank you, so much, in advance.

The following MWE illustrates the issue. The \lVert nearly runs into the tail of the italic f in math mode. I believe the same goes for the bold version. In the MWE, I compare with CMR.

%!TEX program = lualatex
%
\documentclass{article}
\usepackage{unicode-math}

\begin{document}
\[\lVert f \rVert\]
\setmathfont{Libertinus Math}
\[\lVert f \rVert\]
\end{document}

This, for me, produces the following

Here are my versions: (let me know if I missed something important)

Libertinus, Latest git
LuaTeX, Version beta-0.80.0 (TeX Live 2015/Debian) (rev 5238)
luaotfload.sty 2015/12/09 v2.6 OpenType layout system
lualatex-math.sty 2015/09/22 v1.5 Patches for mathematics typesetting with Lu
unicode-math.sty 2015/09/24 v0.8c Unicode maths in XeLaTeX and LuaLaTeX
fontspec.sty 2016/02/01 v2.5a Font selection for XeLaTeX and LuaLaTeX

Thanks again. I'm willing to help, but I have no experience with font editing to speak of. Let me know if there is anything I can do.

Some people and complaining about this, I don’t know for sure what this font was for so logging this here for reference.

Using autoinst [1] it is quite easy to generate type 1 fonts from otf fonts in various encodings. I use type 1 versions of libertine serif for quite some time without problems. It would be nice to see them included in the official package, though. Using fontspec and luatex is not always an option.

[1] https://www.ctan.org/pkg/fontools

With Linux Libertine, Qu ligature (as in Quasi-) works both with XeLaTeX and LuaLaTeX, with Libertinus however it only works with XeTeX. Other ligatures work fine with both engines (see the picture made with LuaLaTeX).

I am using the current version from TeXLive 2015, i.e. 6.2, with last record in the changelog from 30 Jan 2016.

Dear Khaled,
thanks for working on this font. My question: Why do you fork this project under a new name? I think the new name is a bit confusing. I see no reason not modify the Libertine/Bi­olinum fonts directly as a new maintainer. The original author would probably be happy with such solution.

Greetings from Berlin,
...Rolf

Hi
would it be possible to have a libertinus math bold font?

I have also posted about this issue on tex.SE.

For some reason, in the math mode, alternative glyphs are used for any numerals whose size is less than \normalsize, explicitly (that is, through \small, \scriptsize, \footnotesize, and \tiny) or implicitly (for example, through ^ and _), leading to unpleasant kerning for consecutive numerals etc etc. Is this an intended behaviour?

Currently I wrap all sub/superscripts in \mathrm as a workaround.

\documentclass{article}
\usepackage{unicode-math}
\setmainfont{Libertinus Serif}
\setmathfont{Libertinus Math}

\begin{document}

$1234567890^{1234567890}$

$1234567890^\mathrm{1234567890}$

\end{document}

Dear maintainer:
consider the following code:

\documentclass{article}
\usepackage{unicode-math}
\setmathfont[Extension=.otf]{libertinusmath-regular}
\begin{document}
\(\scriptstyle \left(\right)\)
\end{document}

The output, when compiled with xelatex, are two crossed boxes. They should be really parentheses.

The same bug occurs with \mleft, \mright from the package mleftright.
No bug occurs with lualatex. No bug occurs when other fonts are used, e.g. TeX Gyre Pagella.
Any chance for a bugfix?

First of all thanks for your great work, Khaled!

I encountered the following issue concerning Libertinus Math: Using \frac within a \sqrt results in typesetting a root only spanning the numerator as well in XeLaTeX as in LuaLaTeX (Linux, Mac OS X and Windows 10; using most recent versions of each provided by TeX Live 2015).

oldman77

In German typography, ligatures are suppressed across morpheme boundaries. In a word like "Auflage", 'edition', there should be no fl-ligature, as this word is composed from the prefix 'auf-' and the noun 'Lage'. This is the reason for the creation of the selnolig package (https://www.ctan.org/pkg/selnolig).

Suppressing a ligature is often not enough, though, since this will often just cause the regular characters to collide. When suppressing a ligature, it is often necessary to access alternative glyphs. In Linux Libertine, there is an alternative glyph of 'f' that has a shorter arm so that it doesn't collide with a following character. I've found documentation that indicates that this was once accessed through an OpenType feature, but it was later removed, yet the glyph stayed in the font, now accessible in the PUA.

I can therefore use the short armed 'f' by accessing it directly. Example:

\documentclass{article}
\usepackage{fontspec}
\setmainfont{LinLibertine_Rah}[Path=L:/fonts/LinLibertine5.3.0/]
\begin{document}
Auflage -- Au\char"E0E0lage
\end{document}

It's dirty, because this probably breaks hyphenation, and the output 'f' will probably not map to the regular 'f' when copying and searching in the pdf.

But in Libertine Serif, the glyph has been moved into the unencoded area, and now it's not accessible in any way -- not via an OpenType feature, and not by typing it directly.

\documentclass{article}
\usepackage{fontspec}
\setmainfont{LinLibertine_Rah}[Path=L:/fonts/LinLibertine5.3.0/]
\newfontfamily{\libertine}{libertineserif-regular}[Path=L:/fonts/libertine-master02/, Ligatures = TeX]
\begin{document}
Linux Libertine:

Auflage -- Au\char"E0E0lage\\

Libertine Serif:

{\libertine{Auflage} -- ?}
\end{document}

Ideally there should be an OpenType feature that would allow the short armed 'f' to be accessed, one that would correctly map to Unicode 'f' when copying and searching in the pdf.

EDIT: The glyph is named "f.short" in both fonts. It's probably possible with LuaTeX to write some fancy code that will replace "f" with "f.short", but I think an OpenType feature would be a better way to do it.

This needs to be redesigned, both the math and the text ones.

At first: Thank you very much for your great work!

I think the minus sign is not correctly aligned with the fraction bar.
It is already visible in the supplied libertinus-testmath.pdf, but I created an example to make it clearer.

This also causes a problem with \xrightarrow etc.
On some zoom steps, every single minus sign is visible.

In XITS Math, they are almost perfectly aligned.
(The fraction bar is a tiny little bit lower, but \xrightarrow is perfect)

Seen from SourceForge
https://sourceforge.net/p/linuxlibertine/discussion/307868/thread/3ed15788/
Someome repored old Danish has a special ligature for “Qv”, is it possible to implement it?

The original Linux Libertine and Linux Biolinum fonts were licensed under both OFL and GPL+Font exception, but Libertinus is licensed under OFL only, need GPL+FE to follow it.

I see that the file /BUGS lists some of the existing issues for Linux Libertine at SourceForge as either Fixed, Invalid or Ignored, but not all of them. Should we raise the Open ones again on GitHub (if they apply) to keep better track of them?

• 037 MacOSX-10.5: “Defects in several kern-tables” (2011-11-07)
• 039 ☹ and ☺ (smileys) are too similar (2011-11-07) #307, #183
• 063 Several issues in "Combining Diacritical Marks" block (2011-11-07) #131?
• 077 combining iota subscript (2011-11-07)
• 078 U+2620 skull and crossbones and other misc symbols (2011-11-07)
• 084 Greek doesn't bold very well in Biolinum (2011-11-07)
• 086 XeLaTeX, Small Caps and vocals with macron accent (2011-11-07)
• 088 Wrong Icelandic text in multilingual.pdf (2015-07-10)
• 093 The placement of Dagesh in Hebrew letters is incorrect (2010-06-30)
• 094 line spacing is bugged (2010-06-30)
• 101 Problems with OpenType Ligatures in Linux Libertine O (2010-07-31)
• 103 ZWNJ/​ZWJ (U+200c/​0d) do not work as expected (2011-01-06)
• 108 Missing Old/​Extended Cyrillic letters (2011-01-05)
• 109 incorrect hinting of Latin small letter e (2010-09-23)
• 120 diacritics collide in letter o with dot above and macron (2011-11-07) #131?
• 122 Three exclamation marks (2011-11-07)
• 123 Composing vulgar fractions from subscripts and superscripts (2011-11-07)
• 124 Libertine Opentype does not display properly in Word 2010 (2011-11-07)
• 125 Libertine Opentype does not display properly in Word 2010 (2011-11-07)
• 127 Ellipsis and closing quote characters too tight (2011-11-07)
• 128 e-macron is misplaced when 'mark' feature is on (2011-03-28)
• 130 no small scaps for h with dot and a with bar (2011-11-07)
• 133 Greek small capitals and accents (2011-11-07)
• 134 Kerning (2011-11-07)
• 135 Kerning (2011-11-07)
• 136 Fonts package naming (2011-06-08)
• 141 Linux Libertine O has problems with italics, and ligatures (2011-06-08)
• 144 Biolinum Cyrillic Bold (2011-06-03)
• 145 Wrong slot for Indian Rupee Sign (2011-11-07)
• 146 XeLaTex: Problems with Linux Biolinum O with bold small caps (2011-11-07)
• 152 Bad ligatures again (2012-07-09)
• 154 Re - Accute accentuated barred i (2011-07-04)
• 157 Standardize font naming convention (2011-11-07)
• 158 Opera 11.50 displays Libertine Regular as Bold (2011-11-07)
• 159 Kerning of AW versus WA (2011-11-07)
• 166 Kerning of quotation marks “ ” ‘ ’ „ ‚ etc. (2014-08-22)
• 167 website: unintended redirect (2011-11-07) 🚫 Invalid
• 170 Wrong italic form of Serbian lowercase dje (ђ) (2011-11-07), duplicate of 271, hence #246
• 172 Problems with Libertine and Beamer Templates (2012-06-24)
• 174 Too small width of EM-Space (Unicode 2003, GID 1727) ) (2011-09-19)
• 176 Kerning of !!! (2011-11-07) #122
• 180 Kerning between two IPA symbols to small (2011-10-12)
• 183 Space sizes - hair, thin, and more (2011-11-18)
• 197 In xelatex, \usepackage and setmainfont is different. (2011-12-22)
• 201 Issues with Arrows, sub- and superscripts in Word (2012-02-06)
• 202 Italic superscripts and subscripts (2012-03-09)
• 211 No build script in the source (2012-05-10)
• 218 Linux Libertine Italic: vertical symbols (2012-06-02)
• 222 ſ (long s) is too narrow in Mono (2015-07-10)
• 223 oldstyle number bold and italic 6 is like 0 (2013-11-18)
• 224 5.3 messed up printing on Canon Image Runner (2015-09-15)
• 225 LinLib Italic/​Semibold/​Bold inconsistencies (2012-07-15)
• 226 U+0344 Combining Greek Dialytika Tonos (2012-07-20)
• 227 <,> <-> not superscript in superscript texts (2012-08-03)
• 228 Unexpected behaviour of \radical0 (2012-08-09)
• 229 Linux Libertine Mono: not showing up in monospaced font list (2015-09-28) #411
• 230 Ligatures c_h.alt, T_h.alt, f_h.alt, longs_h.alt missing (2012-08-30)
• 231 Missing kerning for é (2012-09-10)
• 232 Missing kerning for é (2012-09-11)
• 233 cfftot1 segfaults on italic fonts (2012-09-16)
• 234 cfftot1 segfaults on italic fonts (2012-09-17)
• 235 ligatures missing in LinLibertine_RBI.otf (2012-09-24)
• 236 Spacin between ťa and ťk (2012-09-25)
• 237 Libertine 5.3.0 - square bracket mismatch (2014-05-16)
• 238 missing smallcapital variant for amacron (2016-07-17)
• 239 Several ligatures produce output that is not copyable (2012-10-06)
• 240 Smal-Caps Problem in Illustrator, Photoshop (Cyrillic/​Greek) (2012-10-09)
• 241 Printing failure (2012-10-10)
• 242 some additions and corrections (2012-10-24)
• 243 Wrong kerning for uni2115 (2012-10-29)
• 244 Font-Squirell doesn't handle regular (2012-11-01)
• 245 Biolinum bold italic is missing (2012-11-02)
• 246 text disappears on sony reader when using the TTF-font (2012-11-02)
• 247 Can’t switch old-style numerals to lining numerals (2012-11-05)
• 249 "Libertine Italic" and "Libertine Initials" seen as dupes (2012-12-06)
• 250 ch and ck ligatures are not activated by +liga as advertised (2012-12-14) – still a valid bug!
• 251 5.3 Printing failure (2013-01-14)
• 252 Accents over capital ezh are ugly (2013-01-17)
• 253 Combining superscript letters extremely small (2013-02-23)
• 254 Poor kerning for multiple exclamation marks (2013-02-27) #122
• 255 U+01B3 hook on the right (2013-03-16)
• 256 U+018E should be turned not reversed (2013-03-17)
• 257 U+0264 should have hooked terminals (2013-03-17)
• 258 different strokes (2013-03-17)
• 259 U+0192 should be upright for Ewe language (2014-04-04)
• 260 U+0165 character has wrong kerning (2013-03-30)
• 261 U+0165 and misplaced caron in Libertine Mono (2013-03-30)
• 262 ligatures in Capitals fonts (2013-04-17)
• 263 Typo in Multilingual sample (2013-04-19)
• 264 Miissing glyph (in LinuxLibertine 5.3.0 Mono [otf, ttf]) (2013-04-20)
• 265 "Mono" as subfamily? (2013-04-25)
• 266 combining Greek hypogegrammeni (2013-04-27)
• 268 Poorly positioned U+0312 COMBINING TURNED COMMA ABOVE (2013-05-12)
• 269 Accents at ligated letters (2013-05-12)
• 270 U+0466–U+046D CYRILLIC * YUS (2013-05-12)
• 271 Wrong U+0452 CYRILLIC SMALL LETTER DJE in italic (2013-05-13) #246
• 272 Diacritics in Linux Libertine SemiBold (2013-05-13) – still a valid bug!
• 273 EZH too thin in Linux Libertine Bold (2013-05-18)
• 274 Comma below different between precomposed and combining (2013-05-25)
• 275 Wrong anchor point type for retroflex approximant (U+027B) (2013-06-08)
• 276 Rounded lower-case italic v and w (2013-06-16)
• 277 add superscript symbols for latex footnotes (2013-06-21) ❌ Ignored
• 278 Hyphenation and triple letter (2013-07-04) 🚫 Invalid
• 279 Support for Serbian glyphs buggy (2015-04-29)
• 280 Two diacritics using combining characters doesn’t work (2013-09-20) #131?
• 286 W-vowel and V-vowel need kerning (2014-03-01) 🚫 Invalid
• 287 inconsistent style of uppercase letters (especially M and W) (2014-03-18) 🚫 Invalid
• 288 Feature Request: Google Font Upload (2014-03-21)
• 289 font height (2014-04-04) 🚫 Invalid
• 290 Biolinum khook is from Libertine (2014-04-04)
• 291 Adhere to Adobe glyph naming AGL(FN) (2014-04-04)
• 292 MUFI 3.0 support (2014-04-04) #101
• 293 Coverage charts in documentation (2014-04-04)
• 294 Restore Greek small caps (2014-12-23)
• 295 Petite caps from unicoded small caps (2014-04-17)
• 296 Biolinum Sytlistic Set 2 (ss02) has Libertine glyphs (2014-05-05)
• 297 Several IPA/​African letters in Biolinum have Libertine serifs (2014-05-20) #42 #96?, #97?
• 298 Line height issue with italics (2014-05-31) #98?
• 299 Enforcing Linux Libertine in Browser results in broken websites due to font abuse (2014-06-12) 🚫 Invalid
• 300 Add IPA symbols (2014-08-27) #438, #429
• 301 Vietnamese diacritics are misplaced and base letters are missing in bold and bold italic (2014-06-30) #131 #157
• 302 Cyrillic + Bold + Italic + XeLaTeX -> Glyphs missing (2014-12-23)
• 303 Modifier letters behaving as combining (2014-07-09)
• 304 Biolinum Bold Italic (Slanted?) (2014-07-29)
• 305 Missing "E" in Linux Libertine Initials 5.0.6 (2014-08-07) ❌ Ignored
• 306 Propotional number spacing in MS Word unsatisfactory (2014-08-13) ❌ Ignored
• 307 U+003A COLON floats above the baseline (compare to U+002E FULL STOP) (2014-09-01) 🚫 Invalid
• 308 Khmer combining character 0x17CC does not combine (2014-08-29) 🚫 Invalid
• 309 Require for Russian Ruble sign (2015-03-10) #457, #255
• 310 Bad repositioning on Windows 8.1 (2015-03-10) ❌ Ignored
• 311 wrong class ("base") of uni0306 in RZI (2014-12-16)
• 312 Two Cyrillic letters looks incorrect (2015-03-10)
• 313 Italic correction is insufficient for brackets (2015-03-12) 🚫 Invalid
• 314 use ttfautohint 1.3+ (2015-04-14) 🚫 Invalid
• 315 Remove Linux from fonts NAME table and filename (2015-04-14)
• 316 Broken links on "Files & License" page (2015-06-25) 🚫 Invalid
• 317 missing U+A78D Latin Capital Letter Turned H (2015-07-10)
• 318 serifs at U+0265 Latin Small Letter Turned H (2015-07-10)
• 319 Italic appearance in mono font (2015-09-30)
• 320 Word 2010: Text disappears when using Biolinum ligature ft like in 'sanft' (2015-10-08)
• 321 Error in GPOS table (2016-03-14)
• 322 Missing kerning between t and æ for small caps. (2016-03-31)
• 323 Missing glyphs for liturgical Latin (easy) (2016-05-27) #85
• 324 U+0171 lacks its accent in monospace font (5.3.0) (2016-06-21) 🚫 Invalid

PS: Converted table to checklist on 15 March 2021 and incorporated updates from comments below up until then.

Hi!

Thank you for implementing math in such a nice font. Though I've found that two glyphs which are very common in Russian math typography are missing: they are slanted "less or equal" (unicode 2A7D) and slanted "greater or equal" (unicode 2A7E). Could you implement them?

The following blocks should be completed:

• Combining Diacritical Marks
• Currency Symbols
• Cyrillic (several non-Slavonic letters)
• Dingbats (circled digits)
• Enclosed Alphanumerits
• Greek and Coptic, except some “Coptic” letters of this block
• Hebrew
• Latin Extended-C and -D
• Letterlike Symbols
• Mathematical Operators
• Miscellaneous Mathematical
• Miscellaneous Technecal
• Number Forms
• Phonetic Extensions
• Phonetic Extensions Supplement

Hi Khaled – This is what I see when I open the Serif Semibold Italic:

It's the only serif that looks like this, the only one that looks like a sans.

I have just verified that bug #297 filed against Linux Biolinum at Sourceforge also applies to Libertinus Sans (Regular and Italic), except for Khook (also see SF#290). The listed glyphs show serifs although this font is s upposed to be sans-serif. I didn’t check PUA and hebrew characters.

I’m not sure about r-fishhook glyphs (U+027E and U+027F) and some others (U+01A6, U+01AC).

The Bold font is completely missing most of these glyphs, some are not bolded, but the ones that are also correctly have no serifs.

When a \tilde is used on a character in the math font, it is aligned too far to the left, appearing unbalanced. The MWE below illustrates the problem.

\documentclass{minimal}
\usepackage{fontspec}
\usepackage{unicode-math}
\setmathfont{Libertinus Math}

\begin{document}
$\tilde{d}$

\setmathfont{texgyrepagella-math}
$\tilde{d}$
\end{document}

Some TeX users are requesting slanted fonts in keeping with Computer Modern tradition of having both italic and slanted. Not sure of mechanical slanting can cut it, but lets track it here.

Hi Khaled,

Thanks for all your efforts!

I usually use the libertine package for my LaTeX documents and it's pretty great. Do you think it would make sense to either adapt some of the features of that package or, maybe better, ask Bob Tennent if he would be willing to add support for libertinus (in xetex/`luatex)?

Thanks,
Rasmus

Does this include serif small caps and sans serif small caps like the libertine package in latex? I can't seem to find them but I could just be dense. Thanks!

If they're not included, then that's a feature request.

I have found two problems tha seem specific to Libertinus when used in LuaLaTeX in math mode :

• Calligraphic letters do no appear
• Inconsistent choice of variable-sized delimiters.

A minimal set of example is in the enclosed archive
LibertinusPbms.zip. (Sorry for this, the Github Thought police accepts neither pdf nor tex files...).

These problems make the Libertinus family currently unsuitable as a drop-in replacement of Latin Modern.

Seems to be missing. Is there any way to make them in bulk? (0 experience editing fonts here)

In working up #46, I realized there is another issue with the +salt set: none of the special shapes are available in the Bold or Bold Italic variants.

Most notably here the descender on the 'h' is missing:

\begin[papersize=a5]{document}
\set[parameter=document.parindent,value=0]
\set[parameter=document.parskip,value=1em]
\font[family=Libertinus Serif,size=16pt]

\font[features="+salt,+ss02"]

\font[]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[style=Italic]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[weight=800]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[style=Italic,weight=800]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}
\end{document}

I accidentally found a people reported that Ó̧ (U+00D3 U+0327) is misplaced in NotoSans-Regular. When I copied the string, I found both Libertinus Sans and Serif has this bug.

Several fonts such as Arial, Segoe UI, Times New Roman, DejaVu fonts, Charis SIL and Doulos SIL can do this, but Libertinus font cannot.

Here is a test for localized form for Cyrillic:


In this case б should be transformed in all faces, other letters should also has italic variant in Libertinus Serif Italic, Libertinus Serif Bold Italic and so on. BTW I found Bosnian sometimes uses Serbian Cyrillic in Bosnia and Herzegovina, is it possible to make it available for Bosnian?

The space between small caps P and A is too large. The space is slightly narrower with the libertine package.

\documentclass{article}
%% With lualatex
\usepackage{fontspec}
\setmainfont{Libertine Serif}[SmallCapsFeatures = {Renderer = Basic, Letters = SmallCaps}]
%%

%% With latex
%\usepackage{libertine}
%%

\begin{document}
\textsc{ipa}
\end{document}

I was skimming a PDF today and noticed what at first I thought was an extra space that had bad it's way into the source inside a word. Unfortunately it turned out the source was clean, the problem was the kerning was inconsistent enough as to be misleading. Here's a sample of what I was seeing:

To my eye the last ’İ in the section heading is so far set off as to be almost separate. Obviously the L prior to it is largely to blame, but on inspection I noted what was going on earlier:

I'm not an expert, but it looks like the kerning pairs are not nearly as refined in the Small Caps as they are in the normal Serif series.

Besides my the L’ pair creating a huge gap (red) it seems like some places are squished (yellow) and others perhaps too open (blue). Or maybe those are fine and the relative weight of the R is throwing me off? I'm out of my depth now, all I know is it doesn't look right.

I just ran into a rather obscure pair of missing glyphs. I'm typesetting a very large book and considering a couple different font options. One of them is Libertinus Serif. My thought was to use the base style set for the body and the the italic version of stylistic alternate set 2 for one particular class of quotes.

In a perfect storm of heavy handed emphasis it just so happened the author has an emphasized bit inside one of these blocks that's also starting a sentence and thus capitalized. And starts with an 'r'. And has the name of a king in the middle. Did you follow that? In summary:

Libertinus Serif → Italic → Bold → SS02 → Capital → letters 'R' and 'K'.

These glyphs didn't come out bold.

Here is a MWE for Sile:

\begin[papersize=a5]{document}
\set[parameter=document.parindent,value=0]
\set[parameter=document.parskip,value=1em]
\font[family=Libertinus Serif,size=16pt]

\font[]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[style=Italic]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[weight=800]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[style=Italic,weight=800]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}

\font[features="+ss02"]

\font[]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[style=Italic]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[weight=800]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}\break
\font[style=Italic,weight=800]{AÂBCÇDEFGĞHIİJKLMNOÖPQRSŞTUÜÛVWXYZ}

\font[]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[style=Italic]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[weight=800]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}\break
\font[style=Italic,weight=800]{aâbcçdefgğhiijklmnoöpqrsştuüûvwxyz}
\end{document}

Note in the middle paragraph two glyphs are not like the others:

Calligraphic characters (the \mathcal range) don't appear to be supported in Libertine Math:

\documentclass[11pt]{article}
\usepackage{fontspec}
\usepackage{amsmath}
\usepackage{amsthm}
\usepackage{amssymb}

\usepackage{unicode-math}

\setmainfont[Numbers=OldStyle,Ligatures=TeX]{Libertine Serif}
\newfontfamily{\Lpnum}[Numbers=Lining,Ligatures=TeX]{Libertine Serif}
\setsansfont[Ligatures=TeX]{Libertine Sans}

\setmathfont{Libertine Math}

\begin{document}
${\cal A B C D E F G H I J K L M N O P Q R S T U V W X Y Z}$ ${\cal H}_i$
\end{document}

produces, with both LuaTeX and XeTeX,

XITS Math has a full range I can use with \setmathfont[range={\mathcal,\mathbfcal}]{xits-math.otf}, but it'd be nice if Libertine Math had its own set (perhaps borrowed from XITS or elsewhere).

Thanks for the otherwise lovely fonts.

Is there any plan to support this character T̈ ẗ like in DejaVu Serif?

Also is there bold and italic mode for Libertine Mono?

Thanks.

Here are a list of glyphs missing from the math font that can be copied from the existing Libertinus Mono font:

• Math monospace digits 0–9 (U+1D7F6 through U+1D7FF, \mttzero through \mttnine).
• Monospace uppercase Latin A–Z (U+1D670 through U+1D689, \mttA through \mttZ)
• Monospace lowercase Latin a–z (U+1D68A through U+1D6A3, \mtta through \mttz)

(These glyphs are low priority, but also low-hanging fruit.)

I'm testing version 6.1 of Libertinus and noticed some problems with the small caps feature.
As you can see on the screenshot, the Qu ligature is missing on the italic and bold versions while in the bold-italic the "fi" does not change into small caps, remaining as a lower-case ligature.
I'm using XeTeX (trough LyX)

The gap beneath the big Sigma in \sum is a bit cramped if the bottom limit has a symbol with cap height. If the top limit has a descender, it also seems to be cramped from above. MWE and illustration follows.

%!TEX program = lualatex
%
\documentclass{article}
\usepackage{unicode-math}

\begin{document}
\[
  \sum_{k=1}^{N-1}\;\sum_{A=1}^{g-1}\;\sum_{H=1}^{p-1}
\]
\setmathfont{Libertinus Math}
\[
  \sum_{k=1}^{N-1}\;\sum_{A=1}^{g-1}\;\sum_{H=1}^{p-1}
\]
\end{document}

Here are my versions: (let me know if I missed something important)

Libertinus, Latest git
LuaTeX, Version beta-0.80.0 (TeX Live 2015/Debian) (rev 5238)
luaotfload.sty 2015/12/09 v2.6 OpenType layout system
lualatex-math.sty 2015/09/22 v1.5 Patches for mathematics typesetting with Lu
unicode-math.sty 2015/09/24 v0.8c Unicode maths in XeLaTeX and LuaLaTeX
fontspec.sty 2016/02/01 v2.5a Font selection for XeLaTeX and LuaLaTeX

Thanks again. I'm willing to help, but I have no experience with font editing to speak of. Let me know if there is anything I can do.

(This is actually a bug in Linux Biolinum which Libertine has inherited - I don't want to go to the trouble of registering on Sourceforge to report something this trivial, but the email address they provide bounced when I tried to report it like that. If you could push it upstream for me I'd be grateful.)

In Libertine Sans Regular and Libertine Sans Italic (and the corresponding Linux Biolinum faces) LATIN SMALL LETTER TURNED M (U+026F) has the wrong width, causing the glyph to overlap with its neighbour.

EDIT: after some googling, it seems that I confused things and the "Linux Biolinum G" is actually a graphite font that I got from here: http://www.numbertext.org/linux/ . So the issue is actually a feature request for graphite versions of the font, since LibreOffice only supports these.

Hi

I'm not sure what's causing this problem, but when I try to use extended opentype features in LibreOffice with the OTF files provided in this repository, it does not work.

As an example, see this screenshot:

The first line has the font "Linux Biolinum G:onum=1", the second line uses "Libertine Sans:onum=1". As you can see, for Linux Biolinum, oldstyle figures and thousand separation is working. For Libertine Sans, I get neither of these effects.

Is there anything I'm doing wrong, is LibreOffice broken, or are the OTF files broken in some way?

A bug inherited from Linux Libertine

\documentclass{article}
\usepackage{fontspec}
\setmainfont[Numbers=OldStyle]{Libertinus Serif}
\begin{document}
\textbf{\emph{6}}
\end{document}

The 6 look like a zero. I have corrected it on my fork of Linux Libertine, but it would be fun to correct it also in Libertinus.

Inside the PUA starting at E000 there are some slots not used for the same character. Have a look at Unicode index E00D and compare e.g. the italic serif version with the regular one.

Is this an intended behaviour?

The kerning between the characters 'f' and 'j' looks correct in sans regular, in that the kerning appears to be the same as for 'fi'. But the kerning is wrong for sans italic:

\documentclass{article}
\usepackage{fontspec}
\setmainfont{libertinesans-}[
    UprightFont = *regular,
    ItalicFont = *italic]
\begin{document}
fi \textit{fi}

fj \textit{fj}
\end{document}

Several romanizations of Chinese dialects has some diacritics with less-known usage for typographic designers.

1. Pe̍h-ōe-jī use O͘ for pronounciation [ɔ], use a short vertical line for tone 8.
2. Foochow, Hinghwa and Kienning Romanizations (Bàng-uâ-cê, Báⁿ-uā-ci̍, Gṳ̄̄ing-nǎing Lô̤-mǎ-cī) use diaeresis below to dictinct some vowels.
3. Kienning Colloquial Romanized use macron twice on a letter for tone 4.

I put the following code into WYMeditor integration example to test them, using Firefox 45.0.1:

<div style="font-family: Libertinus Serif; font-size:2em; line-height:1.1em">
<p>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i> <b>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i></b></p>
<p style="font-feature-settings: 'smcp' on, 'c2sc' on">O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i> <b>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i></b></p>
<p>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i> <b>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i></b></p>
<p style="font-feature-settings: 'smcp' on, 'c2sc' on">O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i> <b>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘ <i>O͘Ó͘Ò͘Ô͘Ō͘O̍͘o͘ó͘ò͘ô͘ō͘o̍͘</i></b></p>
<p>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍ <i>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍</i> <b>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍ <i>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍</i></b></p>
<p style="font-feature-settings: 'smcp' on, 'c2sc' on">A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍ <i>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍</i> <b>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍ <i>A̍E̍I̍O̍U̍a̍e̍i̍o̍u̍</i></b></p>
<p>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂ <i>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂</i> <b>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂ <i>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂</i></b></p>
<p style="font-feature-settings: 'smcp' on, 'c2sc' on">A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂ <i>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂</i> <b>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂ <i>A̤Ă̤Ā̤Á̤À̤Â̤E̤Ĕ̤Ē̤É̤È̤Ê̤O̤O̤Ŏ̤Ó̤Ò̤Ô̤ṲṲ̆Ṳ́Ṳ̀Ṳ̂a̤ă̤ā̤á̤à̤â̤e̤ĕ̤ē̤é̤è̤ê̤o̤o̤ŏ̤ó̤ò̤ô̤ṳṳ̆ṳ́ṳ̀ṳ̂</i></b></p>
<p>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄ <i>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄</i> <b>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄ <i>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄</i></b></p>
<p style="font-feature-settings: 'smcp' on, 'c2sc' on">Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄ <i>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄</i> <b>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄ <i>Ā̄Ā̄Ā̤̄Ā̤̄Ē̄Ē̄Ē̤̄Ē̤̄Ī̄Ī̄Ō̄Ō̄Ō̤̄Ō̤̄Ū̄Ū̄Ṳ̄̄Ṳ̄̄ā̄ā̄ā̤̄ā̤̄ē̄ē̄ē̤̄ē̤̄ī̄ī̄ō̄ō̄ō̤̄ō̤̄ū̄ū̄ṳ̄̄ṳ̄̄</i></b></p>
</div>

Then I click “Preview” button and “print” via PDFCreator, get the following result:
Preview.zip
Seen from above file, I found some of them were failed in Libertinus Serif font, please make enhancements for them.

The glyphs for \ulcorner (⌜ = unicode 0x0231c) and \urcorner (⌝ = 0x0231d) are missing from Libertinus Math. MWE (for XeLaTeX):

\documentclass[12pt]{article}
\usepackage{amssymb,amsmath}
\usepackage{mathspec}
\defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase}
\usepackage{unicode-math}
\setmainfont{Linux Libertine O}
\setmathfont{Libertinus Math}
% \setmathfont[range={"231C,"231D}]{Asana Math}
\begin{document}
Simple test for ulcorner and urcorner:
\[
  T \ulcorner \forall x (Fx) \urcorner \vDash F \ulcorner\exists x (\lnot Fx) \urcorner
\]
\end{document}

These glyphs are present in Asana Math, so for now one can use the \setmathfont[range=...]{} workaround which is commented out in the above example.