14

I have the following entry in my bib-file which contains the accented character \u\i:

@article {Gur1966SpacesUniversalPlacement,
    author = {Gurari{\u{\i}}, V. I.},
     title = {Spaces of universal placement, isotropic spaces and a problem of {M}azur on rotations of {B}anach spaces},
   journal = {Sibirsk. Mat. \v{Z}.},
    volume = {7},
      year = {1966},
     pages = {1002--1013},
}

When I use BibTeX the correct entry is generated. Using Biber results in the following error message:

Undefined control sequence.
  Gurari{\ui 
                  }

It seems that the backslash in \u\i is removed. Is it possible to prevent Biber from doing this? Or more general, what is the proper way of using accented characters with Biber?

Christian
  • 1,012
  • Just use the correct unicode character. – Johannes_B Apr 12 '16 at 08:37
  • By the way, in an up to date version of biber, this should be resolved. – Johannes_B Apr 12 '16 at 08:43
  • @Johannes_B: I tried to use the unicode character "ĭ" but then I ran into troubles with fontenc. It seems that ĭ is not in T1. – Christian Apr 12 '16 at 08:54
  • ĭ works fine for me with \usepackage[T1]{fontenc}. Perhaps your inputenc setting and/or the encoding of the bib is faulty. – Ulrike Fischer Apr 12 '16 at 09:12
  • @UlrikeFischer Apparently, Biber changes \u{\i} into U+0131 U+0306 which of course confuses pdflatex. Here's what I see: Gurari<C4><B1><CC><86> – egreg Apr 12 '16 at 09:20
  • 1
    @egreg That \u{\i} doesn't work is not really a surprise, but the direct input ĭ should work if the (input) encodings are correct and correctly declared. – Ulrike Fischer Apr 12 '16 at 09:27
  • @Johannes_B Really? I still get errors if my .bib files contain accented characters in traditional input rather than their unicode equivalents. At least, I've certainly seen this very recently. (I tend to change the entries as I come across them when they cause problems, which only happens when they are actually cited. So I have a long-term supply of potential errors.) – cfr Apr 12 '16 at 09:36
  • @UlrikeFischer: I am using \usepackage[T1]{fontenc} and I checked that my files are utf-8 encoded but I am getting the error: Package inputenc Error: Unicode char \u8:ĭ not set up for use with LaTeX. when I compile with pdflatex. – Christian Apr 12 '16 at 11:21
  • @UlrikeFischer: Sorry I meant \usepackage[utf8]{inputenc} in the above comment. – Christian Apr 12 '16 at 11:23
  • @user93559 The error Package inputenc Error: Unicode char \u8:ĭ not set up means you have an older version of LaTeX. Update your TeX distribution. On the other hand, Gurar{\u{i}} should work; it's very unfortunate having to use different input, which makes code unportable. :-( – egreg Apr 12 '16 at 11:41
  • @egreg The problem with Gurari{\u{i}} is that biber seems to change it to Gurari\ui, i.e. it removes the second backslash and the parentheses. – Christian Apr 12 '16 at 11:44
  • @user93559 Update your TeX distribution – egreg Apr 12 '16 at 14:51

1 Answers1

13

There's a strange combination of factors that should be solved by different people.

  1. There's no predefined combination \u{i} in t1enc.def, so ĭ should be typed in as \u{\i}

  2. Biber transforms \u{i} into U+012D LATIN SMALL LETTER I WITH BREVE

  3. Biber transforms \u{\i} into U+0131 LATIN SMALL LETTER DOTLESS I U+0306 COMBINING BREVE

So the legal LaTeX input Gurari{\u{\i}} becomes a combination that makes sense to Unicode engines (XeTeX and LuaTeX) because of their normalization rules, but is pretty useless with pdflatex.

Solution: in the author field either type in

Gurar{\u{i}}

or

Gurariĭ

(the latter requires you load \usepackage[utf8]{inputenc}

egreg
  • 1,121,712
  • the problem goes beyond that - the second character (U+0131) does not have coverage on some very standard fonts, like for example, Times New Roman, so the following code, even under xelatex produces a square in the place of the accent: \documentclass{report} \usepackage[utf8]{inputenc} \RequirePackage{fontspec} \setmainfont{Times New Roman} \begin{document} Oleı̆nik. \end{document} – Paulo Ney Apr 18 '18 at 21:08
  • @PauloNey If a font doesn't have the necessary glyphs, it's not the fault of the maintainers of Biblatex. Please make a new question with the details. – egreg Apr 18 '18 at 21:18
  • @egreg You are my hero of the day! (+1 of course). – Henrik Schumacher Aug 14 '19 at 16:38