127

This is with ref to my previous question Package clash in multilingual report.

\documentclass[11pt,table,a4paper]{article}
\usepackage{lmodern}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}

\usepackage{CJKutf8}
\usepackage[english,russian]{babel}

\newenvironment{SChinese}{%
 \CJKfamily{gbsn}%
 \CJKtilde
 \CJKnospace}{}

 \begin{document}
 \selectlanguage{russian}
  Это мой первый многоязычный докладе.
  Инфантильный гипертрофический стеноз привратника - это серьёзное 
 \selectlanguage{english}
  This is my first multilingual report.

 \begin{CJK}{UTF8}{}
  \begin{SChinese}
    这是我的第一个多语种的报告。
  \end{SChinese}
  \end{CJK}

  \end{document}

when I try to compile it, I get following error message.

LaTeX Warning: Unused global option(s):
    [table].

(./data.aux
(/usr/local/texlive/2011/texmf-dist/tex/latex/cyrillic/t2acmr.fd))
(/usr/local/texlive/2011/texmf-dist/tex/latex/lm/t1lmr.fd)

LaTeX Font Warning: Font shape `T2A/lmr/m/n' undefined (Font)         
using `T2A/cmr/m/n' instead on input line 15.


! Package inputenc Error: Unicode char \u8:  not set up for use with
LaTeX.

See the inputenc package documentation for explanation. Type  H
<return>  for immediate help.  ...                                    

l.18 ...�ный гипертрофический стеноз привра...

How can I avoid such error message.

Improve this question
9
  • 6
    Try loading fontenc with \usepackage[T1,T2A]{fontenc}. The T2A option is needed for Russian letters. Also note that lmodern doesn't load any cyrillic characters, hence the name "Latin Modern" (if you didn't already know this).
    – user2473
    Nov 20, 2012 at 4:01
  • 3
    I used \usepackage[T1,T2A]{fontenc} still same error message.
    – Manish
    Nov 20, 2012 at 4:04
  • 1
    Your example compiles fine on my machine. I see you're using TeX Live 2011. Something seems to be acting funny with your fonts. I suggest upgrading to an updated version of TeX Live 2012. In the meantime, perhaps removing \usepackage{lmodern} might help?
    – user2473
    Nov 20, 2012 at 4:07
  • I installed Tex Live 2012 and remove lmodern message but i m still getting same error message.
    – Manish
    Nov 20, 2012 at 6:11
  • 4
    The problem seems to be in the "no-break space" character. Add \DeclareUnicodeCharacter{00A0}{~} to your preamble.
    – egreg
    Nov 20, 2012 at 7:36

13 Answers 13

Reset to default
96

The error you get is due to a "no-break space" character, according to what I can gather by copying an pasting your message.

This character is not usually set up by the [utf8] option and it's invisible to many editors, so it can slip in a document without the typist knowing it.

Solution: add in your preamble

\DeclareUnicodeCharacter{00A0}{ }

if you don't mean to type a no-break space, or

\DeclareUnicodeCharacter{00A0}{~}

if you want that the character stands for what its name says.

UPDATE

Recent (after 2015-01-01) versions of the UTF8 configuration file for inputenc do define U+00A0 as \nobreakspace, so this should be of no concern, now.

Improve this answer
11
  • Sorry, I was wrong in my last comment. The fffe I got was put there by the command I used to see the encoding of the pasted text. Anyway, I cannot get the 00A0 either, so I hold my question. How did you came to it?
    – JLDiaz
    Mar 18, 2013 at 11:36
  • 1
    @JLDiaz Copy paste the error message; the conversion of the char \u8:  not bit gives the following sequence of Unicode points: 0063 0068 0061 0072 0020 005C 0075 0038 003A 00A0 0020 006E 006F 0074, where you clearly see 00A0.
    – egreg
    Mar 18, 2013 at 11:36
  • @egreg Thanks. That's what I did, but I got 0020 instead. I guess that my browser was "too smart" when copying characters to the clipboard. I pasted then at several Unicode converters online, and also in a terminal and used xxd, but I got always 0020 for the 00A0 char.
    – JLDiaz
    Mar 18, 2013 at 12:29
  • 3
    @nuttyaboutnatty You can use \newunicodechar{^^a0}{~} after loading \usepackage{newunicodechar}
    – egreg
    Jul 8, 2013 at 20:51
  • 1
    @Nicolas You don't load inputenc with LuaLaTeX
    – egreg
    May 18, 2016 at 16:52
67

As this is one of the top google hits for this error message, here's a more general answer with an example:

The cause is a unicode character in one of your input files that isn't mapped to an output. This may -- especially if you're using the (unicode-supporting) biblatex/biber system -- be in your bibliography. This is a good place to look for errors as .bib files downloaded from publishers website are often malformed. You can tell if the error comes from the bib file - the line number in the error message will be that of your \end{document}, which makes tracking down the actual error rather tricky (inspecting various aux files doesn't appear to help).

Some of these errors are subtle, like the non-breaking space in the question, or the hyphen (U+2010) character given to me by one journal, which looks identical to the hyphen-minus produced by the keyboard.

Copying the character after the hyphen and searching for it should help - unless your command window or editor "helps" by converting it to the more common equivalent or replacing unicode with blanks - in that case copy it from your .log and search all the input files.

(I'm happy to expand this in response to comments or watch it grow -- it's just an attempt to be helpful to searchers)

Improve this answer
12
  • 1
    I often mistakenly type this character in ViM on Mac OS X by hitting ALT+SHIFT+SPACE, often after writing {\some-command ...}. Thank you for the explanation!
    – csl
    Apr 14, 2014 at 18:15
  • 1
    It's a good strategy. It's often needed to clean the output files after fixing the problem.
    – user2821
    May 13, 2015 at 12:49
  • That is one hell of a subtle error. I indeed had a hyphen (U+2010) instead of the hyphen-minus. Thanks.
    – Roald
    Dec 6, 2016 at 16:31
  • How one may generally view what kind character is causing problems? In my case I somehow got U+FEFF 'ZERO WIDTH NO-BREAK SPACE' (hunted down by modyfying file in git repository and using git diff - but is there a more general solution)?
    – reducing activity
    Dec 22, 2016 at 7:44
  • 1
    Agreed, this is a general case as it could be many different unicode characters causing the issue. In a recent case it was &lowast; which interestingly is an HTML entity but for some reason needs to be declared as unicode &#8727.
    – jeffmcneill
    Oct 8, 2019 at 13:20
29

\usepackage[utf8x]{inputenc}

Ubuntu:

You must install texlive-latex-extra before use it.

Fedora:

You must install texlive-collection-latexextra before use it.

Improve this answer
9
3

You may get this error also if you use different language for bibtex. In that case project.bbl may contain characters in different encoding (e.g latin2).

What you need to do is swap encoding when rendering bibliography to latin2 and switch back to utf8 after.

\inputencoding{latin2}
\bibliography{mybib}
\inputencoding{utf8}

Hope this helps.

Improve this answer
2
  • Welcome to TeX.SX! This is an answer to an (old) question. You should elaborate a little bit on your post and show a fully working example
    – user31729
    Nov 23, 2014 at 20:06
  • I installed texlive-lang-czechslovak package to use slovak translations of entries in bibtex, but was getting errors while compiling project in utf8. I found out that it is caused by different encoding of .bbl file (evidently czechslovak package uses latin2 for .bbl files by default, bc. before using it I had no problems).
    – zub0r
    Nov 25, 2014 at 0:11
3

This happened to me when I did save my .tex file with utf8 but forgot to save also the .bib file with the same encoding (it was still in ANSI).

Instead of returning back to ANSI on my .tex file I just opened the .bib file with Notepad++ and chose to convert to utf8.

Then after compiling everything was working OK.

Improve this answer
3

I have found the same problem but none of the above answers solved it. In the end, I found the code \'{\i} in my .bib file. This was supposed to yield í but was producing a crazy unicode char that broke compilation. This .bib file was exported from CiteULike based on a reference that I entered mannually or copy-&-pasted from somewhere else. I suppose something wrong happened while converting a mannually entered/pasted í to \'{i}.

Improve this answer
3

I had two similar problems:

  1. "Unicode char \u8: "
  2. "Unicode char \u8:." (with dot)

The problems were related to the .bib file (references list).

The first problem was solved based on the \DeclareUnicodeCharacter{00A0}{ } stated by @egreg.

The second one was helped by the @Chris H 's answer. I opened the generated file .log and looked for errors. I found:

! Package inputenc Error: Unicode char \u8:\C3. not set up for use with LaTeX.

Then, I looked the "\C3" string in the generated file .bbl and I found out that the letter "Ó" (the first letter of an author's name, Óscar Oballe-Peinado) was the problem. So, I changed it in the bibliography file for {\'{O}} and voilà!

Despite I'm using "\usepackage[utf8]{inputenc}", it seems not working specifically with the accented first letter of the first name of authors.

Improve this answer
2
2

If this happens in the bibliography, try specifying the language explicitly to bibtextu:

bibtextu -l ru my_paper_with_russian_bibliography

This fixed it for me.

Improve this answer
2

I had this Error because I accidentally saved an included .tex file as ANSI while the master file was in UTF-8.

You can change file encoding in Notepad++ for example. But you will need to copy it from the the ANSI version and paste it into the UTF-8 version.

Improve this answer
2

XeTeX is more suitable than most other TeX engines for unicode : replace the lines

\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}

with

\usepackage{fontspec}

and compile with xelatex myfile.tex.

Improve this answer
2
  • 1
    You mean “XeTeX is more suitable than pdfTeX”, right? Because XeLaTeX is still LaTeX…
    – cgnieder
    Jul 4, 2017 at 10:39
  • @clemens Thanks for the correction. My bad, I had latex in mind. pdfTeX would be too narrow though. I hope the current formulation is better.
    – Skippy le Grand Gourou
    Jul 4, 2017 at 10:48
1

Inserting

% !TEX encoding = UTF-8 Unicode

at the beginning of the file solved this issue for me.

Improve this answer
1
  • didn't work for my case, where I wanted to render an unicode accent character from this author's name Rui M´ximo
    – Micah Stubbs
    Sep 17, 2019 at 19:48
1

My error was related, but slightly different:

! Package inputenc Error: Unicode character ╩│ (U+02B3)

or

! Package inputenc Error: Unicode character ᵉ (U+1D49)  
(inputenc)                not set up for use with LaTeX.

I had these errors because of a bibliographic entries generated in French babel which typesets the edition field, e.g. 1ʳᵉ éd. or 3ᵉ éd. using a raised re or e, resulting in U+02B3 and U+1D49 which apparently are not valid, even though (so far) all the other accented characters, e.g., é in édition are valid...

Because of my complex pandoc settings from Markdown with Libertine font, I don't want to swap to another *Tex engine...

So, my band-aid fix was to use the math mode to raise these letters:

\DeclareUnicodeCharacter{1D49}{$^\text{e}$}
\DeclareUnicodeCharacter{02B3}{$^\text{r}$}

Also, if you're using pandoc, put them in your header-includes: part of the front matter.

Improve this answer
0

An anwser for a slightly different case:

! Package inputenc Error: Unicode character ° (U+B0)
(inputenc)                not set up for use with LaTeX.

Here, the answer is very simple:

\usepackage{textcomp}

This packages defines the \textdegree macro and also sets up ° to use it.

Improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.