It would be awesome if biber could also produce HTML output, similar to bibtex2html.

In lib/Biber/Config.pm, 'use File::slurp;' should be inserted before calling File::slurp::read_file later

Sometimes biber will report an error like the following and then stop working:

data source /var/folders/m6/bn7r45zx6cx55rr9g4qh6s6w0000gn/T/par-agoldst/cache-87e533530c5239fb9f8e3ff008979f1f16ea0e5e//inc/lib/Biber/LaTeX/recode_data.xml not found in .

I don't know when this happens and therefore can't give a reproducible test case, but this happens to others as well (see http://humtex.wordpress.com/2011/11/29/biber-first-aid-for-data-source-not-found/ for example).

The solution is to delete the par-directory. Afterwards everything works fine again. Since this error message is rather cryptic and the remedy seems to be straightforward I was wondering whether this could be improved.

1 down vote favorite

I have a tex file that was working perfectly with several bibliography parts with different sorting options. It is no longer working! Some bibliography parts are not displayed. My OS is Windows 7 Pro 64 bit. I am using:

• TexnicCenter v2.02 stable 64 bit
• Miktex 2.9 (with biblatex updated to v2.9a)
• Biber v1.9 for biblatex v2.9

Here is a minimal example which does not work:

\documentclass[a4paper]{article}

\usepackage{filecontents}

\begin{filecontents}{refs1.bib}
@book{bbb,
title = {bbb},
publisher = {bbb},
author = {BBB, Bbb},
year = {2012}
}
@book{ccc,
title = {ccc},
publisher = {ccc},
author = {CCC, Ccc},
year = {2014}
}
\end{filecontents}

\begin{filecontents}{refs2.bib}
@book{aaa,
title = {aaa},
publisher = {aaa},
author = {AAA, Aaa},
year = {2013}
}
\end{filecontents}

\usepackage[
style=authoryear,
firstinits=true,
backend=biber]{biblatex}

\addbibresource{refs1.bib}
\addbibresource{refs2.bib}

\DeclareBibliographyCategory{ref1_cat}
\DeclareBibliographyCategory{ref2_cat}

\addtocategory{ref1_cat}{bbb,ccc}
\addtocategory{ref2_cat}{aaa}

\begin{document}

Main text. See refs \parencite{ccc,aaa,bbb}.

\printbibliography[title={Bibliography},sorting=nyt,heading=bibintoc]
\nocite{*}

\appendix

\printbibliography[title={Only ref 1 (nyt)},category=ref1_cat,sorting=nyt,heading=bibnumbered]
\nocite{*}

\printbibliography[title={Only ref 1 (ydnt)},category=ref1_cat,sorting=ydnt,heading=bibnumbered]
\nocite{*}

\end{document}

The output is the following:

The bibliography with ydnt sorting scheme is not displayed. Since such kind of examples was working fine before, I am suspecting an incompatibility between biber v1.9 for windows and biblatex which I has recently updated to v2.9a (I guess the previous version was v2.9).

When I compile the tex code, I have the following warning:
Package biblatex Warning: File 'error_biber1.9_bilatex2.9a.bbl' is wrong format version - expected 2.4.
The format of the generated bbl file is v2.3.

I am also including the generated blg file:

[0] Config.pm:318> INFO - This is Biber 1.8
[0] Config.pm:321> INFO - Logfile is 'error_biber1.9_bilatex2.9a.blg'
[81] biber-MSWIN:272> INFO - === %a %b %e, %Y, %H:%M:%S
[82] Biber.pm:333> INFO - Reading 'error_biber1.9_bilatex2.9a.bcf'
[163] Utils.pm:169> WARN - Warning: Found biblatex control file version 2.6, expected version 2.5
[163] Biber.pm:607> INFO - Using all citekeys in bib section 0
[197] Biber.pm:3053> INFO - Processing section 0
[230] Biber.pm:3190> INFO - Looking for bibtex format file 'refs1.bib' for section 0
[233] bibtex.pm:812> INFO - Found BibTeX data source 'refs1.bib'
[238] Biber.pm:3190> INFO - Looking for bibtex format file 'refs2.bib' for section 0
[2169] bibtex.pm:812> INFO - Found BibTeX data source 'refs2.bib'
[2253] Biber.pm:2939> INFO - Overriding locale 'English_United States.1252' default tailoring 'variable = shifted' with 'variable = non-ignorable'
[2254] Biber.pm:2945> INFO - Sorting 'entry' list '' keys
[2254] Biber.pm:2949> INFO - No sort tailoring available for locale 'English_United States.1252'
[2275] Biber.pm:2939> INFO - Overriding locale 'English_United States.1252' default tailoring 'variable = shifted' with 'variable = non-ignorable'
[2275] Biber.pm:2945> INFO - Sorting 'entry' list 'nyt' keys
[2275] Biber.pm:2949> INFO - No sort tailoring available for locale 'English_United States.1252'
[2283] bbl.pm:482> INFO - Writing 'error_biber1.9_bilatex2.9a.bbl' with encoding 'ascii'
[2284] bbl.pm:555> INFO - Output to error_biber1.9_bilatex2.9a.bbl
[2285] Biber.pm:105> INFO - WARNINGS: 1

Is there any solution to this problem? I wanted to get back to the previous version (prior to v2.9a) of the biblatex package under Miktex but could not find how to do that (this is for sure kind of a dumb solution ...).

Does this issue come from the use of biber v1.9 with biblatex v2.9a?

I am trying to convert a biblatex list of reference into a list of bibitems following these instructions:
http://tex.stackexchange.com/questions/154215/biblatex-biber-to-bibitem

It worked flawlessly until I updated biblatex to 2.9 via mixtex, and to biber 1.9. Now the command "biber --output_format=bibtex main" returns an error:
Undefined subroutine &Biber::Output::bibtex::NFC called at C:/.../inc/lib/Biber/Output/bibtex.pm line 129

Running the command with biber 1.8 works.

MWE:

\documentclass{article}
\usepackage[backend=biber,sorting=none]{biblatex}

\begin{filecontents}{references.bib}
@Article{citation1,
author = {Author1},
title = {Title1},
journal = {Journal1},
year = {2008},
volume = {26},
pages = {316-324},
number = {6}
}

@Article{citation2,
author = {Author2},
title = {Title2},
journal = {Journal2},
year = {2007},
volume = {46},
pages = {4589-4597},
number = {15}
}

@Article{citation3,
author = {Author3},
title = {Title3},
journal = {Journal3},
year = {2005},
volume = {280},
pages = {30392-30399},
number = {34}
}
\end{filecontents}

\addbibresource{references.bib}

\begin{document}
Citation 1: cite author2 \cite{citation2}

Citation 2: cite author3 \cite{citation3}

Citation 3: cite author1 \cite{citation1}

\printbibliography
%
\end{document}

Name this file main.tex and run the below script (Windows) which is a .bat file I wrote based on the previous link to create a dummy.bib with the list of refs in bibitems:

set LATEX_PATH=C:\Program Files\MiKTeX 2.9\miktex\bin\x64
set TEX_FILE_NAME=main

"%LATEX_PATH%\biber" --output_format=bibtex "%TEX_FILE_NAME%.bcf"
echo \documentclass{article} > dummy.tex
echo \begin{document} >> dummy.tex
echo \nocite{*} >> dummy.tex
echo \bibliography{%TEX_FILE_NAME%_biber} >> dummy.tex
echo \bibliographystyle{unsrt} >> dummy.tex
echo \end{document} >> dummy.tex
"%LATEX_PATH%\pdflatex" dummy
"%LATEX_PATH%\bibtex" dummy

Compiling the attached tex/bib files with texlive 2013 and up to date biblatex biber results in a SEGV from biber.

Obviously the bib file is malformed, but I think biber should not segfault nevertheless.

\documentclass{scrartcl}

\usepackage{polyglossia}
\setdefaultlanguage{german}
\usepackage{csquotes}
\usepackage[hidelinks]{hyperref}

\usepackage[backend=biber]{biblatex}
\addbibresource{libbtparse.bib}

\begin{document}
\nocite{*}
\printbibliography[heading=none]
\end{document}

@comment test

@article{abcd,
title ="test"}

BLG file

[1] Config.pm:318> INFO - This is Biber 1.8
[1] Config.pm:321> INFO - Logfile is 'address-book.blg'
[75] biber:275> INFO - === Mon Jan  6, 2014, 23:17:28
[76] Biber.pm:333> INFO - Reading 'address-book.bcf'
[130] Biber.pm:607> INFO - Using all citekeys in bib section 0
[131] Biber.pm:675> DEBUG - Adding 'entry' list 'nty' for section 0
[225] Biber.pm:3053> INFO - Processing section 0
[225] Biber.pm:3174> DEBUG - Looking for directly cited keys: 
[236] Biber.pm:3190> INFO - Looking for bibtex format file 'libbtparse.bib' for section 0
[236] bibtex.pm:215> DEBUG - Caching data for BibTeX format file 'libbtparse.bib' for section 0
[236] bibtex.pm:937> INFO - Decoding LaTeX character macros into UTF-8
[236] bibtex.pm:812> INFO - Found BibTeX data source 'libbtparse.bib'
[237] bibtex.pm:93> FATAL - Caught signal: SEGV
Likely your .bib has a very bad entry which causes libbtparse to crash: Inappropriate ioctl for device

Miktex users with a 64bit miktex needs to download biber separatly. When they are looking for biber they often download the biber-cygwin32 version as its link is placed at the top of the sourceforge site and http://sourceforge.net/project/platformdownload.php?group_id=228270 actually directly starts the download of this version.

But this version seems not work with miktex: they get complains about a cygwin.dll missing. The version they actually need is either the one from 32-bit miktex or this
http://sourceforge.net/projects/biblatex-biber/files/biblatex-biber/1.9/binaries/Windows/biber-MSWIN.zip/download

Could the websites and the links be changed so that miktex/windows users no longer run into this trap?

Ulrike Fischer

If I test a bib-file against the datamodel with unknown keys:

@Report{IRS,
  title         = {title},
  author        = {author},
  institution   = {blub},
  blub={bla}
}

Then biber 2.0. stops like in issue 30. biber 1.9 works fine:

 G:\Z-Test>biberexp --tool -V bib-problem.bib
 INFO - This is Biber 2.0 (beta) running in TOOL mode
 INFO - Logfile is 'bib-problem.bib.blg'
 INFO - Looking for bibtex format file 'bib-problem.bib'
 INFO - Decoding LaTeX character macros into UTF-8
 INFO - Found BibTeX data source 'bib-problem.bib'

 G:\Z-Test>biberdev --tool -V bib-problem.bib
 INFO - This is Biber 1.9 (beta) running in TOOL mode
 INFO - Logfile is 'bib-problem.bib.blg'
 INFO - Looking for bibtex format file 'bib-problem.bib'
 INFO - Decoding LaTeX character macros into UTF-8
 INFO - Found BibTeX data source 'bib-problem.bib'
 WARN - Datamodel: Entry 'IRS' (bib-problem.bib): Field 'blub' invalid in data mo
 del - ignoring
 WARN - Datamodel: Entry 'IRS' (bib-problem.bib): Missing mandatory field - one o
 f 'date, year' must be defined
 WARN - Datamodel: Entry 'IRS' (bib-problem.bib): Missing mandatory field 'type'
 INFO - Overriding locale 'en_US' defaults 'variable = shifted' with 'variable =
 non-ignorable'
 INFO - Overriding locale 'en_US' defaults 'normalization = NFD' with 'normalizat
 ion = prenormalized'
 INFO - Sorting list 'none' of type 'entry' with scheme 'none' and locale 'en_US'

 INFO - No sort tailoring available for locale 'en_US'
 INFO - Writing 'bib-problem_bibertool.bib' with encoding 'UTF-8'
 INFO - Output to bib-problem_bibertool.bib
 INFO - WARNINGS: 3

The pdflatex command have the option '-include-directory=DIR'. It would be nice if the biber command could have the same option, in case of multiple .bib files arranged in any directory.

For some reason, biber seems to have propblems with the letter Å (danish a with ring, \AA in LaTeX).

When I try to cite the following entry biber will not go further than the message INFO - Found BibTeX data source …

@book{Madsen.S:1984a,
Address = {Frankfurt a. M.},
Author = {Madsen, Svend Åge},
Publisher = {Suhrkamp},
Subtitle = {Utopischer Roman},
Title = {Dem Tag entgegen},
Year = {1984}}

As soon as I change the Å to a regular A, the entry is handled without a problem.

This happens both with pdflatex (with \usepackage[utf8]{inputenc}) and with LuaLaTeX.

I seem to have installed version 1.8 via TeXLive (the manual is for 1.8), but biber -v says version 1.7, and biblatex 2.8 complains about wrong bbl version.
I checked and the binary on CTAN also reports version 1.7.

I don't know how and by who the binaries are created, that is why I report this issue here.

I wanted to try out the experimental biber/biblatex.

Testing a simple document like this:

\documentclass{article}
\usepackage{biblatex}
\addbibresource{biblatex-examples.bib}
\begin{document}
\cite{angenendt}
\printbibliography
\end{document}

and then running biber leads to this error message

[0] Config.pm:320> INFO - This is Biber 2.0 (beta)
[1] Config.pm:323> INFO - Logfile is 'test-bib.blg'
[76] biber-MSWIN:273> INFO - === %a %b %e, %Y, %H:%M:%S
[77] Biber.pm:340> INFO - Reading 'test-bib.bcf'
[206] Biber.pm:638> INFO - Found 1 citekeys in bib section 0
[273] Biber.pm:3125> INFO - Processing section 0
[329] Utils.pm:188> ERROR - Error loading data source package    'Biber::Input::file::bibtex': Can't use an undefined value as an ARRAY reference at C:  \Users\Nililand\AppData\Local\Temp\par-4e696c696c616e64\cache- f1acf52d1b94bfe170a7dc1159874b7a400e7ebf\inc\lib/Biber/Input/file/bibtex.pm line  595.
Compilation failed in require at (eval 135) line 2.
[330] Biber.pm:110> INFO - ERRORS: 1

Ulrike Fischer

I have been playing around with merging and combining of bib-files.

The following examples works perfectly:

 \documentclass[12pt,a4paper]{scrreprt}
 \usepackage[ngerman]{babel}
 \usepackage[T1]{fontenc}
 \usepackage[utf8]{inputenc}
 \usepackage[backend=biber]{biblatex}
 \addbibresource{bib.bib}
 \addbibresource{biblatex-examples.bib}
 \begin{document}
 \cite{test}
 \cite{angenendt}
 %\nocite{*} abc
 \end{document}

When I compile it and then run
biber --output_format=bibtex test-utf8.bcf
I get a bib-file which contains the two bib-entries from the two bib-files.

But when I add \nocite{*} the resulting bib is empty. ;-(. (Tested with 1.8 and 2.0)

Btw: biber --help claims that for --output_format the options

"bibtex" and "biblatexml" are only supported in (and only make sense in) "tool" mode (see --tool option)".

I don't understand this. bibtex seems to work fine without --tool. Also I don't see how I can use the control file of a document in --tool-mode.

Biber 1.8 Solaris/Intel/64 fails with missing bbl.pm.

ERROR - Error loading data source package 'Biber::Output::bbl': Can't locate Biber/Output/bbl.pm in @inc (@inc contains: /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc/lib /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc CODE(0x128f208) CODE(0x128f760)) at (eval 132) line 2.

Biber executable (downloaded a few hours ago from sourceforge) md5 sum: 69d525a2f7d02410ed0fdd9b21f45b4e

Solaris version: SunOS 5.11 oi_151a7 i86pc i386 i86pc

Removing the par- directory, and running biber again does not solve the problem.
I executed the command with truss, and it seems like it never tries to create the files:

truss biber …

grep CREAT /tmp/log | grep par | grep Output

open("/var/tmp//par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc/lib/Biber/Output/base.pm", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 6
open("/var/tmp//par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc/lib/Biber/Output/dot.pm", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 6

biber wiklund_system_thinking
INFO - This is Biber 1.8
INFO - Logfile is 'wiklund_system_thinking.blg'
INFO - Reading 'wiklund_system_thinking.bcf'
INFO - Found 33 citekeys in bib section 0
ERROR - Error loading data source package 'Biber::Output::bbl': Can't locate Biber/Output/bbl.pm in @inc (@inc contains: /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc/lib /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc CODE(0x128f208) CODE(0x128f760)) at (eval 132) line 2.

INFO - ERRORS: 1

rm -rf /var/tmp/par-6a656e6b696e73/

-bash-4.0$ biber wiklund_system_thinking
INFO - This is Biber 1.8
INFO - Logfile is 'wiklund_system_thinking.blg'
INFO - Reading 'wiklund_system_thinking.bcf'
INFO - Found 33 citekeys in bib section 0
ERROR - Error loading data source package 'Biber::Output::bbl': Can't locate Biber/Output/bbl.pm in @inc (@inc contains: /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc/lib /var/tmp/par-6a656e6b696e73/cache-306dbebacaac9b761bf326b67d43a0a81875ea3f/inc CODE(0x128f328) CODE(0x128f880)) at (eval 132) line 2.

INFO - ERRORS: 1

Hi

I am trying to update the Debian package to 1.8, but now the package does not build anymore, due to a crash in the tests:

main::(t/biblatexml.t:12):      chdir("t/tdata");
  DB<1> r
XPathContext: cannot unregister namespace
 at /tmp/biber-1.8/blib/lib/Biber/Input/file/biblatexml.pm line 159.
        Biber::Input::file::biblatexml::extract_entries('biblatexml.bltxml', 'ARRAY(0x3767e90)') called at /tmp/biber-1.8/blib/lib/Biber.pm line 3193
        Biber::fetch_data('Biber=HASH(0x1735ca8)') called at /tmp/biber-1.8/blib/lib/Biber.pm line 3058
        Biber::prepare('Biber=HASH(0x1735ca8)') called at t/biblatexml.t line 39
# Looks like your test exited with 255 before it could output anything.

The line numbers are slightly off since we patch the mozilla::ca completely out and pass the path to the system certificate file, but this is the only things that might be relevant.

Do you have any idea about that?

Thanks

Norbert

When the author list of two entries is different but longer than maxnames and the first author is the same, something like "Author et al. [1] and Author et al. [2]" is produced which looks fairly weird. So in my mind, \textcite{authoretal1,authoretal2} should be merged to "Author et al. [1,2]" even though the co-authors are different. There should at least be an option for this. When the author list is completely identical, the output is correctly merged BTW.

Moved over from https://sourceforge.net/p/biblatex/bugs/178/

Biber will crash if there are any entries with an extra comma in
the author field. For example,

author = {Sagan, , Carl},

causes Biber to segfault with

FATAL - Caught signal: SEGV
Likely your .bib has a very bad entry which causes libbtparse to crash: Inappropriate ioctl for device

and it must be corrected to

author = {Sagan, Carl},

bibtex-mode in Emacs reports that the construct is syntactically valid though.

Using recent git commit d9d65aa and also tested with 1.9 (+1 commit), biber seems to stop without a clear error message. I'm not a heavy biber user (this is my only document I've used it for) so I'm not much use a debugging this!

Anyway, the output:

biber report
INFO - This is Biber 2.0 (beta)
INFO - Logfile is 'report.blg'
INFO - Reading 'report.bcf'
WARN - Warning: Found biblatex control file version 2.5, expected version 2.6
INFO - Found 50 citekeys in bib section 0
INFO - Found 8 citekeys in bib section 1
INFO - Found 0 citekeys in bib section 0
Use of uninitialized value $name in lc at /usr/share/perl5/Biber/SortList.pm line 87.
Use of uninitialized value $lssn in concatenation (.) or string at /usr/share/perl5/Biber.pm line 682.
Use of uninitialized value $lname in concatenation (.) or string at /usr/share/perl5/Biber.pm line 682.
Use of uninitialized value $name in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value $ssn in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value $lname in concatenation (.) or string at /usr/share/perl5/Biber.pm line 661.
Use of uninitialized value $lssn in concatenation (.) or string at /usr/share/perl5/Biber.pm line 661.
Use of uninitialized value $name in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value $ssn in string eq at /usr/share/perl5/Biber/SortLists.pm line 77.
Use of uninitialized value $name in lc at /usr/share/perl5/Biber/SortList.pm line 87.
Use of uninitialized value $lssn in concatenation (.) or string at /usr/share/perl5/Biber.pm line 682.
Use of uninitialized value $lname in concatenation (.) or string at /usr/share/perl5/Biber.pm line 682.
INFO - Processing section 0
INFO - Looking for bibtex format file 'cbm-pub.bib' for section 0
INFO - Decoding LaTeX character macros into UTF-8
INFO - Found BibTeX data source '/home/cbm/work/texinputs/cbm-pub.bib'
make: *** [all] Error 255

I can't tell what is the error message here, is it the biblatex warning?

Here's the output from biber 1.8 (which works fine)

biber report
INFO - This is Biber 1.8
INFO - Logfile is 'report.blg'
INFO - Reading 'report.bcf'
INFO - Found 50 citekeys in bib section 0
INFO - Found 8 citekeys in bib section 1
INFO - Found 0 citekeys in bib section 0
INFO - Processing section 0
INFO - Looking for bibtex format file 'cbm-pub.bib' for section 0
INFO - Found BibTeX data source '/home/cbm/work/texinputs/cbm-pub.bib'
INFO - Looking for bibtex format file 'cbm-students.bib' for section 0
INFO - Found BibTeX data source '/home/cbm/work/texinputs/cbm-students.bib'
INFO - Looking for bibtex format file 'cbm-talks.bib' for section 0
INFO - Found BibTeX data source '/home/cbm/work/texinputs/cbm-talks.bib'
INFO - Overriding locale 'en_US.UTF-8' default tailoring 'variable = shifted' with 'variable = non-ignorable'
INFO - Sorting 'entry' list 'none' keys
INFO - No sort tailoring available for locale 'en_US.UTF-8'
INFO - Processing section 1
INFO - Looking for bibtex format file 'cbm-pub.bib' for section 1
INFO - Looking for bibtex format file 'cbm-students.bib' for section 1
INFO - Overriding locale 'en_US.UTF-8' default tailoring 'variable = shifted' with 'variable = non-ignorable'
INFO - Sorting 'entry' list 'none' keys
INFO - No sort tailoring available for locale 'en_US.UTF-8'
INFO - Writing 'report.bbl' with encoding 'ascii'
INFO - Output to report.bbl

The following minimum working examples describes the problem:

\documentclass{scrartcl}

\usepackage[utf8]{inputenc}

\usepackage[
style=ieee,
backend=biber,
]{biblatex}

\addbibresource{test.bib}

\usepackage{filecontents}
\begin{filecontents}{test.bib}
@INPROCEEDINGS{test,
author = {Mustermann, Max},
title = {{E}in {T}itel mit {Ö}ffnungen},
year = {2014},
month = apr
}
\end{filecontents}

\begin{document}

\nocite{*}
\printbibliography

\end{document}

The problem is that the letter 'Ö' is not being written as a capital letter in the pdf output file, although it is put into curly braces. I'm using biber 1.8 under MikTeX 2.9.

See also the (German) discussion about this problem at http://www.mrunix.de/forums/showthread.php?t=76277

The doc gives the wrong parameter for visual output: it says '--output_format=dot' instead of the correct '--outformat=dot' (the --help option gets it right).

If you use the fields related and relatedtype in an item, biber has problems grouping authors. A working example can be found here: http://tex.stackexchange.com/questions/128031/false-grouping-of-author-in-bibliography-with-biblatex

With the following document

 \documentclass{article}
 \usepackage[english]{babel}
 \usepackage[utf8]{inputenc}
 \usepackage[autostyle]{csquotes}
 \usepackage[vform=romanised]{biblatex}

 \addbibresource{problem.bib}

\begin{document}
\cite{testromanise} 

\printbibliography
\end{document}

and this bib

  @article{testromanise,
 author={Li, Wang},
  author_romanised={LiRom, Wang},
  title={title}
   }

I get a warning:

  [288] Entry.pm:498> WARN - Variant enabled datafield author/romanised/english in key 'testromanise' already has a value  - overwriting

The bbl looks ok.
With larger bibliographies one gets a lot of these warning when using vform=romanised.

While using Biber 1.7 on Linux (Gentoo, using the ebuild from the mrueg overlay) in tool mode, I get the following error:

$ biber -D -O refs_pre.bib -g blx2bib-pre.conf --tool refs.bib
INFO - This is Biber 1.7 running in TOOL mode
INFO - Config file is 'blx2bib-pre.conf'
INFO - Logfile is 'refs.bib.blg'
INFO - DEBUG mode: all messages are logged to 'refs.bib.blg'
INFO - Looking for bibtex format file 'refs.bib'
INFO - Decoding LaTeX character macros into UTF-8
INFO - Found BibTeX data source 'refs.bib'
Can't use an undefined value as an ARRAY reference at /usr/lib64/perl5/vendor_perl/5.16.3/Biber/Internals.pm line 940.

The minimal example files used are refs.bib:

@online{Fukuda-polyfaq,
  AUTHOR = {Komei Fukuda},
  DATE   = {2004-06-18},
  TITLE  = {Frequently asked questions in polyhedral computation},
  URL    = {http://www.ifor.math.ethz.ch/~fukuda/polyfaq},
}

and blx2bib-pre.conf:

<?xml version="1.0" encoding="UTF-8"?>
<config>
  <sourcemap>
    <maps datatype="bibtex" map_overwrite="1">
      <!-- transform DATE into YEAR and, possibly, MONTH; then drop DATE -->
      <map>
        <map_step map_field_source="DATE" map_final="1" />
        <map_step map_field_source="DATE" map_match="^(\d{4})" map_field_set="YEAR" map_field_value="$1" />
        <map_step map_field_source="DATE" map_match="^\d{4}-(\d{2})" map_field_set="MONTH" map_field_value="$1" />
        <map_step map_field_source="MONTH" map_match="01" map_replace="jan" />
        <map_step map_field_source="MONTH" map_match="02" map_replace="feb" />
        <map_step map_field_source="MONTH" map_match="03" map_replace="mar" />
        <map_step map_field_source="MONTH" map_match="04" map_replace="apr" />
        <map_step map_field_source="MONTH" map_match="05" map_replace="may" />
        <map_step map_field_source="MONTH" map_match="06" map_replace="jun" />
        <map_step map_field_source="MONTH" map_match="07" map_replace="jul" />
        <map_step map_field_source="MONTH" map_match="08" map_replace="aug" />
        <map_step map_field_source="MONTH" map_match="09" map_replace="sep" />
        <map_step map_field_source="MONTH" map_match="10" map_replace="oct" />
        <map_step map_field_source="MONTH" map_match="11" map_replace="nov" />
        <map_step map_field_source="MONTH" map_match="12" map_replace="dec" />
        <map_step map_field_set="DATE" map_null="1" />
      </map>
    </maps>
  </sourcemap>
</config>

Only an empty file refs_pre.bib is generated as output next to the logfile refs.bib.blg:

[0] Config.pm:425> INFO - This is Biber 1.7 running in TOOL mode
[0] Config.pm:427> INFO - Config file is 'blx2bib-pre.conf'
[0] Config.pm:428> INFO - Logfile is 'refs.bib.blg'
[59] biber:252> INFO - === wo jul 17, 2013, 12:59:47
[59] Biber.pm:255> DEBUG - Adding 'entry' list 'tool' for pseudo-section 0
[63] Biber.pm:3118> DEBUG - Looking for directly cited keys: 
[72] Biber.pm:3130> INFO - Looking for bibtex format file 'refs.bib'
[72] bibtex.pm:189> DEBUG - Caching data for BibTeX format file 'refs.bib' for section 0
[72] bibtex.pm:911> INFO - Decoding LaTeX character macros into UTF-8
[72] bibtex.pm:786> INFO - Found BibTeX data source 'refs.bib'
[73] bibtex.pm:870> DEBUG - Cached Text::BibTeX entry for key 'Fukuda-polyfaq' from BibTeX file 'refs.bib'
[73] bibtex.pm:197> DEBUG - All cached citekeys will be used for section '0'
[73] bibtex.pm:505> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Setting field 'year' to '2004'
[73] bibtex.pm:505> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Setting field 'month' to '06'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '01' -> 'jan' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '02' -> 'feb' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '03' -> 'mar' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '04' -> 'apr' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '05' -> 'may' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '06' -> 'jun' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '07' -> 'jul' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '08' -> 'aug' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '09' -> 'sep' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '10' -> 'oct' on field 'month'
[74] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '11' -> 'nov' on field 'month'
[75] bibtex.pm:414> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Doing match/replace '12' -> 'dec' on field 'month'
[75] bibtex.pm:462> DEBUG - Source mapping (type=user, key=Fukuda-polyfaq): Deleting field 'DATE'
[75] bibtex.pm:225> DEBUG - Added all citekeys to section '0': Fukuda-polyfaq
[75] Biber.pm:3141> DEBUG - Directly cited keys not found for section '0': 
[75] Biber.pm:1926> DEBUG - Populated 'entry' list 'tool' in section 0 with keys: Fukuda-polyfaq
[75] Biber.pm:1940> DEBUG - Checking sorting cache for list 'tool'
[75] Biber.pm:1955> DEBUG - No sorting cache entry for 'tool'

\documentclass{article}
\usepackage[utf8]{inputenc}

\usepackage{filecontents}

\begin{filecontents}{\jobname.bib}
@misc{ref,
  author = {Mart{\'\i}},
}
\end{filecontents}

\usepackage{biblatex}
\addbibresource{\jobname.bib}

\begin{document}
\nocite{ref}
\printbibliography
\end{document}

The above yields an error, because biber will put NFD UTF8 in the bbl file, and LaTeX can't handle that. Without the inputenc it will work, because then the bcf file won't say utf8. With [backend=bibtex] this works, so there are bib files that look like this.

biber.tex says that output aways will be NFC.

Here is an example:

% biber -D these-gc
INFO - This is Biber 1.9
INFO - Logfile is 'these-gc.blg'
INFO - DEBUG mode: all messages are logged to 'these-gc.blg'
INFO - Reading 'these-gc.bcf'
INFO - Using all citekeys in bib section 0
INFO - Processing section 0
INFO - Looking for bibtex format file './library.bib' for section 0
INFO - Decoding LaTeX character macros into UTF-8
INFO - Found BibTeX data source './library.bib'
FATAL - Caught signal: SEGV
Likely your .bib has a very bad entry which causes libbtparse to crash:
zsh: segmentation fault (core dumped) biber -D these-gc

If I try one more time (with the same files):

% biber -D these-gc
INFO - This is Biber 1.9
INFO - Logfile is 'these-gc.blg'
INFO - DEBUG mode: all messages are logged to 'these-gc.blg'
INFO - Reading 'these-gc.bcf'
INFO - Using all citekeys in bib section 0
INFO - Processing section 0
INFO - Looking for bibtex format file './library.bib' for section 0
INFO - Decoding LaTeX character macros into UTF-8
INFO - Found BibTeX data source './library.bib'
*** Error in `/usr/': free(): invalid size: 0x00000000063f8b30 ***
zsh: abort (core dumped) biber -D these-gc

How to debug this problem?

\documentclass{article}
\usepackage{biblatex}
\addbibresource{bib.bib}
\begin{document}
\cite{tilde}% 
\end{document}

with the bib entry

@article{tilde,
 author={J.~ Muster},
 title={test},
 }

gives a fatal biber error (2.9 and 4.0):

..
FATAL - Caught signal: SEGV
Likely your .bib has a very bad entry which causes libbtparse to crash:
Drücken Sie eine beliebige Taste . . .

The space after the ~ is important. And I know that the tilde is wrong here ;-)

This document

\documentclass{article}
\usepackage{biblatex}
\addbibresource{bib.bib}
\begin{document}
\cite{lowercaseI}, \cite{lowercaseII}
\printbibliography
\end{document}

and this bib-entries

@article{lowercaseI,
 author={lowercase,a.},
 title={test lowercaseI}
 }

@article{lowercaseII,
 author={{l}owercase,a.},
 title={test lowercaseII}
 }

gives the biber message

[316] Utils.pm:146> WARN - Couldn't determine Last Name for name "lowercase,a." - ignoring name
[324] Utils.pm:146> WARN - The field 'author' in entry 'lowercaseI' cannot be null, deleting it

and lowercaseI has no author. Protecting the author name with braces works but it is intentional? (with bibtex the name works fine).

(the real author in a large bib which triggered this problem writes her name intentionally in lowercase).

I think that the sortinit field in bbl files is only meant for the feature with \bibinitsep in biblatex.
In that case it should use some canonical form of the first letter, instead of just the first letter. Here is an example:

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage{filecontents}

\begin{filecontents}{\jobname.bib}
@misc{da, author = {DA}}
@misc{db, author = {DB}}
@misc{dc, author = {DC}}
@misc{dd, author = {DD}}
@misc{ea, author = {EA}}
@misc{eb, author = {ÉB}}
@misc{ec, author = {ÉC}}
@misc{ed, author = {ED}}
@misc{ee, author = {EE}}
@misc{f1, author = {f1}}
@misc{f2, author = {F2}}
\end{filecontents}

\usepackage[style=authoryear]{biblatex}
\addbibresource{\jobname.bib}

\setlength{\bibinitsep}{\baselineskip}

\begin{document}
\nocite{*}
\printbibliography
\end{document}

Here the Ds are in one group, but the Es are divided because some of them have an accent. The Fs are also divided because of case.
I expected gaps only before "EA" and "f1".

Fixing the accent case correctly would mean getting a canonical form of the first "sort unit" in the string (thinking about hard cases like traditional Spanish where for example "ch" is sorted as a separate "letter", although I hear that is not done like that anymore). I don't know if that is feasible. If not, just removing all accents would make it better, I think. (And I say that as a Swede, even though Swedish is one of the few languages that this is not good enough for, with our alphabet ending with XYZÅÄÖ so any Swedish list with Å and Ä in it will lack one gap.)

I would like to generate HTML from bib files, similarly to how BibTeX can generate HTML if given Bibhtml style files (http://mirror.klaus-uwe.me/ctan/biblio/bibtex/contrib/bibhtml/bibhtml.html). Is this already possible? If not, will it ever become possible?

This works fine:

\documentclass{article}
% \usepackage[utf8]{inputenc}

\usepackage{tipa}
\usepackage{filecontents}

\begin{filecontents}{\jobname.bib}
@misc{explanation,
  title = {\textipa{[""Ekspl@"neIS@n]}}
}
@misc{schwa,
  title = {\textipa{/f{\textschwa}f/}}
}
@misc{g,
  title = {\textipa{/g\textg/}}
}
\end{filecontents}

\usepackage{biblatex}
\addbibresource{\jobname.bib}

\begin{document}
\nocite{*}

\printbibliography
\end{document}

(Although maybe it's a questionable idea to have markup like that in titles.)

With inputencoding utf8 activated "explanation" will still work, since the text is copied as it is.

But "schwa", where longer input with commands for tipa, will yield an error, as the bbl file now will contain

      \field{labeltitle}{\textipa{/fəf/}}
      \field{title}{\textipa{/fəf/}}

Why is biber doing all this recoding into characters when using utf8? The small advantages are for cases when it would have been better to put the utf8 characters in the bib file to begin with anyway. Would there be any big disadvantages of copying what's in the bib file instead? The assumption would be that users should use utf8 in the bib file if they do in the tex file, and if they write special commands instead it's there for a reason. When is there an advantage of having textschwa mapped in recode_data.xml?

The last case doesn't yield an error, but gives erroneous output with utf8, since it becomes just \textipa{/gg/ in the bbl file. ("g" looks different inside \textipa{...}, and \textg is a special command then to get the normal g.)

Even if I'm wrong in all of this being a problem, I think \textg shows a small problem at least. The entry for \textg seems to be the reason that "g" is in encode_exclude in lib/Biber/LaTeX/recode_data.xml. That can't be right, can it? (Trying to understand that occurence of "g" was what led me to this.)

A bib entry like

@Book{ex,
  author =   {A. U. Thor},
  title =    {Title},
  year =     2014,
  pages =        {pp. 1--2}
}

with explicit (unnecessary) "pp." is converted into

     \field{sortinit}{T}
      \field{labeltitle}{Title}
      \field{title}{Title}
      \field{year}{2014}
      \field{pages}{pages}

in the bbl, so the pages field has the wrong contents. There are similar strange results when using other values that include "pp." or "p.", like

  pages =        {p. 17, pp. 617--716},

becoming

      \field{pages}{p. 17\bibrangessep p. 17}

It seems like there is some special handling of "pp." that isn't working.

(This comes from http://tex.stackexchange.com/questions/167803/page-numbers-in-bibliography-not-correctly-printed-if-list-contains-pp-12-34 )

Nick Ross noted this and I can confirm it using the biber 1.9 (technically 65e3cef) in Fedora 20.

# biber foo
INFO - This is Biber 1.9 (beta)
INFO - Logfile is 'foo.blg'
Undefined subroutine &File::Slurp::read_file called at /usr/share/perl5/Biber/LaTeX/Recode.pm line 112.
make: *** [all] Error 25

When I run with one of the development bibers (1.9/2.0)
biber --tool -V bib-problem.bib

on this bib-file:

@book{test,
author={A},
keywords={keyword},
}

Then biber doesn't output a bibertool.bib and the log-file stops in the middle:

[0] Config.pm:318> INFO - This is Biber 1.9 (beta) running in TOOL mode
[1] Config.pm:321> INFO - Logfile is 'bib-problem.bib.blg'
[59] biber-MSWIN:273> INFO - === %a %b %e, %Y, %H:%M:%S
[141] Biber.pm:3181> INFO - Looking for bibtex format file 'bib-problem.bib'
[146] bibtex.pm:957> INFO - Decoding LaTeX character macros into UTF-8
[148] bibtex.pm:831> INFO - Found BibTeX data source 'bib-problem.bib'

Removing or commenting the keywords line solves the problem:

[0] Config.pm:318> INFO - This is Biber 1.9 (beta) running in TOOL mode
[1] Config.pm:321> INFO - Logfile is 'bib-problem.bib.blg'
[61] biber-MSWIN:273> INFO - === %a %b %e, %Y, %H:%M:%S
[145] Biber.pm:3181> INFO - Looking for bibtex format file 'bib-problem.bib'
[150] bibtex.pm:957> INFO - Decoding LaTeX character macros into UTF-8
[152] bibtex.pm:831> INFO - Found BibTeX data source 'bib-problem.bib'
[158] Utils.pm:170> WARN - Datamodel: Entry 'test' (bib-problem.bib): Missing mandatory field - one of 'date, year' must be defined
[159] Utils.pm:170> WARN - Datamodel: Entry 'test' (bib-problem.bib): Missing mandatory field 'title'
[192] UCollate.pm:66> INFO - Overriding locale 'en_US' defaults 'normalization = NFD' with 'normalization = prenormalized'
[194] UCollate.pm:66> INFO - Overriding locale 'en_US' defaults 'variable = shifted' with 'variable = non-ignorable'
[195] Biber.pm:2924> INFO - Sorting list 'none' of type 'entry' with scheme 'none' and locale 'en_US'
[197] Biber.pm:2928> INFO - No sort tailoring available for locale 'en_US'
[199] bibtex.pm:166> INFO - Writing 'bib-problem_bibertool.bib' with encoding 'UTF-8'
[200] bibtex.pm:180> INFO - Output to bib-problem_bibertool.bib
[201] Biber.pm:106> INFO - WARNINGS: 2

This is a (low-priority) feature request, as the correct behavior may
differ from style to style. Would it be possible to give users an
option to stop Biber considering two such dates as equal for the
purposes of assigning them an extrayear? (Perhaps there's already an
option for doing so? I can't find it.)

Here's a MWE:

\documentclass[12pt]{article}
\usepackage{csquotes}
\usepackage[american]{babel}
\usepackage[style=authoryear]{biblatex}
\usepackage{filecontents}
\begin{filecontents}{\jobname.bib}
@Collection{harley:ancient:cart,
title = {Cartography in Prehistoric, Ancient, and Medieval
Europe and the Mediterranean},
maintitle = {The History of Cartography},
editor = {Harley, J.~B. and Woodward, David},
date = {1987},
publisher = {University of Chicago Press},
volume = 1}
@MVCollection{harley:hoc,
title = {The History of Cartography},
date = {1987/},
editor = {Harley, J.~B. and Woodward, David},
volumes = {3},
publisher = {University of Chicago Press},
location = {Chicago}}
\end{filecontents}
\addbibresource{\jobname.bib}

\begin{document}
\null\vfill
\nocite{*}
\printbibliography
\end{document}

And here's the output:

I wonder whether the following output isn't unambiguous enough for
some styles:

Thanks.

I am no expert so sorry for the rather verbose and not-useful 'blabla'
I had an issue when running bib files with biber:
When entering the keywords field in .bib file biber generates an empty bbl file.

I am using bib latex 2.9a with biber 1.9 on mac os texlive 2014 distribution.
(running the terminal commands xelatex/latex and biber)
(I got the same error when using TeXshop.app)

bib entry example:
@book{aaa,
title = {aaa},
publisher = {aaa},
author = {AAA, Aaa},
year = {2013},
keywords = {aa},
}

when I remove the keywords entry line everything runs smoothly. I found actually useful the keyword filtering ability and use it in many different occasions.

any ideas?

I switched from bibtex8 to biber and I found that my .bib abstracts had things like $'s and &'s in them. The % sign however caused huge errors and broke my document. I think this is quite strange because the only way that would happen is if biber demanded that my abstract was valid latex code. Somehow it is able to deal with URLs perfectly fine though.

I spent hours trying to trace down the problem and to have it be a problem in my bibliogrophy abstract was terrible. My solution is one that you already implement. Treat all fields exactly like you treat a URL, or at least just the abstract. If Biber requires special characters to be escaped, it has to be on Biber to do so. It is completely unreasonable to expect that input from a 3rd party conforms to your internal needs. It is up the application to take whatever text is given and make it conform to internal requirements for processing. If you need help with this let me know, but there is almost no documentation associated with Biber or Biblatex that Latex reserved characters are not handled correctly by this application.

Seen in build log with t/utils.t

[ 876s] Use of uninitialized value in concatenation (.) or string at /home/abuild/rpmbuild/BUILD/biblatex-biber-1.7/blib/lib/Biber/LaTeX/Recode.pm line 234.
[ 876s] Use of uninitialized value in concatenation (.) or string at /home/abuild/rpmbuild/BUILD/biblatex-biber-1.7/blib/lib/Biber/LaTeX/Recode.pm line 235.

also from 10 tests of t/tool-bltxml.t 9 do fail ... guess there is a racy condition

Hi all

At Debian we would very much like to package Biber, but unfortunately we have a problem with one of Biber's dependencies.

Biber depends on Data::Diver, but this Perl package comes with no copyright file. This means that distributions like Debian (and many others) cannot ship the Data::Diver Perl package.

So far the Data::Diver author has not responded to the SourceForge bug ticket.

Does Biber really need to depend on the Data::Diver package? If not, is it possible to exclude its use? This will allow Debian and many other distributions to include Biber in their repositories.

Thank you for your kind consideration.

Best regards

Danai Sae-Han

I invoke biblatex using e.g. mincrossrefs=50 to prevent the automatic addition of cross-referenced entries not otherwise cited. (I would prefer to have them added but abbreviate the references to them but I don't understand biblatex well enough to achieve that.)

If I've understood the manual's explanation of mincrossrefs correctly, this should prevent cross-referenced entries from being included as standalone entries unless referenced 50 times by child entries. However, this doesn't work. If, for example, I cite 2 entries which reference a third entry, the third entry is included separately in the bibliography. This would not be quite as bad if the child entries were formatted to just reference that entry (though it would still be a bug) but the information from the parent entry is also included in the formatting of the child entries. In the case of @inBook or @bookinbook, the results are especially awkward as you can easily end up with half a dozen entries which are very slight variations on a theme e.g. A in X, B in X, C in X... as well as X itself one after another in an author-year formatted bibliography.

Minimal example

test.tex:

\documentclass{article}
\usepackage{csquotes}
\usepackage[backend=biber,mincrossrefs=50]{biblatex}
\bibliography{test}
\title{}
\author{}
\begin{document}
\cite{anselm-m}
\cite{anselm-p}
\printbibliography
\end{document}

test.bib:

@bookinbook{anselm-m,
crossref = {anselm-cptt},
author = {Anselm},
title = {Monologion},
pages = {1--87}}
@bookinbook{anselm-p,
crossref = {anselm-cptt},
author = {Anselm},
title = {Proslogion},
pages = {88--112}}
@book{anselm-cptt,
author = {Anselm},
title = {Complete Philosophical and Theological Treatises of {A}nselm of {C}anterbury},
booktitle = {Complete Philosophical and Theological Treatises of {A}nselm of {C}anterbury},
address = minnea,
publisher = {The Arthur J. Banning Press},
year = 2000,
translator = {Jasper Hopkins and Herbert Richardson}}

Dynamic set entries do not appear in output when using all citekeys (at least with biblatex).

.bcf:

<!-- ... -->
<!-- CITATION DATA -->
<!-- SECTION 0 -->
<!-- ... -->
<bcf:section number="0">
  <bcf:citekey type="set" members="key1,key2">set1</bcf:citekey>
  <bcf:citekey order="0">*</bcf:citekey>
</bcf:section>
<!-- ... -->

A work around is possible by breaking it up using (order is important):

<bcf:section number="0">
  <bcf:citekey order="0">*</bcf:citekey>
</bcf:section>
<bcf:section number="0">
  <bcf:citekey type="set" members="key1,key2">set1</bcf:citekey>
</bcf:section>

Although this requires manual interaction for biblatex to use the workaround. Recreating this issue using biblatex is in this gist: https://gist.github.com/agrede/11076859

It seems that question marks in .bib entries trigger unnecessary warnings from biber when the input encoding of the .tex file is something else than UTF-8.

There is more information about this on TeX - LaTeX Stack Exchange: Bib entry cannot be encoded in ansinew. My observations / guesses are in the comments.

I noticed that the Swedish letter Å is recoded incorrectly when using biber --tool. I'm encoding from utf8 to ascii (i.e. TEX-macro) and got the following output:

\r{A}

when it should be

{\AA}

The same happens to both upper and lowercase Å.

Example bib file as input:

@Article{Test123,
title = {{Experimental Biber Recoding}},
author = {ÅÄÖåäö},
}

with the command
biber --tool --input_encoding=utf8 --output_encoding=ascii test.bib
outputs
@Article{Test123,
AUTHOR = {\r{A}"{A}"{O}\r{a}"{a}"{o}},
TITLE = {{Experimental Biber Recoding}},
}

but should be

@Article{Test123,
AUTHOR = {{\AA}"{A}"{O}{\aa}"{a}"{o}},
TITLE = {{Experimental Biber Recoding}},
}

The letters äöÄÖ work fine. Biber 1.9

Biber v1.8 (TeXLive 2013) generates this bibliography from the two files below:

Antonuccibad, R. R. (1985). “Spectropolarimetry and the nature of NGC 1068”. In: ApJ 297, pp. 621–632.
Antonuccigood, R. R. J. (1985). “Spectropolarimetry and the nature of NGC 1068”. In: ApJ 297, pp. 621–632.

One initial is missing here as well as in mwe.bbl. I couldn't get the GitHub version of Biber to work, but my guess based on the source is that the offending line is

line 1027: lib/Biber/Input/file/bibtex.pm

$nd_namestr =~ s/(\w)\.~(\w)/$1. $2/g;

It cannot change both ties to spaces. I'd appreciate it greatly if someone who understands the source (and Perl) better could look into this.

mwe.tex

\documentclass{article}
\RequirePackage[style=authoryear, backend=biber]{biblatex}
\ExecuteBibliographyOptions{firstinits=true, uniquename=allinit}
\addbibresource{mwe.bib}
\begin{document}
\cite{1985ApJ...297..621A-good}
\cite{1985ApJ...297..621A-bad}
\printbibliography
\end{document}

mwe.bib

@article{1985ApJ...297..621A-good,
  author        = "Antonuccigood, R. R. J.",
  title         = "Spectropolarimetry and the nature of NGC 1068",
  year          = 1985,
  month         = 10,
  journal       = "ApJ",
  volume        = 297,
  pages         = "621--632",
}
@article{1985ApJ...297..621A-bad,
  author        = "Antonuccibad, R.~R.~J.",
  title         = "Spectropolarimetry and the nature of NGC 1068",
  year          = 1985,
  month         = 10,
  journal       = "ApJ",
  volume        = 297,
  pages         = "621--632",
}

Unlike bibtex8, biber doesn't uppercase the sortinit field. This might make sense with sortcase enabled, but given the definition of \blx@initsep I doubt that PL envisioned this field as case-sensitive.

Related TeX.SE comment:

http://tex.stackexchange.com/questions/150068/how-to-split-the-bibliography-alphabetically/150100#comment341854_150100

Using biblatex 2.7a and biber 1.7, a call to \DeclareLabeldate in the
preamble fails if the call includes optional entry types within square
brackets. Any entry of the same type as that included in the
declaration will lack a labelyear field in the .bbl file, and biber -D
reports:

[1681] Biber.pm:1610> DEBUG - labeldate information of entry
james:ambassadors is unset

I also see several lines like this in the output when biber runs:

Use of uninitialized value in string eq at /tmp/.../inc/lib/Biber.pm
line 1552.

Use of uninitialized value in string eq at /tmp/.../inc/lib/Biber.pm
line 1572.

If you use the obsolete \DeclareLabelyear instead, then you do get a
labelyear in the .bbl file, but it's not the correct one, but rather
the one provided by the global definition of \DeclareLabeldate in
biblatex.def instead of the one from the declaration in the document preamble.

\DeclareLabeldate works fine so long as you don't specify entry types
in square brackets, and \DeclareLabeltitle seems to work fine both
with and without optional entry types.

I'm not certain when this began, perhaps with the switch to
\DeclareLabeldate from \DeclareLabelyear. It all worked properly in
biber 1.5. It may be that there is something peculiar to, and wrong
in, my setup?

I'll include a minimum working example below.

Thanks.

*** minexam.tex ***

\nonstopmode
\listfiles
\documentclass[a4paper,12pt]{article}
\usepackage[T1]{fontenc}
\usepackage{textcomp}
\usepackage[latin1]{inputenc}
\usepackage[german,french,american]{babel}
\usepackage[autostyle]{csquotes}
\usepackage[style=authoryear,backend=biber,alldates=comp,%
babel=other,bibencoding=ascii]{biblatex}
\usepackage{ifthen}
\usepackage{setspace}
\usepackage{vmargin} \setpapersize{A4}
\setmarginsrb{1in}{20pt}{1in}{.5in}{1pt}{2pt}{0pt}{2mm}
\usepackage{url}
\DeclareLabeltitle[book]{%
\field{origyear}
\field{title}
\field{shorttitle}}
\DeclareLabeldate[book]{%
\field{eventyear}
\field{origyear}
\field{year}
\field{urlyear}}
\appto\bibsetup{\sloppy}
\appto\citesetup{\sloppy}
\setlength{\dimen\footins}{9.5in}
\setlength{\parindent}{0pt}
\newcommand{\cmd}[1]{\texttt{\textbackslash #1}}
\bibliography{tiny}
\begin{document}

This citation of a \textbf{Book} entry with two dates shows
\cmd{DeclareLabeltitle[book]} working and \cmd{DeclareLabeldate[book]}
failing: \autocite{james:ambassadors} --- this being the origdate of
the entry, italicized as the labeltitle, rather than provided as the
labeldate, absent both here and in the reference list below.

\printbibliography
\end{document}
%%% Local Variables:
%%% mode: latex
%%% TeX-master: t
%%% End:

*** tiny.bib ***

@string{cup = {Cambridge University Press}}
@string{oup = {Oxford University Press}}
@string{uchp = {University of Chicago Press}}
@string{hup = {Harvard University Press}}
@string{pup = {Princeton University Press}}
@string{ucp = {University of California Press}}
@string{sup = {Stanford University Press}}

@book{james:ambassadors,
title = {The Ambassadors},
date = 1996,
origdate = 1909,
author = {James, Henry},
publisher = {Project Gutenberg},
url = {ftp://ibiblio.org/pub/documents/books/gutenberg/etext96/ambas10.txt},
annote = {Presents an online edition of a book which, not
being inherently an online text, still uses a Book
entry. The origdate field is the date of the print
publication of the text that is now online.}
}

minexam.bbl:

\refsection{0}
\sortlist{entry}{nyt}
\entry{james:ambassadors}{book}{}
\name{labelname}{1}{}{%
{{uniquename=0,hash=c478ac19353c3947855cd8192165811f}{James}{J\bibinitpe
riod}{Henry}{H\bibinitperiod}{}{}{}{}}%
}
\name{author}{1}{}{%
{{uniquename=0,hash=c478ac19353c3947855cd8192165811f}{James}{J\bibinitpe
riod}{Henry}{H\bibinitperiod}{}{}{}{}}%
}
\list{publisher}{1}{%
{Project Gutenberg}%
}
\strng{namehash}{c478ac19353c3947855cd8192165811f}
\strng{fullhash}{c478ac19353c3947855cd8192165811f}
\field{sortinit}{J}
\field{labeltitle}{1909}
\field{annotation}{Presents an online edition of a book which, not being i
nherently an online text, still uses a Book entry. The origdate field is the dat
e of the print publication of the text that is now online.}
\field{origyear}{1909}
\field{title}{The Ambassadors}
\field{year}{1996}
\verb{url}
\verb ftp://ibiblio.org/pub/documents/books/gutenberg/etext96/ambas10.txt
\endverb
\endentry
\endsortlist
\endrefsection

I am using biblatex 2.7a with biber 1.7. I have found biber writing nothing to bbl, and consequently failing to make any effect on the document (citations or bibliography). The console output is as follows:

INFO - This is Biber 1.7
INFO - Logfile is 'women_ordain.blg'
INFO - Reading 'women_ordain.bcf'
INFO - Found 17 citekeys in bib section 0
INFO - Processing section 0
INFO - Looking for bibtex format file 'pandita.bib' for section 0
INFO - Decoding LaTeX character macros into UTF-8
INFO - Found BibTeX data source '/home/pan/texlive/2013/../texmf-local/bibtex/bib/pandita.bib'
Can't use string ("usetranslator=true") as an ARRAY ref while "strict refs" in use at /tmp/par-70616e/cache-f37ab610b7d79b2720a8ee3732849c6821705520/inc/lib/Biber/Utils.pm line 755.
INFO - Overriding locale 'en_US.utf8' default tailoring 'variable = shifted' with 'variable = non-ignorable'
INFO - Sorting 'shorthand' list 'shorthand' keys
INFO - No sort tailoring available for locale 'en_US.utf8'
INFO - Overriding locale 'en_US.utf8' default tailoring 'variable = shifted' with 'variable = non-ignorable'
INFO - Sorting 'entry' list 'nty' keys
INFO - No sort tailoring available for locale 'en_US.utf8'

The bib entry containing "usetranslator=true" is as follows:
@book{horner:bd,
options = {usetranslator=true},
translator = {Horner, I.~B.},
title = {The Book of the Discipline {(Vinaya Piṭaka)}},
shorttitle = {Discipline},
year = {1938-66},
address = {Oxford},
publisher = {The Pali Text Society},
volumes = {6}
}

I built biber 2.0 (the development version on sourceforge) for MSWIN64 with strawberry perl 5.18.2.2 and when I try to test it on the test files from sourceforge I get the following error:

INFO - This is Biber 2.0 (beta)
INFO - Logfile is 'test.blg'
ERROR - 'test.bcf' failed to validate against schema 'D:\media\tmp\par-63616c76696e\cache-4af7dcd5d9fc8de20fab5d40baf993ac2125c4a9\inc\lib/Biber\bcf.rng'
INFO - ERRORS: 1

The log file has the following in it:

[0] Config.pm:322> INFO - This is Biber 2.0 (beta)
[2] Config.pm:325> INFO - Logfile is 'test.blg'
[43] biber-MSWIN64:274> INFO - === %a %b %e, %Y, %H:%M:%S
[53] Utils.pm:160> ERROR - 'test.bcf' failed to validate against schema 'D:\media\tmp\par-63616c76696e\cache-4af7dcd5d9fc8de20fab5d40baf993ac2125c4a9\inc\lib/Biber\bcf.rng'
[54] Biber.pm:110> INFO - ERRORS: 1

The same happens with the official binary build 1.9 from sourceforge.

Biber seems to run fine on my projects, so I'm not sure this is relevant but I thought I'd let you know.

This ticket is for tracking testing issues with the biber 2.0 multiscript branch.

biber appears to be substituting an incorrect character for the \- sequence from bib database files. This is a change of behaviour. I'm not sure what it substituted before, if anything, but entries containing this sequence did not cause errors in the past. However, I can't say exactly when the problem was introduced since it depends on using one of the affected entries.

If using pdflatex with the utf8 option of inputenc, this causes errors as the unicode character is not recognised:

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

That is, biber is introducing unrecognised characters into the bbl which cause subsequent compilation runs to fail, report errors or otherwise behave badly. (I get no PDF on one system but a PDF despite the errors on another.)

http://tex.stackexchange.com/q/194207/39222 includes details. Ulrike Fischer suggested filing this as a bug report.

The following example file can be used to reproduce the problem:

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[backend=biber]{biblatex}
\usepackage{filecontents}
\begin{filecontents}{\jobname.bib}
  @book{mybook,
    title       =   {Great Book},
    author      =   {Author, Some},
    year        =   1055,
    publisher   =   {Oxford University Press/Clar\-endon},
    address     =   {Oxford}}
\end{filecontents}

\bibliography{\jobname}

\begin{document}

  \autocite{mybook}

  \printbibliography

\end{document}

Note that I'm aware from egreg's comments that this isn't a good example since I should not be using \- here but there certainly are places one might want to use it in bibliography database files and biber ought not substitute something problematic by default in any case.

I attempt to build everything I use from source. For my sins, I have been trying to improve how beyond linuxfromscratch builds texlive. Using biber-1.8 (to match what appears to have been shipped in texlive-2014) works with UTF-8 input, but only if I restrict the "key" values to characters used in W. Europe such as accented vowels or the german sharp-s. If I use a letter such as c-caron (č) in the key, e.g.
@Article{majdič,
then biber fails with:
Cannot decode string with wide characters at /usr/lib/perl5/5.20.0/x86_64-linux/Encode.pm line 215.
(works ok with 'majdic').
Google found one match for this against biber, but no clear solution, as if that user had mis-keyed the text.
Note that all UTF-8 from the actual entries within the bib file displays correctly when I restrict the characters in the key.

It also found a lot of other matches in general, suggesting that the behaviour of Encode.pm had changed in the last few years.

If this is the wrong place to ask, advice on where to ask would be appreciated.