I have a UTF8 workflow: xelatex, using biblatex and biber for the bibliography, and I think I have come across a funny behaviour from biber. It seems to be converting silently to a percent-coding string (no trace in any log) a URL that contains one single non-ASCII character.

See the MWE with output. The URL contains one simple ü, which is being converted to %C3%BC. I would expect and wish the URL to be kept as it is. Can you spot any mistake? As you can see I've tried the biblatex options that affect biber's, but nothing seems to work, I still get the bloody %C3%BC.

Just in case, I manage my .bib file with BibDesk and I've checked it is saved as UTF8 too.

%!TEX TS-program = xelatex
%!TEX encoding = UTF-8 Unicode


Author = {{Warburg Institute}},
Title = {Exhibition: \emph{Verknüpfungszwang}},
Url = {http://warburg.sas.ac.uk/exhibition-verknüpfungszwang},
Urldate = {2016-12-12}}


Hello and bla bla bla bla bla bye


output of above MWE

I've checked that the .bib file contains the unicode alright, but it has been converted in the .bbl

  • 3
    With the newest biber (2.7) you can use the option ` --nouri-encode` Dec 15, 2016 at 16:58
  • Ulrike you're a star. Is there a way to pass this biber option from the tex file? Dec 15, 2016 at 17:08
  • I don't think so. The option is shiny new (and undocumented ...). But you could make a feature request. github.com/plk/biblatex/issues Dec 15, 2016 at 17:32


Browse other questions tagged .