plk

It might be considered a bug I think. Dynamic sets are treated as citation keys which explains this ordering. Looking into it.

Hmm, the doc says "\defbibentryset implies the equivalent of a \nocite command, i. e., all sets which are declared are also added to the bibliography." This makes: ``` \defbibentryset{biyi2013}{Yi2014--,Yi2013--} \defbibentryset{bili2013}{Li2014--,Li2013--}...

It's a rather esoteric issue, `sortcites=true` changes it and so does any other sorting than `sorting=none`.

Is this driven by the "announcement" field in the CTAN upload? I never use this. I agree we should do more in this direction.

There are Unicode aware casing modules for Perl but not for title casing as this is not like other casings as it's partly semantic rather than purely syntactic. I doubt...

I wonder if we should rather look and see of CLDR has something that will work for all Unicode. Hacking this syntactically is probably just too hard. http://cldr.unicode.org/development/development-process/design-proposals/consistent-casing

This is basically impossible - the .bbl for biblatex contains custom macros defined in `biblatex.sty` and not generic TeX. So, the coupling is very tight and ignoring the version would...

@josephwright, @aboruvka - this is an interesting suggestion - `\@` is clearly correct here and perhaps should be default except that this does depend somewhat on the assumptions of the...

Can you please try this with the current 3.4 DEV version (and biber 2.5 DEV version if you use biber) - I believe this might be fixed.

@josephwright - do you think you will have time to look at this?