Refactor mergedate settings and authoryear citations

[moewew]

Cf. #520 (see also #148).

Currently mergedate is still slightly broken and the format between citations and bibliography differs in year labels. This refactor brings bibliography and citations in line by using the same commands and fixes most of mergedate's fixable behaviour.

We simply use \printlabeldateextra in citations, this greatly reduces clutter.

[moewew]

Note that this is quite a large change. In my tests there were some differences in output compared to v3.7, but all of these were desirable as far as I can tell (and as far as I am concerned). I assume the test suite will also report different output, so that needs careful inspection.

[plk]

Just ran the test suite and am checking output now.

[plk]

First thing to note is the changed bibtex output where there seems to be repeated labelyear (from 01-introduction.tex):

[moewew]

Ah, OK I will have to have a look at BibTeX. The tests are based on labeldatesource which is much safer than just comparing labelyear and year with Biber, but I didn't think about BibTeX.

[moewew]

If you can list all the other problems that arise I will have a look at those as well.

[plk]

The only other difference I see for the biber tests is that entries with month/day in the .bib always have this now in the bibliography - I'm not sure about this:

Even for normal entries which are not online-style variants:

[moewew]

This is my interpretation of what mergedate should do now. Previously entries like this would only have the month in the bibliography with the default mergedate=compact (this is the documented behaviour of mergedate in 50-style-authoryear), but since we allow for days now we might as well include them.

The BibTeX problem could be solved by adding labeldatesource to the .bbl. I am currently trying to modify the .bst to do so. But I am quite clueless with .bst files, so if you have an idea...

[plk]

Ok, yes, I think you are right. In that case, the biber test files look good - great work, many thanks. @josephwright - you are the bibtex expert here - can you help? We really would like to fix this soon so we can release.

[moewew]

My suggested changes to biblatex.bst are in #606 now. It seems to work alright, but I would really appreciate someone with experience with BST files having a look.

[plk]

Tests look good, I have committed the test changes and uploaded a new 3.8. Seems we are close to being able to release? I think it's fine to put in a static label[end]year value for bibtex.

[moewew]

Yes, but there are a few open issues left.

• Sorting of braced (corporate) authors with alphabetic style and BibTeX #601 would be nice to have fixed before a release.
• I didn't get any further with Why is origlanguage a field while language is a list? #594, so unless you have an idea that would have to be deferred.
• Write operation prevents font kerning #598, \DeclareFieldFormat and spaces between arguments #586, \ifnumeral chokes on a single Cyrillic letter #559 I'm clueless about and they are only 'small things', but can be really annoying when noticed.
• Detect footnotes with beamer #581 could hopefully be solved easily by someone who knows about beamer.
• There is also beamer and \newrefsection: entry \blx@keyaliases@1 not found #544 with beamer interaction.
• Regarding the date issue Datetime. Ranges: Unknowns, Opens, and Unspecifieds. #540 is still open, and maybe John finds some quirks in the solution to Datetime. Missing frills and datetime output, authoryear style, and mergedate. Bugs. #520 just committed.
• Not sure how we left defernumbers causes problem in very specific case #568.
• For weird punctuation problems there is Ellipsis (\dots) not being recognised as dot(s) with fontspec #514 (we probably need to change the spacefactor of the ellipsis here)
• and \uspunctuation, hyperref link quirk. #499.
• variant of @set #470 just crept up again.
• Also forcing hyperref commands even if hyperref is not detected #585 was more complicated than I thought - mainly conceptually.

Of course that is just the list of issues that caught my eye when scrolling down the tracker.
#587 and #518 can easily be deferred.

[plk]

@moewew, @josephwright - can we clean up the last few things soon - we really should release I think - it's a bit overdue.

[moewew]

I think I could live with some of these things being deferred to the next version.
I started working on #585 and #514 and have at least a first attempt working. In #585 I would appreciate your feedback and ideally also Ulrike's response. In #514 I probably need Joseph to have a look.