@UlrikeFischer @DavidCarlisle I'm wondering what the best solution is to fix this mparhack problem. As far as I can see the mparhack "fix" relies on messing with \hb@xt@. I think at this point I'm more inclined to remove the autoloading of mparhack, leave it as an open bug (as it is not really memoir related) and mention mparhack and the needed patch in the manual. I don't know how many users use twocolumn mode combined with marginpar.
@daleif I have the consent of the author to adapt the package and will upload a version with Frank 's fix. If you test files it would be nice if you could check if it works everywhere.
@daleif well in the long run, marginpar should be fixed not to required such a package. But mparhack works since 20 years, so it can't be a complete disaster to keep it a bit longer.
@daleif it's a hack (the clue is in the name) but I think we can safely make it work as before and not disturb memoir which would be a good thing for users (and politically:-) as @UlrikeFischer says longer term it would be good to have a fixed marginpar in the kernel (and that might be needed so you can do correct pdf tagging anyway) so you may want to hold off any memoir adjustments until then.
@DavidCarlisle not a problem. Luckily I only load mparhack if twocolumn class option is active. So we don't hit that many users.
BTW: note that \tag* issue with amsmath on github. Not quite why what the problem its, but I noticed it works just fine if one uses align instead of equatuion
Good morning everyone! Quick question relative to tex.stackexchange.com/a/595446/38080: is there a standard way to define a function in l3fp? As the original declare function in the question?
@JosephWright while at it, do I need to use parens around \x in fpeval? Like -(\x) to take into account the possibility that \x is negative, or it is taken care of?
@UlrikeFischer getting rid of the if page.lang == mr tests in various places and adding a new entry on the general translations file:
# optional list of localised numbers 0-20 (fr alphabetic just for testing)
navigational-elements_digits:
fr: "*,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t"
mr: "०,१,२,३,४,५,६,७,८,९,१०,११,१२,१३,१४,१५,१६,१७"
Is the online style special in biblatex? See tex.stackexchange.com/q/595478/3929, why is changing the formatting for author or title not affected for an online entry?
@UlrikeFischer if I change \DeclareFieldFormat[online]{author}{#1\addcomma} to \DeclareNameFormat[online]{author}{#1\addcomma} in the OP example I seem to get family= outputteed to the bibliography
@daleif well sure, name formats are bit more complicated than simple fields, as they need to define the format using the various name parts, instead of #1 you should use commands like \namepartfamily, but if the OP only wants to change a comma, it would probably better to adjust a delimiter or one of the bib macros.
@JosephWright I just killed my aws account so latexcgi.xyz is no more (well actually it can be resuscitated during the next 90 days) It would start to cost real money from next month.
@DavidCarlisle I'm not sure if it is new, but if one compiles a document with hyperref (and the default pdfpagemode UseOutlines) in latex online, then one get a rather agressive sidebar:
I meant support for a user to store a style. If someone wanted to replace `\num[color=blue,add-decimal-zero,minimum-integer-digits = 4]{1}` by `\num[style=mystyle]`, could they do it?
@UlrikeFischer In tagpdf, if I want to mess everything upimplement something creative by reassigning a bunch of nodes to a marked content sequence after they were created, do I have to do anything else beside reassigning their mc attributes?
@JosephWright yes I know that I can define a key like this, but could a user define it with \sisetup{mystyle/.style={...}} or with some other sensible interface similar to the pgfkeys one?
@UlrikeFischer Currently there is nothign built-in but it's doable
@UlrikeFischer Point of course is that pgfkeys is OOP from the get-go, so there is no difference between defining keys and setting them - we went for distinct steps so you need some form of indirection
@JosephWright I didn't doubt this. I only wondered if there is already some setup to copy. Imho it wouldn't be so good if one package defines "create-style" and the next "storestyle" and the third "new-style" ;-)
@UlrikeFischer Honestly I'd go for some separate interface to \keys_define:nn - it would be pretty easy to arrange
@UlrikeFischer Probably one for the team list or LaTeX-L
@UlrikeFischer Easiest would be \DefineStyle (generic) or \DefineSIStyle (dedicated), the latter taking two arguments: \DefineSIStyle{mystyle}{...}, which would then pass to \keys_define:nn { siunitx } { #1 .meta:n = {#2} }
@UlrikeFischer Like I said, it's not hard in general; if we provided a generic document/design 'set stuff' interface, we could have a matching 'provide style' one