@JosephWright Do you have a second for an l3build question?
I think we're currently having issues with some spurious build failures because the current setup only copies over biblatex.sty into the test directory and does not also take all the other files like biblatex.def and bbx/standard.bbx with it.
We recently had a commit with changes in biblatex.def and these changes are not applied in tests.
I tried to see which bit of build.lua is responsible for this (I thought maybe installfiles), but I couldn't get l3build to copy them over.
@JosephWright Should something like installfiles = {"./tex/latex/biblatex/*"} work to get all files in that subdirectory?
What about files that live in a subdirectory of ./tex/latex/biblatex/, i.e. ./tex/latex/biblatex/bbx/, can we get them as well (flattened, i.e. not in their subdirectory)?
@PauloCereda yes, font choice matters. I almost lost my trip to NYC because the font in the passport number of my daughter does not show a big difference between O and 0, and I compiled the ESTA wrongly.
Yep. Happened to several Spaniards this summer; a whole set of passport sporting PAO as the first letters that lot of people (/me included) got as PA0.
@Rmano Well, they were designed to be human-readable and machine-readable … quite a while ago. And a QR code would cover much area without adding information any officer could scan by eye. With machine readers extracting digital information from passports being ubiquitous, one could rethink it. But changing international standards is hard.
@TeXnician Yes, I suppose that's the reason. Although the human-readable part should really avoid this kind of mistake. (I tried to compile the ESTA sending the passport image, but that failed --- I should have suspected the problem. But alas... at the end the new ESTA arrived, although just 5 minutes before the flight closing).
@AlanMunn I know. But there have been alterations (in fact, only additions iirc) to international requirements for passports after that. Still, it's not global standard to have electronic passport checks, so it would be hard to sacrifice space that's human-readable in favor of QR codes even today.
@PauloCereda You have been in the right countries then. In the 2010s I was travelling in southern Asia and there were some (smaller) airports where my passport was literally copied manually by pen into some sheet. The officers took ages to let passengers through.
@samcarter Me too. I hate whiteboards with a passion. Luckily our building also houses the math and stats departments, so all the teaching classrooms have blackboards, although the custodial staff are terrible about maintaining the chalk supply.
@AlanMunn -- Oh! Didn't have those in my day. Thick felt sheets glued sidewise to a backing of thick cardboard. Then one would take the erasers outside and pound them on the sidewalk to clean out the dust.
@AlanMunn A new professor arrived some time ago and -- imagine -- he wanted to replace the blackboards in our offices with whiteboards! After quite the outcry, a compromise was found that each office could decide if the wanted to keep the blackboard or get a new whiteboard.
Today's note about math fonts: The hat ("2C6) has in some fonts (lucida, latin modern) a very wide bounding box. The base glyph in widehat ("302) is sometimes very different from the hat glyph (pagella). There is also a third hat ("5E) that I have no idea of how (if) it should be used, and it looks weird in most fonts. Screenshot follows.
(I do not see the big bounding box in latin modern in lualatex, though)
It seems that the "2C6 is not used for \hat in unicode-math.
@DavidCarlisle OK, so "2C6 is not used at all? Do you know anything about the usage of "5E?
@DavidCarlisle After today's discussion, it looks like there will be something similar for ConTeXt, but perhaps that the variants from \widehat will move to \hat as well.
@mickep well it's COMBINING CIRCUMFLEX so not supposed to have an advance width itself just combine with previous, what do other combining accents do in opentype, I haven't looked recently?
@mickep we'll probably write something up for mathml, although that is markup so usage is different <mover><mi>x</mi><mo>^</mo></mover> the mover is responsible for making it small and accent position so a full size ^ is natural, U+0302 has the danger of combining with the markup and putting an accent on the > of <mo> (which of course is just visual effect not affecting the xml parsing, but looks weird) of course internally a renderer might use x U+0302 to typeset the mover...
@DavidCarlisle I'm interested in reading that, if/when available. Hans is also writing something up, but on unicode math, and one section is about accents.