stackoverflow.blog/2024/08/05/…

Lightweights

1986 - hah!

I use en.wikipedia.org/wiki/Thermal_ellipsoid routinely - released 1965 :)

@daleif Can I draw your attention to a small typo in the empheq documentation? In the second framed box on page 20 there's EmpheqMainEnv instead of EmphEqMainEnv.

@JosephWright I use lots of Newtonian mechanics. Released 1687 :-)

@campa not a computer application though. The key point here was which is the oldest unaltered software you still regularly use.

@JosephWright ^^ /shrug

but let me guess, it's diff?

@yo' you did peek, didn't you?

@Skillmon 40 year old tool that can't be improved? Well, Pluss-Knuth wouldn't be interesting enough for SO Blog...

@yo' I honestly didn't know how old diff was, I'd have thought it would be older, tbh.

@Skillmon I'm not surprised. It's the time when all these computer-theoretical things were bursting out.

Apparently my local git repo for mathtools/empheq hasn't been properly updated after the renaming of the master branch to main. So I get some errors on git pull/push.

Anyone remember what the proper way of makeing the renaming permanent on the local repo.

git branch -m master main
git fetch origin
git branch -u origin/main main
git remote set-head origin main

@daleif ^

@daleif I have the above as a shell script in my path and have run it on hundreds of repos by now all the ho-tex ones, and latex3 ones and my own, still keep finding the odd one that I'd missed

@DavidCarlisle I get the following on the third command

I keep getting this one:

git push
fatal: The upstream branch of your current branch does not match
the name of your current branch.  To push to the upstream branch
on the remote, use

    git push origin HEAD:master

To push to the branch of the same name on the remote, use

    git push origin HEAD

To choose either option permanently, see push.default in 'git help config'.

It should be in the main branch

@daleif oh let me check what branches are in empheq

@daleif is your local checkout based on the latex3 upstream (ie after it was moved theer) what does git remote -vvv report?

The error you show says the remote has a branch called master (as far as I can see) but that's not true of latex3/mathtools at guthub

@daleif oh actually I think I see. What happens if you just do git remote set-head origin main

@daleif Failing that, there is the obligatory refernce git solution for all such problems:

git remote -vvv
origin	[email protected]:latex3/mathtools.git (fetch)
origin	[email protected]:latex3/mathtools.git (push)

If I do a pull, I get this

git pull
Your configuration specifies to merge with the ref 'refs/heads/master'
from the remote, but no such ref was fetched.

@daleif git remote set-head origin main is supposed to tell it to use main

that is you are supposed to see

$ more .git/refs/remotes/origin/HEAD
ref: refs/remotes/origin/main

but simply re-cloning the thing would work, as xkcd rightly says

I get that like as well

@daleif sorry then I'm out of ideas, it works for me, I'd re-clone it then commit any local additions you have made since then

@DavidCarlisle I renamed the folder and re-cloned. @campas typos are now in the dtx

@daleif that worked, my local version now has, on main:

commit 8296a5e37d719ff6e7d298013b285d6f91e1e1b7 (HEAD -> main, origin/main, origin/HEAD)
Author: daleif <[email protected]>
Date:   Thu Aug 8 12:21:31 2024 +0200

    fixed typo reported in chat by campa

Say we have the following code in our preamble

\newcommand\AAA{%
  \parbox[b]{5cm}{\fontsize{10pt}{12pt}\selectfont ZZZ \\ ZZZ}%
}
\AddToHook{shipout/background}{%
  \put(6cm,-10cm){\AAA}
  \put(6cm,\dimexpr-10cm-0.1pt){\color{magenta}\rule{1cm}{0.1pt}}
  \put(6cm,\dimexpr-10cm-0.1pt+12pt){\color{magenta}\rule{1cm}{0.1pt}}
}

Wouldn't we then expect that the two ZZZ sits on the two magenta lines. Any one know of packages that can mess with \fontsize? Because in my letter class the above has the top ZZZ slithtly (about 0.5pt) above the top magenta line. Unsure why

Aha, it is the bera package

Ahh, it is setting the linespread

So if one is after a specific line spacing regardless of the current value of the \linespread/\baselinestretch, what exactly is the calculation.

I have a macro to recreate our Uni logo (or rather the text part of it), and it needs to match some spacing in the logo given just the size of the logo (we know that the line skip needs to be half that size, regardless of \lineskip. I already have the version take the font scaling option into account, but seems to have forgotten about \baselinestretch

@daleif \def\baselinestretch{1}\fontsize{10pt}{12pt}\selectfont ?

@UlrikeFischer that is probably the easiest. For now I just devided the lineskip input by \baselinestretch

@daleif or use \baselineskip (the tex primitive) not 12pt for your line offsets

@DavidCarlisle not understood. 12pt is variable here.

if in the hook you had \fontsize{10pt}{12pt}\selectfont ... \put(6cm,\dimexpr-10cm-0.1pt+\baselineskip) then that would add 12pt scaled by whatever linespread you have

Confirmation that time travellers are real: on the main site, a user reports to have TexStudio version 22.3 (current texstudio version is 4.8.2)

@samcarter they also reported to have miktex 22.3 and their log show 24.1 ;-(. Never trust users ...

@DavidCarlisle Right. My issue is probably the oposite. I have a macro that has to calculate a fontsize from a given input, and the lineskip have to measure that input no matter the value of \baselinestretch. for now I'm doing the equivalent to \fontsize{ \fpeval{some calc} \l_tmpa_dim }{ \fpval{1/\baselinestretch} \l_tmpa_dim } where \l_tmpa_dim holds the target fixed size lineskip.

@UlrikeFischer ... confused user is a much more disappointing explanation! :)

@daleif why don't you want to simply reset \baselinestretch first?

@UlrikeFischer side effects. Users not being careful about grouping might do \MyMacro{20pt}\selectfont Something \normalsize..., which have now reset the \baselinestretch to 1. If I'm in control of the use then of course that is the better solution.

@DavidCarlisle Aren't you using Emacs as well? Do you have dome completion setup for expl3 syntax? The standard M-/ does not see _ as part of a word so cannot toggle through expl3 named vars

@daleif I did have one somewhere once but not currently. I can't remember who wrote it.

@daleif this claims it should work out the box

@daleif xcoffins is in l3experimental, but the expl3 bit isn't. however, the expl3 doc refers you to the xcoffins doc for some stuff, so not entirely clear whether that bit of it is experimental or not :-).

@cfr Every function in interface3 is stable

is it possible to stop l3build manifest adding symbols to file names without rewriting whichever function? (rewriting the functions works, obviously, but seems like overkill just to remove the symbol?) if I use a modified intro, l3build still adds the symbols, even though they don't refer to anything. (and they don't group sensibly as far as I can tell.)

@JosephWright even the bits documented in l3experimental? but I really meant: it isn't clear from the docs.

[I'm sure the symbols group sensibly if you haven't redefined other things. but they don't group sensibly if you've messed around with the setup.]

@cfr That's not in interface3 :)

@JosephWright ??

@cfr which symbols?

@cfr oh you mean: Files listed with a ‘†’ marker are included in the TDS but not CTAN files, and files listed with ‘‡’ are included in both

@DavidCarlisle yes, sorry. those symbols. can I get rid of them easily? the result doesn't make any sense.

@DavidCarlisle I realise I could just include that explanation, but the symbols aren't applied consistently to all files, so it doesn't really make sense.

@cfr I must say I hadn't noticed that feature before, and I just tried l3build manifest in a couple of repos and they didn't get added. (but the text describing them does)

@cfr ah if I try it in ho-tex/oberdiek they get added (presumably as that has tdsfiles set)

@DavidCarlisle I suppose I could omit the tds archive, but then the .sty doesn't get added (presumably because it's derived).

@cfr just looking at the code I guess you can use manifest_setup in a l3build-manifest-setup.lua to over-ride the lists but ... I think I see why I nevr use this at all but it seems the key hint is

% In order for \texttt{manifest} to detect derived and typeset files, it should be run
% \emph{after} running \texttt{unpack} and \texttt{doc}. If \texttt{manifest}
% is run after also running \texttt{ctan} it will include the files included
% in the CTAN and TDS directories as well.

@cfr so if you run l3build manifest first before l3build ctan it won't flag the ctan files (did I get that right?)

Hello everyone, sorry for being pedantic but why the radical symbol has little gap.

using MathJax .

@CroCo Usually it doesn't look like that. Do you have.a minimal complete example?

@mickep it happens in math stackoverflow.

@CroCo that isn't tex

@DavidCarlisle but it says "MathJax is a cross-browser JavaScript library that displays mathematical notation in web browsers, using MathML, LaTeX and ASCIIMathML markup."

@CroCo still not TeX...

@CroCo yes exactly mathjax is a javascript library that parses a subset of latex math markup (but not other latex commands such as sections or tables) and makes a similar rendering using web platform rendering. it doesn't use tex at all.

I thought it internally uses Tex and upload it as a picture.

@CroCo you can see it's not an image as you can select text from it, eg that 9 is really a 9 (unlike the form in this chat which is simply an image)

@CroCo no it does not use server side at all, it does all the processing in your browser, in javascript.

but why this happens. Is this really hard to render the radical?

sorry I just knew now the question has nothing to do with Tex.

@DavidCarlisle just learned something new. I thought it is merely an image.

@CroCo it's usually a lot better than that but the tick is a glyph from the font, and the rule is a drawn rule, not from a font, even in tex where you have full control over both you can sometimes see a small glitch where they join. On the web where each browser is implementing it differently using different fonts on different platforms, it gets a bit harder.

@CroCo do you have a link to a posting that looks that bad?

it isn't that bad though.

I was just curious.

@CroCo I see no gaps there:

@CroCo where did you see the form you originally showed, I have never seen mathjax make output that bad, but as I say it depends on the browser and operating system in use...

If I zoom in more, I see the gap unequivocally.

At 110% view size. Chrome.

@CroCo firefox at 500% zoom you see a small notch but no gap at all

interesting. so the issue from the browser?

Edge at 200% zoom

@CroCo chrome at 400% , (edge is chrome with a microsoft badge, so same) firefox wins:-)

@DavidCarlisle what browser?

Firefox is flawless, although I can see little bump.

which I don't see any issue with Tex.

@CroCo yes but that is at 500% so not surprising

In my browser, it is noticeable in 110%.

@CroCo you will in some fonts, the original tex computer modern font "tick" has a tiny horizontal slab at the top that makes joining the rule easier but if you look at your example here the font just has a sloping tick with a rounded end, if you close the gap and start the rule earlier you are almost bound to get a small visible artifact at the join if you zoom in enough

I see. thanks alot.

@CroCo yes but you are joining a sloping, rounded antialiased glyph from the font with a fixed unaliased rule, you can do a lot better than chrome does there but there are limits. (as in tex the way to avoid all such artifacts is not to use a character at all and draw the whole thing as a single path (say in tikz in tex or svg on the web) that's possible but has other issues.

I guess it is not that bad in rendering the radical symbol.

@DavidCarlisle I think that is probably sensible.

@DavidCarlisle I don't think it affects what gets flagged. they don't get flagged anyway. I think the difference is whether they get included at all. I want them included. I just don't want any of them flagged. but really, this is just too complicated and not worth it. I already have manifest_setup which overrides the lists. but I still get symbols in the lists, even though they don't match anything.

oh, except I have to figure it out for the font cases unless I don't make tds archives. damn.

@cfr yes you could put in a feature request at l3build to have a simple option to drop those without having to rewrite internal functions