@Nasser -- You have \underset{n=1}{\overset{\infty }{\sum }}. A more usual form of this expression is \sum_{n=1}^{\infty} or, if not in display mode, \sum\limits_{n=1}^{\infty}. Did you write this yourself, or was it created by some other (automated) method?

@barbarabeeton This was auto-generated by computer algebra system Maple. I have thousands of such long equations that overflow the edge of the page and no way to break them. I can't do this manually, since each time I run my program, they will be overwritten by new output from Maple and Mathematica Latex output. Here is the current version shorturl.at/clFPR and the PDF is there so you can see. Thanks.

I think the page letter size should have been made wider. 8.5 inches is just not wide enough for math. But too late to change paper size now. This is something that should have been fixed 100 years ago.

I tried once to do landscape mode document, instead of default Portrait , but it did not look good.

@Nasser If the equations are of a particular form, perhaps you could add specialized breaking heuristics? I expect this is customizable, like everything else.

@Nasser well every country except 1 (or perhaps 2?) stopped using letter size paper years ago although typeset math journals traditionally are typeset in much smaller page sizes. But do you print your document or is it pdf-only? in which case you should not be constrained to any paper size.

@Nasser there is always:

@DavidCarlisle thanks! I actually remembered long time ago you gave this trick, but could not find it any more when I searched for it. I'll go an try it now and see how it works on my current document and will let you know.

@DavidCarlisle for this example, Maple uses \left\{  how should I change you code to handle this? as it is still one long equation going over the edge. Thanks.

\documentclass[12pt]{article}
\usepackage{amsmath}

\begin{document}

$\displaystyle\let\left\relax\let\right\relax
\left\{\left\{w(x,y)\to c_1\left(\frac{b e^{-\lambda  x}}{a \lambda }-\frac{e^{-\beta  y}}{\beta }\right) \exp \left(-\frac{e^{-\lambda  x} \left(e^{-\beta  y}\right)^{-\frac{\delta +\gamma }{\beta }} \left((\beta -\delta ) \left(c \gamma  \left(e^{-\beta  y}\right)^{\frac{\delta }{\beta }} \left(\frac{a \lambda  e^{\l…

@Nasser Could you instruct your computer software to write parts of the equation as separate expressions?

@Nasser the \{ isn't an issue, the large fraction is. Been through all this before:

E.g. \alpha = <ridiculously long exponent>

@Nasser Clearly you should make notes. :-)

You've been working on this since 2014? You poor man.

@barbarabeeton oh, don't worry, I don't really participate in editor wars, too, I won't ever mean anything about that bloated OS called editor by some seriously :)

Could anyone try the equivalent of C:\texlive\2019\texmf-dist\fonts\opentype\google\roboto>otfinfo -T Roboto-Black.otf ?

When hyperref is building the bookmarks. How does it know which level each type has?

Reason for asking: Attempting to hack pdfpages addtotoc to include a non numbered entry to the toc. I have it working, but it gets the wrong level in the bookmarks.

My solution is to use a custom section like entry and have it do stuff.

But it gets level 0 and not 1 in the .out file.

@UlrikeFischer yes

$ otfinfo -T Roboto-Black.otf
otfinfo: bad table name
otfinfo: Type otfinfo --help for more information.

@daleif it has settings like \def\toclevel@section{1}

@DavidCarlisle you too ... I get this for every font I tried ;-(.

@UlrikeFischer isn't the usage

$ otfinfo -T Math Roboto-Black.otf
otfinfo: Roboto-Black.otf: no 'Math' table

@UlrikeFischer thanks, and I of course missed the Package hyperref Warning: bookmark level for unknown references defaults to 0. warning i the log

@DavidCarlisle oh. Then it is perhaps a autoinst bug (that's where I got the error first).

@UlrikeFischer -t gives

$ otfinfo -t Roboto-Black.otf
 149725 CFF
     28 FFTM
    450 GDEF
  23858 GPOS
  12606 GSUB
     96 OS/2
   1198 cmap
     54 head
     36 hhea
   5000 hmtx
      6 maxp
   1281 name
     32 post

@DavidCarlisle I wasn't actually trying roboto but another font. So I guess I will have to find out which table name autoinst is trying to find.

@UlrikeFischer does otfinfo --dump-table 'OS/2' Roboto-Black.otf work for you or does the / mess up the windows command line

@DavidCarlisle the slash seems not to be a problem, but the quotes: otfinfo --dump-table 'post' fails while otfinfo --dump-table post works.

@UlrikeFischer you might want to play with line 1512 of the autoinst script.....

@DavidCarlisle I removed in the autoinst script the quotes in two places and now it works. I wonder if this is an otfinfo (on windows) or an autoinst bug.

@DavidCarlisle and 1233

@UlrikeFischer windows command line is a pain (all the troubles I had getting curl to work. If autoinst can avoid the problem more chance of it being fixed there)

@DavidCarlisle well the question is if it will fail elsewhere if the quotes are removed.

@UlrikeFischer I can't see how the post one can fail, OS/2 might be an issue in some commandlines but I don't think it would be a problem in any *nix shell, and if it works on windows,....

@DavidCarlisle I don't think that it is a command line thing. It is the quotes in the arguments, otfinfo seems not to strip them otfinfo -T 'post' calibri.ttf fails, and all similar calls where a string argument is expected.

@DavidCarlisle bamsoftware.com/hacks/zipbomb

@UlrikeFischer yes but with a unix shell the commandline shell removes the quotes so the application never sees them (which is why you cab use post or 'post' or "post" but the windows commandline passes the quoted argument to the application so it has to handle them itself

@DavidCarlisle double quotes work in windows too otfinfo -T "post" ..., only the single quote fails.

@UlrikeFischer perhaps suggest that to the autoinst maintainer then

@DavidCarlisle I already wrote a bug report, but will add a suggestion for the double quotes. the tabls package is not compatible with array ;-).

@DavidCarlisle, @UlrikeFischer, @egreg Looks like time for a dev release :)

What is wrong with this MWE:

DeclaeMathSizes does not seem to alter anything

reference: tex.stackexchange.com/q/499495/3929

Hmm,seems to have forgotten that the first arg should match normal size

Try this instead

switch between 12pt and 14pt class options.

Does not seem to be memoir related. extarticle has the same issue, nothing happens if we apply DeClareMathSizes to 14pt or 17pt

@DavidCarlisle Two is right. Canada uses US Letter paper too.

How the F am I suppose to know \@xivpt is actually 14.4 and thus \DeclareMathSizes{14}... never hits.

As soon as I know how that works I will create an actual issue. (I wish there was a transparent manual for how to use all this GitHub stuff, but meanwhile I was able to clone the official repository. I just wish these things were a bit more intuitive.)

@daleif I was just about to mention that....

@daleif you are supposed to use bitmap fonts at magstep sizes, none of this newfangled scalable font at whatever size you like freedoms.....

@daleif You are supposed to use \@xivpt. ;-)

Only took me 30min to figure that out. Well back to BlackBoard...

@DavidCarlisle "And to this day, David still has the 2e kernel opened inside the emacs buffer..."

@PauloCereda doesn't everyone?

nope :)

I thought everyone had the 2e source code as their screen saver

@daleif LOL

@PauloCereda Just checked I did have latex.ltx open in one of the three emacs I have open at the moment

@DavidCarlisle I rest my case, your honor. :)

@DavidCarlisle people should stop using native thingies with emulated thingies and emulated thingies with native thingies, also native thingies with native thingies, and probably emulated thingies with emulated thingies as well. :)

@PauloCereda just use emacs for everything

@DavidCarlisle ooh

@UlrikeFischer I always knew that you secretly prefer geese over ducks. ;-)

@marmot it is not a geese, it is a "Stockente".

Historical question: what was TeX's birthday? Assuming that's well defined.

AUCTeX seems to have no idea how to indent LaTeX 3 code.

@DavidCarlisle ^^

Is it better to read tug.org/TUGboat/tb30-1/tb94wright-keyval.pdf before tug.org/TUGboat/tb31-1/tb97wright-l3keys.pdf ?

Or just try reading the latter directly?

@FaheemMitha or why don't you read both and the question ends? :)

@FaheemMitha texdoc errorlog starts 10 March 1978 although presumably there was code before that if it was fixed on that date

@PauloCereda that leads to further questions, which order do you read them in, which takes priority if they are in conflict, can you read them in parallel on a split screen, ...

Um. That was a serious question.

@FaheemMitha how bad can it be to read both (what other answer could there be, I don't recall what's in either but neither is very long)

@DavidCarlisle Actually, the first is quite long.

@egreg good high level document markup in your amsthm answer:-)

@FaheemMitha Well one is more general, and gives a sense of how all the various approaches work. The second deals with l3 specifically. It seems very sensible to read both and in date order.

@AlanMunn Um. Ok. I'm a lazy student.

Someone please reassure me that this stuff is actually useful.

@DavidCarlisle or read none, which would solve both conflicts. :)

@PauloCereda I suspect that was my approach

@FaheemMitha I also found the following question incredibly helpful, especially Ryan's answer about pgfkeys.

@AlanMunn Thank you. I'll take a look.

@DavidCarlisle ooh copyright infringement

@PauloCereda Very helpful.

@FaheemMitha useful for what? Sorry but can't you decide yourself this by skimming the first two paragraphs?

@UlrikeFischer Nope.

@PauloCereda I think someone thinks @AlanMunn is being more helpful than us.

@DavidCarlisle bah he's using some sort of linguistic tomfoolery to trick people. :)

Wow, an actually complete and marginally comprehensible example.

@PauloCereda speaking of languages did you like by Brazilian Portuguese answer?

@DavidCarlisle link please

@DavidCarlisle I like the language usage. :)

@PauloCereda very suitable for a linguist like wat I is

@DavidCarlisle :)

I take it all these key packages work in similar ways. Is that a fair statement?

Trying to read the documentation, but at the moment it's gibberish.

@FaheemMitha in what way? They split things at commas then into key-value pairs?

@JosephWright They split things at commas? All of them?

I must have missed that part.

@FaheemMitha Almost all, at least: there are one or two parsers for doing keyvalue pairs on a line by line basis. But almost everyone expects foo1 = a, foob = b to be treated as two keyval pairs

@JosephWright Ok.

Hmm, it looks like the first of those packages was keyval.

@DavidCarlisle Nice, isn't it?

@FaheemMitha More or less, yes

@FaheemMitha Written by our own @DavidCarlisle

@JosephWright I thought it was probably not a different David Carlisle.

@JosephWright ooh

@FaheemMitha In my article, I didn't cover every keyval implementation; some are a bit niche

@FaheemMitha tex.stackexchange.com/questions/26771/…

@JosephWright The first article?

@FaheemMitha Yes; first one is general, makes some recommendations, second is focussed purely on the l3keys approach

@JosephWright That's a lot of packages.

@JosephWright Yes, I see. To repeat what I asked earlier, do you recommend reading both of them?

@FaheemMitha If you are looking at keyval packages, I'd say choice of three is worth considering: keyval, pgfkeys, l3keys

@JosephWright Ok. Thank you.

@FaheemMitha Start with the first. If you are interested in l3keys at that stage, read the second

I get the feeling I've probably seen these in action many times without noticing.

@FaheemMitha Oddly, I'd say l3keys is overall the most powerful :)

@JosephWright Ok. Thank you for the advice.

@JosephWright Why oddly?

@FaheemMitha Yes: keyval drives for example graphicx, pgfkeys all of pgf, l3keys anything using expl3 with keys (e.g. siunitx)

@FaheemMitha I wrote it

@JosephWright Um. Ok.

@FaheemMitha I joined the team on the back of the draft version of l3keys

@JosephWright What year was that?

@JosephWright @FaheemMitha And if you have plenty of TikZ code independently, then pgfkeys is a natural choice. No need to learn two systems.

@AlanMunn I haven't used Tikz much. I created a few diagrams some years ago.

LaTeX 3 seems like it might be the future, so it's probably worth at least considering.

Though I find the LaTeX 3 syntax quite intimidating.

Right now I'm just trying to make sense of what I'm reading. It seems more work than I was expecting. But that's typical of TeX, at least for me.

@FaheemMitha Then yes, l3 keys and l3 in general might be worthwhile.

@AlanMunn Ok. I'm looking at keyval to start with. Under the assumption that the earliest would be the simplest.

Keep getting harassed and harassed and harassed by the same user over and over. Sad. Really sad.

@egreg my answer is nicer

@marmot would you like more users to harass you?

@DavidCarlisle You surrendered too early

@egreg don't teach people to shoot themselves in the foot

@PauloCereda see any of the cricket?

@DavidCarlisle no...

@PauloCereda don't mention anything to Will but...

bbc.co.uk/sport/cricket/48870770

@DavidCarlisle ooh

@DavidCarlisle No.

@marmot Please flag any content that you feel is abusive: applies in general, as it's the only way to see a pattern of behaviour (and for that to be available to SO staff as well as local mods)

@JosephWright I did that. The pattern is always: 1. "you stole my code!" After this is ruled out: 2. "you stole my idea" (instead of saying "yes, you're right, your code is not from my post") and after this is ruled out, other accusations come, including being called "liar". (And this user does get "inspired" by other posts they do not mention.)

@Nasser -- Sadly, the link shorturl.at/clFPR simply connects me to the shortener page, not to anything useful. But I have to say that using \overset and \underset for limits is terrible practice, and whoever caused a computer algebra system to do that should be ashamed of themselves.

@marmot Works on master.

@HenriMenke Thanks! (How did that happen?)

@barbarabeeton sorry about the link not working. will not use short URL again. Here it is again. 12000.org/my_notes/pde_in_CAS/maple_2019_and_mma_12/index.htm and the PDF file link is at the top left of the page. After about page 130, you'll start seeing the CAS latex output. Below each yellow box. Most overflow the page. Maple Latex output is not very good. Maplesoft does not seem to maintain Latex any more. I think they went to different technology than Latex many years ago.

\dimen0=1pt
\dimen1=2pt% <- oops, no space
\ifdim\dimen1=0pt\relax
    \dimen1=\dimen0
\fi
\showthe\dimen1 % prints 1.0pt

The assignment takes place, even though the branch isn't taken.

@AlanMunn Look who's luring people into l3 now ;-)

@PhelypeOleinik :) I'm nothing if not pragmatic.

@PhelypeOleinik Although given that the packages I have in 'development' use TikZ so I'd probably be using pgfkeys myself.

@PhelypeOleinik I really have to overcome my prejudice against the l3 syntax conventions.

@Nasser -- That is indeed pretty bad. I'm not able to interpret the meaning of anything, but the presentation in a number of cases looks questionable. It also looks like there's a blank line between the end of every yellow box and the display that follows it -- vertical spacing is greater above than below the displays. Regarding Maple, maybe you could find some other LaTeX users, band together, and complain. (Or at least find out what they're doing.)

@AlanMunn You'll get used to it, if not by love, by hate, with me and my annoyance potential being an official advertiser ;-)

@barbarabeeton there was lots of complaining about Latex in Maple. It is a lost case. Here are some links FYI mapleprimes.com/posts/43828-Better-LaTeX-Output-From-Maple and mapleprimes.com/posts/… and many more. myself have no experience to fix Maple Latex output manually as I am not very good in Latex internals to be able to do this.

@PhelypeOleinik I'm happy to have your cheerleading!

@AlanMunn :-)