@MarcelKrüger: I'm not able to reproduce the strange overlapping integrals even when using \intop and various settings. For example, with NoLimitSubFactorand NoLimitSupFactor unset, and with the long class code not set I get the following:

@mickep Am I right that the box is the width of the glyph? I guess my main point of confusion is the handling of italic correction in ConTeXt. My understanding of italic correction is that it encodes the amount an italic glyph extends outside of it's box (to the right), but I get the impression that ConTeXt tries to modify the width of the glyphs to include this in the width.

Then I would assume that the italic correction is zero since no correction is needed anymore, but it seems like ConTeXt still uses the italic correction as some kind of independent positioning indicator...

@MarcelKrüger The box is the width of the glyph, yes. Don't you have the same in latin modern? I do not see in the goodie file that the glyph of the integral is changed. I think italic correction is less used in ConTeXt than in LaTeX. As far as I remember it is described in some microsoft specification when it should be applied, and I think that is more or less followed. In the example image above, no italic correction is applied (that would have shown with a number in blue).

In fact, isn't the use of italic correction to place limits on the integral just some old hack needed because the integral sign in computer modern was so leaning?

@mickep I'm currently working with Computer Modern 8bit fonts and not with Latin Modern, but when I'm looking at Latin Modern I get similar values. The latinmodern-math font file lists the width as 66.5% of the font size, while in ConTeXt I see a width of 11.98141pt for a font size of 11.99341pt, so quite a bit higher. Since the font file has italic correction at 33.2% of the fontsize, the width in ConTeXt appears rather similar to the width+italic correction in the font.

@mickep The spec for MATH tables specifies that the scripts should be separated horizontally by the italic correction, but with the *2 line I get separation of two times the italic correction from the font.

This is how it looks in fontforge. The 665 seems to correspond to the width of the orange box above, no?

I probably miss something.

@MarcelKrüger Forgot to ping you.

@mickep Ups, I looked at the wrong glyph. Sorry about that. (The glyph I get in math is of course the bigger variant which is obviously wider... Then the widths fit much better.)

@mickep Well, that's only done in the limits case AFAICT, so I don't get that for nolimits case. But even then, this applies half of the italic correction to both the upper and the lower limit, to together they are still separated by the full (on in this case 2*full) italic correction.

@MarcelKrüger Hm, ok. I can discuss things with Hans once more, probably today. But do you really want to rely on italic corrections for the placement of limits in integrals? And so, if in \nolimits, you want to add half of it at the top and substract half of it at the bottom?

@mickep I would like to have at least the option to use italic correction for that (both for limited compatibility and since that's one of the specified uses of italic correction in the MATH OpenType spec) For real limits (as in above and below the sign) I'm perfectly happy with the current behavior (which adds half at the top and substracts it at the bottom).

For \nolimits I think that the existing flexibility with \mathnolimitsmode and the parameters is pretty great, but all the distances seem doubled. This isn't a big issue, I could easily use the flexibility to adjust that, but it seems odd and I don't understand why it's required.

Does anybody know of a good dupe-target for "command already defined, use \renewenvironment to avoid the problem"?-type of problems? (I'm searching for a duplicate for tex.stackexchange.com/questions/501821/… )

@samcarter Is it something like this you are after?

New sanskrit package version released! 🎉

@mickep Goes in the right direction. I was hoping on something which would mention \renewenvironment

@samcarter Ah, yes, I see.

@wilx Yeah! Let's have a party :)

🍾

There is literally just one line change. :D

@MarcelKrüger We'll see what comes out of it.

Some characters simply do not look good in hats (yes, rsfs A, I'm looking at you!)

@mickep Not everybody can rock wearing a hat like wmagazine.com/culture/… :)

Btw, I'm also in search of a question. Do we have an explanation why shortening \begin{equation}...\end{equation} to anything like \beq...\eeq or \eq{...} is a bad idea? Can't find it now :(

@samcarter Oh, very nice!

@yo' Not for equations as such, but this answer explains why this obfuscates the code tex.stackexchange.com/a/326879/36296

@samcarter that's a good one, thanks! I also found tex.stackexchange.com/a/100167/11002 but that's about align. And then there's tex.stackexchange.com/q/316776/11002 that would deserve a proper do-no-do-it answer.

@yo' -- \beq and \eeq do work, but that made users eager to shorten \begin{align}, etc, which do not work. I think there's an explanation in the amsmath documentation or one of the other documentation files in the amsmath package distribution on CTAN.

@barbarabeeton do work is not really the same as isn't a bad idea though. I have a curious user asking how to do it and I want to discourage them :)

@yo' -- Oh, agreed. But Mike Downes was persuaded to restore the ability to use \beq and \eeq because they were allowed in ur-LaTeX, although he didn't want to. He didn't think it was a good idea, but was overpowered. It isn't a good idea, but the only justification I can think of is the analogy.

@yo' Unfortunately such syntax is probably on the rise as users get used to them from tools like cran.r-project.org/doc/manuals/R-exts.html#Mathematics