Although I believe that your syntax makes very much sense, it appears as if the existing \substack command is simply using \\ for next line.

Apart from that, the outcome looks good! Maybe one would like to test with \left \lfloor \frac{n}{2} \right\rfloor as well to see how that behaves?

Not really sure how to go about adding the \\ as next-line behavior to the substack input due to the python's handling of escape characters, I'm currently still looking into it.

Added the \left \lfloor \frac{n}{2} \right\rfloor as a test with the current implemention.

Maybe the parser expert @anntzer has some input?

I would expect pyparsing's DelimitedList would be helpful here

Note, however, that it looks like it is in a slightly awkward transition from the delimited_list function to the DelimitedList class, the former being deprecated upon the introduction of the latter in pyparsing 3.1.0 (which we are incompatible with for other reasons right now, but should fix)

As such may need something like:

if pyparsing.__version__info__ < (3, 1):
    from pyparsing import delimited_list as DelimitedList
else:
    from pyparsing import DelimitedList

Maybe this could be written using max, which also takes a key argument.

max_width = max(toks[1:], key=lambda c: c.width).width

Something like that? Or even simpler:

max_width = max(map(lambda c: c.width, toks[1:]))

which seems more readable to me.

Thanks! That was neat

Took me an embarrassingly long time to understand!:sweat_smile: I think it separates each element of vlist with Vbox(0, thickness * 2)?

This is probably one of those instances I'd prioritise readability.

vlist_ = [None, Vbox(0, thickness * 2)] * len(vlist)
vlist_[::2] = vlist
vlist_.pop()

This should achieve this same thing. (Another option is to append the Vboxes in the for loop just above.)

@oscargus: I'd defer to your opinion here. What do you think?

Maybe one could just do:

vlist = [val for pair in zip(vlist, [Vbox(0, thickness * 2)]*len(vlist)) for val in pair]
del vlist[-1]

?

(But, yes, I find double (and triple) list comprehensions quite hard to follow to start with.)

I take it that there will be another update to this and @ksunden was only sorting out the parsing?

But this line should go.

This should then be modified to use \\, but ideally there should be no need for other changes.