@SašoŽivanović The new version of Forest you sent me works great - thanks! I need a lot of #s in the following, though:

\documentclass{article}
\usepackage{prooftrees,amssymb,turnstile}
% Sašo Živanović: tex.stackexchange.com/a/296771
\newcommand\hiddenparcommand{\par\noindent}
\newcommand\hiddencommacommand{, }
\forestset{%
  declare keylist register={split here levels},
  declare keylist register={split here interjects},
  declare toks register=split here toks,
  split here levels={},
  split here interjects={},
  to widest/.style={
    tikz+={\path (\forestregister{tempdima}, \forestoption{y}) -- (\forestregister{tempdimb}, \forestoption{y});},

This does split the tree multiple times. It also adds a lot of space. I guess it simply hides the upper part of the tree each time it draws it - the upper part is still there taking space, but just isn't drawn. Is that right?

@cfr That's right. I fixed that after sending you the .sty. Right now, I've got just a single remaining whitespace: after line 10. I have no idea why. I'm sending you the new .sty, maybe you figure out something.

@cfr Also, for this specific usage, an extra complication: closures might better belong with the line above?

@cfr Seems the right number of hashes ... ;-)

It seems to work with crocket's forest tree: tex.stackexchange.com/a/300597.

@SašoŽivanović Closures are the least of it. The only way to do it, I think, is to split the tree much, much later. Otherwise, anything which moves causes havoc. level is probably not the right way to split for these trees: it will have to split by line number, which will therefore have to be done late. At least, I think so.

@cfr Good job, all the way! I hope I can ask you that you edit the answer to work with 2.0.3 once its out. There are some bugfixes wrt to tree drawing.

@SašoŽivanović If you remind me to do it and which answer ;).

@cfr Yes, I agree, line number is the way to go. Btw: the splits are happening late. Very late. Actually, they are not so much splits, technically speaking, as multiple spellouts. ;-)

@cfr Will do! Thanks.

@SašoŽivanović OK. Yes. The splitting is happening late. But the where to split is being decided very early. So it depends what you mean by 'split': the point at which the tree is divided, or the point at which the division is effected.

@cfr Ahh I see. Yes of course.

Or maybe I should do it by y....

But that would have the closures problem, I think.

@SašoŽivanović When are you expecting 2.0.3?

prooftrees has a couple of bugs. But my current version isn't compatible with 2.0.2.

@cfr This weekend, if optimistic. But next weekend if realistic. ;-)

@SašoŽivanović Great. You have a long weekend or not?

@cfr Until monday.

@cfr I guess it should be a complex nodewalk, somehow: first by line no, and then plus closures plus whatever else ...

@cfr And then to the next level: automatic pagebreaks. ;-))))

@SašoŽivanović Monday's a bank holiday here, too: Friday, Sunday, Monday.

No in Scotland, I don't think. Only Friday and Sunday there.

@cfr Uff, you're right: so, until Tuesday! ;-) Getting more optimistic!

@SašoŽivanović Is that even possible? Never mind sane. I know that's what crocket wants, but crocket's not bothered enough to update TeX to get Forest 2 so manual can't be that bad ;).

@cfr ;-(

@cfr ;-) It was meant as a joke ...

@SašoŽivanović Jokes are dangerous. crocket can visit this room ;).

[And will in a bit, I expect: I left a link to the spacey split tree I posted above.]

crocket, if you visit, don't read this! ;-)

And I deliberately didn't use @ ;-)

@SašoŽivanović Maybe you can persuade Joseph to deep freeze that part of the conversation ;).

@cfr ;-)

@cfr I lost my first green tick thanks to you and prooftrees. Well done! :) It looks really nice.

@AdamLiter Ah...Yes, well, sorry about that :).

@cfr The package looks really nice, so I'm glad you have the green tick now! :)

@AdamLiter That's very kind.

@SašoŽivanović The code I posted in this answer does not work with the latest release candidate you sent me. Is this expected collateral damage or not?

@cfr It works with my current version. Can you test if that's the case for you as well? (Sent over email.)

@SašoŽivanović Yes, thanks. It seems to work fine with rc5.

@SašoŽivanović By the way, you don't by any chance have any idea about the following do you?

@cfr Nope ...