I need to check how the function is implemented

or even simpler tex.setbox('global', 2, head, 7)

@Patrick ... try this: \directlua{tex.setbox('global', 2, head, 7)}

the error message is wrong, the node is there :-)

anyway ... what do i care about error messages wanted to break the linebreaking ...

what's the best way to avoid using \sloppy document-wide to fix overfull \hbox problems?

@donc_oe have shorter boxes:-) It depends why you are getting overfull boxes, for example it may be that you don't have the right language hyphenation

@FrankMittelbach This is a (minor) implementation bug. It takes the last argument and checks if it is a node list

@Frank you can take a look at the source if you like: ltexlib.c "setbox()" and "vsetbox()" and lnodelib.c "nodelist_from_lua()"

@donc_oe better to ask that sort of question on site but the first one use url package and \url{....} so it breaks at /` for the second you need to allow a break in the hyphenated word I think some babel packages use "- for that or if not -\hspace{0pt} would work

well, i will post a question with code example i think

@PatrickGundlach yes. nothing erious I agree, but it took me a long time to understand (and minimal understanding of lua doesn't really help)

@PatrickGundlach as I aid, this is not really my concern. My real concern is to make this linebreak_filer callback not to add anything to the current vlist

@FrankMittelbach You are dealing with functionality that not many people have used before. I went through the ordeal of trying out lots of those functions (and requesting them in the first place)

@PatrickGundlach could be the case ... on the other hand that stuff like linebreakfilter etc goes back to my paper from 91

by which I mean there I asked for it :-)

and I implemented a solution for TeX3 actually, but couldn't convince Don back then to put it in

@Frank so you have an empty hbox/vbox that you return from linebreak_filter, right?

What is wrong with ...

@PatrickGundlach seems that is necessary according to documentation (has to be an hlist

...\hbox(0.0+0.0)x0.0, direction ???
...\glue(\parskip) 0.0 plus 1.0

then?

except for the plus 1.0?

I don't want it because I have to get rid of it again

I can handle it on TeX level by parsing everything into a vbox then then throwing the vbox away but ... would like to avoid this

@Frank but isn't this the box you have just created?

@Frank (there is my limited TeX knowledge)

@PatrickGundlach it is, but he wants to return nothing, not an empty box

@Patrick I generate the bloody hting only because that is what linebreak_filter wants as a return value

I understand

then luatex goes on an puts whatever i put as return into the current vertical list

and adds parskip even "urg" :-) what I want is to suppress all that

Then why don't you use the post_linebreak_filter (and set it to false) - then you don't put anything in the mvl (IIRC)

Hm if I set \arraystretch to another value, >{\centering\arraybackslash}m{<myvalue>} doesn't work anymore. :(

Any hints? :)

I can try that, if that works it is my answer, but from how i understand it this is not going to do that ... I'll see in a minute

@FrankMittelbach I've got it wrong: don't set the post_linebreak_filter to false, instead return false

@PauloCereda doesn't work as in causes segfaults, or doesn't work as in not horizontally or vertically centering?

@DavidCarlisle Not vertically centered. :) My bad, I forgot to add it to my sentence. :)

I must be careful, both authors of array are here. I can be killed (or severely injured). :)

@PauloCereda well I could have guessed:-) Probably it is all working but your visual assessment of vertical centering isn't the same as latex's

@DavidCarlisle Can be. :)

@PatrickGundlach nearly :-) that gets rid of the empty hbox ... now I only have to find the place that adds the parskip

any hind for this one?

@FrankMittelbach Is this also introduced after the linebreak_filter?

@DavidCarlisle It the old recurring problem, "how to v-center an image + text inside a tabular environment". :)

@FrankMittelbach no

@PauloCereda m causes the box with that cell to be vertically centred and arraystretch causes the height of a strut at the beginning and the depth of a strut at the end to be changed, but the actual effect of those three things on the visual position of the text in the box isn't always intuitive. similar to the fact that \[10pt] in a tabular does something but not always what people think it does.

@DavidCarlisle Ah! :)

@PatrickGundlach it seems to be added after linebreak_filter but even setting post_linebreak_filter to return false doesn't make it go away.

the reference manual doesn't give any hint.TeX parskip is added when linebreaks ands and its result is appended on the current vertical list

@Frank perhaps this is helpful (at least it is for me)

I use it all the time and it creates a dot /graphviz file with the nodelist in the given boxnumber/head

I then can see what function introduced what nodes. But if you know that already then it is probably of not much help

@DavidCarlisle: got a nice workaround with Martin's adjustbox. :)

@PatrickGundlach nice. But yes, I guess I understand that as long as it isn't new node types that have no equiv in current eTeX/pdfTeX

but I think I may have solved the riddle, the parskip is actually not added by the end of linebreak, but by the beginning of the next paragraph if the current vlist is not empty

so basically the answer is there is nothing much one can do about it other than resolve it on the TeX side

@FrankMittelbach If you get this working it's going to make a mess of the diagram you posted earlier:-)

@DavidCarlisle a) it is already working b) not really it just means that all one has to do to get some decent platform for complex automatic layouting is to wait for luaTeX to be fully stable

It also means that I missed out on this one, I have asked for this for 20 years and now it seems finally to be available in one successor to TeX.

@DavidCarlisle: ^^

@FrankMittelbach I collect material and box it together with Lua anyway, so I can use tex.linebreak() as often as I want with different parameters

I do everything in Lua except for the \directlua{ start_program() } call.

But I wouldn't call that TeX anymore

@PatrickGundlach as long as it is able to interpret TeX input in full power ...

@FrankMittelbach that is exactly what I don't do ;-)

I only use TeX as a paragraph builder and a pdf writer

@PatrickGundlach what do you mean by "collect material and box together with lua ..."

what is your input?

@PatrickGundlach well, I have also written graphics in PostScript ...

I read my (plain) text from XML directly (libxml2) and I have an external style file (set of rules). Everything else is done with simple Lua algorithms and boxes /glue

then I just call node.write() to push it to the PDF file

@PatrickGundlach wrong side of the house then :-)

That way I never use TeX's reading routine

@FrankMittelbach Yes, but it saves me lots of headaches

@PatrickGundlach but wouldn't satify processing all the TeX documents out there and for good or worse the flexibility that it offers their authors

@FrankMittelbach It's of course a completely different application - that's why I said that "i wouldn't call that TeX anymore" - it is completely incompatible with existing TeX documents

The style file I apply is very free (turing complete). I have a few building blocks (rules, text areas, tabular material and boxes) and the rest is up to the author of the style file

@PatrickGundlach yeah looks like it

Mein gut

@N3buchadnezzar Ciao

Why is every question on the site how to do bad typhography?! I am going mad. Mad I tell you!

I love bad typography!

@N3buchadnezzar The one with ten line chapter titles in boldface and small line length? :P

@N3buchadnezzar what exactly is typhography?

Me ToO $\cfrac{1}{2}$,$5+3$ !

I've told my students: if I catch anyone using `\\` in his documents, they get lower score

@egreg The miniTOC in the footnote was also good. :)

@FrankMittelbach Hard to tell, but if it either contains comic sans or makes your eyes bleed it is bad typhography.

@PatrickGundlach so they obviously don't have to do tables or complicated math for you

@PatrickGundlach You have a latex course? Me wants!

@Frank that's an exception, of course

Quid est veritas?

What is the truth ?

:P

@N3buchadnezzar Yes - I need money ;-)

@N3buchadnezzar sorry I was picking o nthe extra "h" :-)

@PatrickGundlach LOL

@PauloCereda I use that quotation too, when talking about logic. The students, at first, don't remember the source.

Ad mortem rectum coitus.

@egreg Oh. :)

@DavidCarlisle: Your answer takes the cake. Undelete yours and I'll delete mine.

@PatrickGundlach but isn't good typhography about doing the right thing in the right circumstances?

Well is a clean tex code or a clean document most important?

@N3buchadnezzar yes but you have to write your document in xml, split it up into paragraphs in lua and just set the boxes with latex, other than those minor adjustments it's all exactly as specified in the latex book.

@FrankMittelbach Yes, but not {\bf Introduction}\\[5mm] instead of \section{Introduction}

... and I see things like that lot of times

@DavidCarlisle I`ll just write it in notepad then, type it into mathtype, paste it into word and hand it in =)

@Werner really I must have misunderstood his comment:-)

@N3buchadnezzar I've seen worse :-)

@FrankMittelbach Wow =)

@DavidCarlisle The title just changed to horizontal spacing...

@Werner well I undeleted he can accept whichever one he wants:-)

@FrankMittelbach You see, I've had less difficulties in converting to LaTeX some Plain TeX papers than purportedly LaTeX ones to "real LaTeX".

@DavidCarlisle laughing out really loud

@PatrickGundlach sites like this are partly the reason for this ... just throw a little question here and there instead of actually learning something to use properly

I had to rotate a text inside a tabular. Only \rotatebox and it's done? Where's the catch? I'ts too easy.

Well, I do not know how much you learn from just asking people to draw things for them in tikz :p

@FrankMittelbach There are a lot of good questions inbetween.

Draw me a rubber duck for my thesis now! I have an partey to attend! Thesis needs to be done in tow deys man1

@PatrickGundlach no doubts

and it is fun to answer the interesting ones ... but what I mean is something different. If people can't be bothered to read even something like lshort but instead ask what it could possibly mean tha \foobar throws an error "undefined csname" ... hmmm

Latex manual is over 10 pages, would not read 0/10.

How surprising that then {\bf My Section}\\ \null \\ \\ \\ becomes part of the folklore

\vspace{3ex}

\begin{center}
{\large \bf 2. Preliminaries on coalgebras, comodules  and the associated
pseudocompact algebras}
\end{center}

 \vspace{2ex}

@egreg Which part is the LaTeX code I can not tell.

it can very well be the L3 syntax

@egreg yes .. but he or she did know something, like ex is a vertical dimension :-)

I know this is silly, what is the main difference between \vspace , \vspace* and ex and em ?

@percusse The same author used this gem

\def\ya{\mbox{\rm (a) \hspace{0.3em}}}
\def\yb{\mbox{\rm (b) \hspace{0.3em}}}
\def\yc{\mbox{\rm (c) \hspace{0.3em}}}
\def\yd{\mbox{\rm (d) \hspace{0.3em}}}
\def\ye{\mbox{\rm (e) \hspace{0.3em}}}

For writing enumerated lists, of course.

@egreg A typical OCD patient maybe?

I`ve had some ugly cases in my youth too. Do not try to say you all, have been perfect at tex from the get go.

@FrankMittelbach He's a very good mathematician, actually. But he wrote in Plain TeX and the conversion to LaTeX hadn't been so good. :)

One should note the nicety of closing the group just before \end{center}

@N3buchadnezzar It took me a while to get that depth is not the third dimension of a box.

@egreg to me this looks a little bit like I don't like the layout but i can't change the LaTeX behavior (which is not trivial in 2e) so i use visual markup

Why are my two line titles so bad?

@egreg perhaps the center env was added later?

@percusse ^^ I do not like that the xyz coordinate system is wrong in tikz...

I guess it is better than to center the entire document.

@FrankMittelbach Of course the instructions for submitting papers for the Proceedings were: just use the article class and amsmath.

@N3buchadnezzar I think I don't understand what you mean

@N3buchadnezzar Well, this would add \makeatother to \centering. Not very useful. :)

@egreg all your code snippets cam from there except perhaps \bf

:-)

@FrankMittelbach What do you say of this one?

\title{Semigroup$\!$ Rings$\!$ that$\!$ are$\!$ Inside$\!$ Factorial\\
  $\!$and$\!$ their$\!$ Cale$\!$ Representation}

And I forgot that the "manual sections" paper started so

\documentstyle[12pt,twoside,xy]{article}
\input{amssymb.sty}

That was in 1994

I like very much this one

\begin{thm}\rm Let $R$ be a PVMD. Then $R$ is a GK-domain if and only
if ${\cal D}(R)=\{(JP_1\cdots P_r)_t,\, J \,\hbox{is a }\,
t\hbox{-invertible}\,\,t\hbox{-ideal}\, \hbox{and}\, P_1,\ldots, P_r
\,(r\ge1) \,\hbox{ pairwise}\,\\ t\hbox{-comaximal} \hbox{ prime}\,
t\hbox{-ideals}\,\}$.
\end{thm}

Notice the \\ to break the line manually: for strange reasons the line breaking was not optimal. :)

@egreg Oh please, @egreg stop hurting our eyes

@PatrickGundlach This was my translation

\begin{thm}
Let $R$ be a PVMD. Then $R$ is a GK-domain if and only if
$\mathcal{D}(R)$ is the set of ideals of the form $(JP_1\cdots P_r)_t$,
where $J$ is a $t$-invertible $t$-ideal, and $P_1$,~\ldots,
$P_r$ \textup{(}$r\ge1$\textup{)} are pairwise $t$-comaximal
prime $t$-ideals.
\end{thm}

I have read your tugboat article (I don't remember when it was published) on that subject (was it a tugboat article? My memory is dysfunctional at the moment)

@PatrickGundlach Yes, those are from that paper, but it was only a selection. I have many more. :)

Does anybody know the name of the macro for me to increase the space between the part number and part title in the TOC? So far, I know it starts with \cftpart..., but I can't find it.

@PauloCereda Ouch! Parts are a headache with tocloft: they don't use the same structure as the lower level sections

@egreg Indeed! I'm lost in the manual! But I found it, it's this one: \setlength{\cftpartnumwidth}{25pt}. :)