TeX, LaTeX and Friends

« first day (4708 days earlier) ← previous day next day → last day (384 days later) »

Skillmon

6:12 AM

@DavidCarlisle so no Dodos.

3 hours later…

Paulo Cereda

9:09 AM

I definitely didn't get any secret emails, specially from @DavidCarlisle

David Carlisle

@PauloCereda ducks are good at secrets and truth

Paulo Cereda

@DavidCarlisle ooh

2 hours later…

daleif

10:51 AM

Anyone know if there is an upper limit to how wide a tabularx can be? A PhD student is creating a very large poster (1m by 2m), landscape in three columns. If the width if a tabularx in one of the columns exceeds a certain size, the headers/cols collapse and I do not understand why

David Carlisle

@daleif i wish people wouldn't use tabularx so much:-) It has some guards to avoid internal calculations exceeding maxdimen, but its arithmetic is just primitive tex dimen arithmetic and can be easily stressed, probably you can force it to make a smaller initial guess for X widths, do you have an NWE (send by mail if it's not very minimal)

Ulrike Fischer

@daleif without an example it is difficult to say if it is a feature, but tabularx for this sounds wrong anyway.

@DavidCarlisle ^^ improving my english spelling

David Carlisle

@daleif just using p{.33\textwidth} and not using X would be a lot easier for TeX:-)

@UlrikeFischer wunderbar

daleif

The interesting part is, it works with \begin{tabularx}{0.96\linewidth} but fails at \begin{tabularx}{0.97...}.

Here the column is about 1380pt

David Carlisle

11:09 AM

@daleif what does \tracingtabularx log for the working and not working cases?

daleif

11:32 AM

(tabularx) Table Width Column Width X Columns
(tabularx) 12813.13077pt 1380.0pt 12
(tabularx) Don't exceed \maxdimen
(tabularx) Reached minimum width, backing up.

@DavidCarlisle ^^^ and the X columns basically have no proper width

MWE:

\documentclass{article}
\RequirePackage{geometry}
\geometry{a0paper}
\usepackage{xcolor,array}
\usepackage{tabularx}
\usepackage{booktabs}
\newcolumntype{C}{>{\centering\arraybackslash}X}
\begin{document}

\rule\textwidth{3mm}

% 1360pt works
\begin{minipage}{1380pt}

\tracingtabularx

\setlength{\tabcolsep}{0pt}

\begin{tabularx}{\linewidth}{
p{4.2cm}
|p{8.9cm}
*{11}{|C}
}
\toprule
&
&
\multicolumn{2}{c}{huh}
& A
& B
& C
& D
& E
& F
& \shortstack[c]{G\\H}
& \shortstack[c]{I\\J}
& K
\\
\midrule

It seems to work just fine if I remove the 4 lines in \TX@arith dealing wit don't exceed ...

Cheating:

\makeatletter

\patchcmd\TX@arith{%
\ifdim\TX@col@width>\@tempdimb%
}{%
\ifdim\TX@col@width>\maxdimen%
}{}{\ERROR}

\makeatother

David Carlisle

11:52 AM

Thanks, I'll check later. Reached minimum width, backing up. means it's decided it needs to make negative widths so it gives up and just makes them 1em so something gets printed, but it's possible its arithmetic got confused before that.

daleif

12:09 PM

Well it seems to start out with a col width of the first arg to tabularx (here 1500pt) and then sets \@tempdimb = \maxdim / 12 which is less than 1500pt and thus stops right away. Not sure what the proper approach for it is here.

Ulrike Fischer

12:20 PM

@daleif perhaps in \TX@endtabularx one could set \TX@old@table to something slightly different to \maxdimen? \TX@old@table\dimexpr\maxdimen-1sp\relax @DavidCarlisle

David Carlisle

1:07 PM

@UlrikeFischer hmm possibly, yes

2 hours later…

David Carlisle

2:53 PM

@barbarabeeton you don't remember the ppppp author's name do you?

barbara beeton

@DavidCarlisle -- I think I can retrieve it. You like that example, huh?

David Carlisle

I had a colleague who had to fight with a publisher for additional p at one time.... @barbarabeeton

I wonder if it was related to ams.org/journals/mcom/2015-84-293/S0025-5718-2014-02865-2/… ... @barbarabeeton

barbara beeton

@DavidCarlisle -- Don't think so, but not sure. Author's first name is Larry, and I think last name begins with E. (I may remember the name at 03:00.)

David Carlisle

I had a phd colleague a year above me who's thesis appears in the refs there, that was published by j algebra not ams, but that area (finite groups) does like to use lots of different font styles. he had 5 (I think) p in his thesis.

barbara beeton

3:13 PM

@DavidCarlisle -- It may be this book: Invariant Theory of Finite Groups by Larry Smith and Mara Neusel. (If so, I was wrong about the "E".)

David Carlisle

oops I know them:-) @barbarabeeton

barbara beeton

@DavidCarlisle -- Is that a good example of appearance = meaning?

mickep

@barbarabeeton Not Lawrence C. Evans?

barbara beeton

@mickep -- Thanks for the suggestion, but I don't think so. The name of the co-author of the book I cited sounds familiar.

Ulrike Fischer

@DavidCarlisle are you back to the "find literature"?

mickep

3:40 PM

@barbarabeeton Well, worth a try. 🤷🏻‍♂️

barbara beeton

@mickep -- David's "oops, I know them" is the clincher.

mickep

@barbarabeeton Ah, I did not read carefully enough.

David Carlisle

4:18 PM

@UlrikeFischer no, I'm not playing that game

Ulrike Fischer

@DavidCarlisle then why are you looking for p's?

David Carlisle

@UlrikeFischer barbara mentioned remembering an author of an ams monologue who used 5 lower case p styles, and I feared it was a friend of mine

Ulrike Fischer

@DavidCarlisle oh ;-)

David Carlisle

4:56 PM

@UlrikeFischer it wasn't who I feared but it was someone I knew and met at a few conferences (and a nice week in Dagstuhl castle a long while ago)

1 hour later…

mickep

6:08 PM

@DavidCarlisle The book by Neusel and Smith linked above also includes several types of upper case P.

David Carlisle

@barbarabeeton ^

barbara beeton

@mickep -- You've got a copy? Can you identify the page(s) on which the "p"s (and "P"s) are defined? It would be really useful to get a scan showing the evidence.

mickep

6:50 PM

@barbarabeeton The department library had it. I can check next week.

barbara beeton

7:06 PM

@mickep -- Thanks.

mickep

@barbarabeeton But I don't know if it is worth scanning. They should trust people...

barbara beeton

@mickep -- While I agree, the UTC has their own ideas about what constitutes evidence, much less proof.

David Carlisle

@mickep would you trust @barbarabeeton?