TeX, LaTeX and Friends

3:13 AM

@PauloCereda We missed you!

@Zarko Serial downvoting was reversed

kan

7:29 AM

@yo' Yo! It's been a really long time! How are you?

Hello everyone...

7:45 AM

@kan As if you never left. @PauloCereda's not finished his thesis, @egreg answering questions about % and stealing ticks, this chat discussing linguistics or Word or football or anything that isn't TeX.

4

7:58 AM

@DavidCarlisle You forgot the main topics: ducks and pizzas!

@CarLaTeX how could I forget

@DavidCarlisle hahaha

Zarko

8:14 AM

@CarLaTeX, I just notice, that downvoting was reversed :)

@DavidCarlisle And you forgot the bashing of Germans for using long words without spaces ;-)

3

@ChristianHupfer whowoulddosuchaterriblething?

@Zarko :)

@DavidCarlisle :D That example does not count

Zarko

@AlanMunn, I took liberty and rewrote ray salmon question \newcounter v v v v v v v v v v v v v v v v. Hopefully is now more clear. To be honest, now I don't understand, what is wrong to use enumerate environments :-(

kan

8:24 AM

@DavidCarlisle LOL! I missed you all though!

I'll try to get back to being active here... it was a lot of fun.

8:49 AM

BTW, @Moriambar has just met Herr Professor Paulinho van Duck

@PauloCereda @UlrikeFischer ^^^^

could someone without texlive 2017 pretest try this

\documentclass{article}
\usepackage{adjustbox,tikz}
\usepackage{hyperref}
\showoutput
\def\showbox0{}

\begin{document}\parskip1cm
\newcommand{\rubylink}{\href{en.wikipedia.org/wiki/Ruby_(programming_language)}{Ruby}}

% text and hyperref link
1 abc \rubylink cde

\setbox0\hbox{\scalebox{2}{abc \rubylink cde}}\showbox0

% scaled by xetex.def code
2 \box0

\setbox0\hbox{\tikz[scale=2,transform shape]\node[inner sep=0]{abc \rubylink cde};}\showbox0

% scaled by tikz
3 \box0

\setbox0\hbox{%

and confirm you get

@DavidCarlisle I'll try TL'16

@DavidCarlisle Which engine?

@JosephWright xetex, sorry

9:01 AM

@DavidCarlisle Same in 16 and 17

@DavidCarlisle Different between 15 and 16

@JosephWright so that nests the graphics state 3 times with q/Q and reflects with a -1,-1 translatiion twice, the aim if the game is to simplify that in any way while keeping 4 working and not breaking 5....

@DavidCarlisle TL'15 ^^^

@JosephWright oh so scaling worked in tl15, hmm is that because we/I/someone changed xetex.def or had dvipdfmx changed (I am out of disk space for work reasons so had to purge a few non work things like old TL :-)

@DavidCarlisle Do you want me to work on this? There were some simplifications for xetex.def but I don't remember if they were related to scaling

@DavidCarlisle To be clear, the TL'15 result is right, yes?

@DavidCarlisle Also seems to apply to the expl3 drivers so they must need adjusting too

@JosephWright yes the current scale code scales the text but leaves the link annotation where it was. I was thinking of giving up (after pruning the tikz code down to that) and asking on xetex list but if you want to cast fresh eyes on it that would be good. I decided I'm not confident about the interaction of literal pdf with the pdf:bcontent special wrapper.

9:13 AM

@DavidCarlisle OK, I'll look at it today (I have some notes in the expl3 version about the way these different things work, or at least what I've observed)

@JosephWright OK thanks

@DavidCarlisle @JosephWright I tried with my system: in TL12 graphicx were good, and tikz bad, TL13-15 both good, and in TL16 and TL17 graphicx good bad. I could check the drivers later.

@UlrikeFischer thanks, it looks like we have enough good/bad test points now that we should be able to get it working again somehow:-)

@DavidCarlisle Yes. But it is curious: like you I thought that the displacement have been there always and so didn't check in older systems even when I couldn't find no old posts and questions. Prejudices are hard to avoid sometimes.

@DavidCarlisle Certainly: main issue is lack of real clarity in driver docs

@UlrikeFischer Could you check against pdfTeX for TL'15? There's something different about the starting point ...

9:24 AM

@JosephWright @DavidCarlisle: with pdflatex I get for all system "graphicx good - tikz bad" (and I checked the fls for local files ;-).

@UlrikeFischer Yes, but do you get a different size for the link outline: tighter height-wise, overlapping onto the "c" length-wise?

@JosephWright Look quite similar to me.

@UlrikeFischer I mean TL'17

@JosephWright The first image is TL15 the second Tl17 (both viewed in adobe)

Have to go now, if we should have something to eat over the weekend

@UlrikeFischer Yes, sure: I'm talking about:

XeTeX vs pdfTeX in TL'17

@UlrikeFischer Good plan!

@DavidCarlisle ^^^ Before any scaling at all

barbara beeton

9:33 AM

@TorbjørnT. -- sorry, don't know. i think i saw someone with a camera, but whether this was for an official recording or personal wasn't clear. nobody used a microphone though, so even if there are videos, i'm not sure the sound will be good in many cases.

@JosephWright yes but that's possibly Ok in that the links are driver-dependent breakable links anyway so no direct control over them from tex, be good if they were more similar though.

back in a bit...

@DavidCarlisle XeTeX one has changed since TL'15

@DavidCarlisle ^^^ Earlier version: I probably want to work out why that's changed before looking at the scaling, in case it has a link

Torbjørn T.

10:08 AM

@barbarabeeton OK, thanks.

10:55 AM

@JosephWright @JosephWright @DavidCarlisle yes, but the font has changed between both versions. You could use \XeTeXLinkBox{\rule{1cm}{0.5cm}} for a better defined link area. (The \XeTeXLinkBox is needed because of this (from the hyperref readme) "When XeTeX generates a link annotation, it does not look at the boxes (as the other drivers), but only at the character glyphs.")

^^ tl15 +xelatex

^^^ tl17 + xelatex

When using `\XeTeXLinkBox{\rule{1cm}{0.5cm}} I see no difference.

James Bond

@PauloCereda I was looking for you in this room. New username for me, lol. Quack! Quack!

Paulo Cereda

@JamesBond oh my!

@JamesBond: hello Jasper. :)

James Bond

@PauloCereda I just watched the first 20 Bond movies, 6 more to go! =)

@JosephWright @DavidCarlisle

^^ tl15 + xelatex

^^tl17 + xelatex

@JosephWright and ^^^ tl17 + xelatex + OT1-cmr font. So it is the font change.

11:49 AM

@UlrikeFischer thanks for the tests

12:03 PM

@DavidCarlisle @JosephWright @AlanMunn Is to berate a very common english verb (i.e. frequently used)? I think I encountered it for the very first time a few minutes ago. ... and no, nobody berated me ;-)

@ChristianHupfer reasonably common, but not of course in our household:-)

@DavidCarlisle Your wife does not berate you then for spending too much time on TeX.SE then? ;-)

12:38 PM

@DavidCarlisle Well the change which broke scaling was this here tug.org/svn/texlive/trunk/Master/texmf-dist/tex/latex/… but I can't find anything that would explain what did trigger the change.

@UlrikeFischer Did occur to me too: suggests something is not quite right I guess (but separate from the matter at hand)

@DavidCarlisle @JosephWright and I can't find the commit (43025) which changed \Gscale@start in xetex.def in the announce list: tug.org/pipermail/tex-live-commits/2017-January/thread.html. I start to wonder if it was meant be released -- someone should perhaps ask Karl.

@UlrikeFischer That was my suggestion, from memory: ideally the pdfTeX and XeTeX drivers can share quite a bit of code

Johannes_B

What would we do with that question? How to make the subsections bold in Springer LaTeX template for Data Mining and Knowledge Discovery journal

12:54 PM

@JosephWright I don't quite see were there is code sharing here: the definition of \Gscale@start in pdftex.def is quite different. And reverting to the older xetex.def definition solves the problem.

@Johannes_B Well, you gave the advice already: The editors/developers of the article style/class file/ do not want bold subsections, so don't do it...

@UlrikeFischer I had some recollection of a change around that time, but I was looking in the log of our github. I think there were some emails bouncing around of "pending chnages" when we took it over, I'll see what I can find..

@UlrikeFischer Not in those .def files but in the expl3 version, which is where I've been working on actually having some documentation, etc.

Johannes_B

@ChristianHupfer I cannot put that as an answer. The question as is is quite clear and can be answered, though the answer would need to add: Don't do it.

Q: Make my infinite squares math-proof

@UlrikeFischer @JosephWright * The second issue raised by the sx question was that even with expl3
or etex arithmetic the example failed with xetex. This turned out to
be an issue with xdvipdfmx which generates a non-invertible
transformation matrix when given very small numbers with its own
x:scale special. Joseph has just updated xetex.def in github to do as
pdftex.def does and use a literal postscript special to specify
scaling which then avoids the problem (or at least defers it: the pdf
renderer may have underflow problems with the specified scaling

27

I created a command that makes a square of text in ever-decreasing sizes to create the effect of infinite decrease. But if I increase the starting width, or change the scale factor, TeX has an arithmetic overflow or its memory is exceeded. I wish I had the math skills to understand why this is so...

12:58 PM

@DavidCarlisle Ah yes, that was it

So we traded working hyperlinks for fractal scaling, seems OK

@DavidCarlisle Yes, but the approach is right in pdfTeX, so there must be another place that the actual issue stems from (likely bcontent or similar)