@DavidCarlisle Sure thing!

@PauloCereda we have expectations ^^^^

@DavidCarlisle He ducked away.

@egreg clearly unreliable

@egreg is the latex3 question answerable? (I decided to pass:-)

@DavidCarlisle I too.

@egreg @JosephWright can take it in the morning:-)

@egreg actually I left a comment not sure it'll help.

@JosephWright Do you mean for the unicode letters setup? I can't imagine any problems there. unicode-math.sty will currently change things anyway, but I'd be happy to delete that code if it's no longer required…

@yo' Maybe now :) Are you still around?

@AndrewCashner Weather catastrophe? Y'all have no idea!

My back hurts from the shoveling the snow!

@AlanMunn I wonder how is Michigan doing with the snow? #ONStorm

@kan We have almost no snow at all. We got about an inch last night and it was quite warm today (around -2 C)

@AlanMunn oops, I went out to shovel more... :)

@DavidCarlisle I tried with a generic sales pitch

I would like to know if that is a normal github issue: github.com/posquit0/Awesome-CV/issues/31

GM's article in the latest TUGboat is fantastically entertaining (ping @DavidCarlisle and @JosephWright).

@Johannes_B Hmm, I would guess not :)

@WillRobertson Freaking templates ;-)

@WillRobertson Yes, I'd seen it during production (was asked to look it over, and write something slightly easier to access about the 'romannumeral trick')

@JosephWright Any ideas you'd like to add to expl3? :) The more \exp_not:n-wrapped functions we have, the more sense it makes to re-introduce the "d" expansion. (Although maybe with a different letter to indicate "result" or similar)

@WillRobertson Example where d would be needed (other than the two cases in l3keyval that I had to take out during the change)

@WillRobertson, @DavidCarlisle Speaking of keys, I really should look at the performance of l3keys as it's poor compared to pgfkeys and the new options package

@JosephWright I'll get back to you…eventually :)

@JosephWright But the easy answer is everywhere that the result is returned in an \exp_not:n :)

\tl_set:Nd \l_myarg_stripped_tl { \tl_trim_spaces:n {#1} }

(Can't actually remember if that function can be converted into a form that uses 2-expansion resulting, but that's the general use case -- define a macro in terms of the result of a function.)

@WillRobertson I had a question about pstool's interference with DeclareGraphicsExtensions -- is there any option to make it work properly? It seems to me that I need to declare both eps and pdf as allowed extensions for the main document -- is that correct?

@yo' Ahhh... I haven't looked at that code for a long time. I can't remember what it was supposed to do; I have a feeling that the package pre-dates the auto-EPS conversion system, so I never considered that EPS would be used directly.

@WillRobertson needless to say, the package is a saviour (for those of us bond to pdflatex), once you learn to use it. Never mind, it's probably not a big issue, I just wanted to make sure I don't miss anything...

@yo' I'm glad to hear it still works!!! I made some updates with dubious utility at some point that I thought made it much more fragile.

@WillRobertson it has some pecularities (like that my class checks for "missing maketitle" and other things, pstool of course breaks this). I would have actually welcome a switch like \ispsfragfig which would allow me to easily check whether it's the main document or the child process and modify the header accordingly, but it's not a big deal I think...

@yo' That's a good idea :) But couldn't you just do \ifpdf ?

@yo' I guess \ifpdf is not as clear and can't be easily made a no-op if necessary

@WillRobertson I could probably simply check for the value of \pdfoutput, right? :D

@yo' Still, I think it's a good idea.

@WillRobertson Acutally, good, but not 100% helpful; with \ifpdf or \pdfoutput, I could do the check for the documentclass as well, which is impossible with \ifpsfragfig since it won't be defined before \usepackage{pstool} in the main document

@yo' True :) I'll continue to think about it. Don't expect anything soon!

github.com/wspr/pstool/issues/27

@WillRobertson I don't, as I said, I solved my problems, I just wanted to make sure I don't overlook some simple sophisticated way :)

@yo' What were you generating the figures in originally? I'm a bit worried that the latest versions of matlab have broken eps output for use in pstool.

@WillRobertson whatever the authors of the articles I typeset/copy edit send :)

it's mostly that they use psfrag and latex-dvips-pspdf which is impossible in my workflow

@WillRobertson thanks. I commented there with an idea of how to solve it smoothly.

@yo' I could also add a way to just use a different preamble...

I have to admit, I get very confused when I see your real name and it's not yo.

@WillRobertson that's way to complicated for everybody I think. It's IMHO clearer to have an if switch

@WillRobertson neither is my real name. Sorry for that :(

@yo' haha, why sorry? :)

@yo' Your real name was shown on the email that github just sent me

@WillRobertson oh a triple confustion! /me likes :D

@yo' @WillRobertson it's all my fault that psfrag doesn't work with pdflatex:-) I should probably see how easy it would be to re-enable the original mechanism. The original idea was ghostscript on the EPS to obtain the coordinates of the labels, and then use a picture-mode to overlay the text at the coordinates. I changed it to do it all in one pass by injecting some postscript code to do the label positioning in the main postscript interpetation so no gs was needed, that's what breaks pdftex

@DavidCarlisle but how could it work in pdflatex when pdflatex does not support EPS figures?!

@DavidCarlisle I'd first do a survey to see how much EPS is generated these days and whether it's worth the time :) My anecdotal feeling is that fewer and fewer tools are supporting that workflow.

@yo' I'm guessing he means that the EPS is auto-edited to suppress the labels and then is converted to pdf in -shell-escape ?

@WillRobertson I only said I should probably do it, I wasn't actually planning to do anything:-)

@DavidCarlisle Good :)

@WillRobertson needless to say, many scientists love EPS simply for psfrag: they do not properly label their axes in the external tool they use, they simply put a,b,c,... as labels of things, and then apply psfrag's tricks :)

no I'm sorry, it's lunchtime for me (exceptionally early today), see you later

I have to go home for dinner as well… I've worked late by accident again

@yo' @WillRobertson pdflatex will shell out to ghostscript to do eps topdf conversion. It could, at that stage inject the psfrag header so the eps text was not typeset and the coordinates were written out. Then pdftex could input the generated pdf and overlay latex text, so the overhead would be no more than current eps inclusion

@DavidCarlisle oh, right.

@DavidCarlisle I had to do a lot of work in order to repair colortbl “features”. :)

@RomainPicot: coin coin!

@PauloCereda Quack :)

@RomainPicot ooh <3

@PauloCereda Did you see the last phdcomic ? :)

@RomainPicot I did. :D

@PauloCereda and your thesis draft speak to you or no? :P

@RomainPicot Actually my thesis draft is singing this song to me:

:D

@PauloCereda Do not feed your thesis with Cookies !

@RomainPicot coooooooooookieeeeeeeeees

:)

@PauloCereda only with chocolate of course

@RomainPicot :)

Duck: a device for turning chocolate into thesis.

@yo' not only duck turn chocolate into thesis :)

@RomainPicot in my case it was coffee really... and cocoa

@yo' you're a mathematicien so coffee of course :)

@yo' awwww <3

@RomainPicot duck(x) ∧ likes(x, chocolate) → thesis(x)

@PauloCereda ooh <3

@RomainPicot If I manage to attend a TUG conference, I will wear a T-shirt with @yo's sentence. :)

@PauloCereda You can wear a T-shirt with this sentence even without being in a TUG conference :)

@PauloCereda oH! :)

@RomainPicot oooh

btw, I surely won't be in Toronto: it ovelaps my summer camp

@DavidCarlisle, @WillRobertson I see some developments are happening with XeTeX: we should think about a 'serious' chat with Jonathan K.

@JosephWright, @yo': I missed the cue yesterday, but hopefully it's time to redeem myself!

:)

@egreg about time you did something useful

@JosephWright yes that's been progressing on the branch for a while (not sure what I did but I get emailed commit log messages it seems:-)

@DavidCarlisle And I didn't even blame you in my answer. Maybe I'll edit it.

@DavidCarlisle If I want to include files with a package, so that when loading the package, you can then insert these files in the document, how should I do that? Is there any guide I can read?

@Alenanno images? or other tex files?

@DavidCarlisle In this case images yes. Should I build a system of directories so that when, if, I upload it to CTAN, the files are available?

@Alenanno either way just need to make sure they are installed as part of the package not just as documentation or source, easiest is to make a tds.zip file so that texlive knows how to unpack. see for example how the mwe package makes example-image available (even if you do not use that package)

@DavidCarlisle Thanks, I'll look into that. I didn't know what package to explore to see how they do it.

Is tds.zip the standard for "folder with images" in this case?

@Alenanno no need for a complicated directory tree, mwe for example puts all teh omages and latex .sty in the same directory.

@DavidCarlisle I see. Thanks for the info, I think I'd still use a separate directory because I think it's more organized, but it's good to know it still works.

@Alenanno it should be an archive in the directory "tex directory system" layout as used by texlive/miktex etc rather than the ctan layout shich mixes documentation and sources in a flat structure. see the TDS entry here: ctan.org/pkg/mwe?lang=en

@DavidCarlisle I see it uses .eps... SVGs are still a pain in Latex, aren't they?

@Alenanno you need separate directories if you distribute fonts as they have a separate search path but images are searched along the TEXINPUTS path, the same as package files, so actually having one directory is simpler in many ways

@DavidCarlisle Oh.

@Alenanno svgs would be as easy as eps if people routinely wanted to generate svg, dvisvg can just copy in an svg while generating an svg for the whole page, just as dvips can copy in an eps while generating postscript. But if you want to input svg and write out postscript or pdf, you need to have a convertor in the loop somewhere

@DavidCarlisle .eps are way easier to handle. I noticed that when I was inserting a logo in a document of mine.

@Alenanno yes but I guess you were not generating an svg for the whole document?

@DavidCarlisle I don't think so? What do you mean "for the whole document"?

@Alenanno if you run latex to get a dvi file then you can use dvisvgm to generate a series of svg (one per page) if using that driver including svg files into the document is easy as they can be more or less just copied in. that is the natural analogue of using latex and dvips which can easily include EPS files as it is generating postscript so the EPS file does not need to be understood, just copied into the generated output.

@DavidCarlisle Ah ok, then no. Just one small logo. :D

@Alenanno so you want to input svg and output pdf, which means you need inkscape or some such program has to be in the loop to do the conversion.

@DavidCarlisle I'll have to take a look at that... :)

@DavidCarlisle So slow. ^^^^

@DavidCarlisle No wonder I have a golden tikz badge and you don't.

@egreg but when I do get mine it will be well deserved

@DavidCarlisle @egreg Yesterday @yo' helped me with the following code. Why does it work, since it seems that \delfin has not ever been defined. Clearly the fact that it appears as part of a delimited \def enough, but I'm not sure why.

\documentclass{article}
\makeatletter
\newcommand\delsc[1]{\d@sc#1\delfin}
\def\d@sc#1.#2\delfin{#1.\textsc{#2}}
\makeatother
\begin{document}
\delsc{foo.bar}
\end{document}

@AlanMunn well, when TeX core sees \d@sc, it looks for the first ., then for \delfin. It never expands these two tokens

If it fails to find one of these, it gives an error "use of \d@sc does not mach it's definition"

@AlanMunn You can use moreorless any token, you can have things like \def\d@sc#1.#2\d@sc{...} -- then when you call \d@esc a.b\d@sc, the second one is never expanded.

@AlanMunn it doesn't have to be defined, it just has to be there to delimit the argument

According to pgfplots manual (1.13, p.48) the folowing is supposed to work. Does it for you guys? If not, is it a bug?

\documentclass{article}
\usepackage{pgfplots}
\pgfplotsset{compat=1.13}
\begin{filecontents}{\jobname.dat}
	T      I1       I2
	305   0.158   0.184
	310   0.147   0.175
	316   0.135   0.163
\end{filecontents}
\begin{document}
\pgfplotstableread{\jobname.dat}\EXPDATA
\begin{tikzpicture}
	\begin{axis}
		\addplot table [x=T, y=I1] {
			\EXPDATA
		} ;
	\end{axis}
\end{tikzpicture}
\end{document}

! Undefined control sequence.
\EXPDATA ->\pgfpl@@

@DavidCarlisle Same here.

@Johannes_B stick to my packages, they don't have bugs.

@DavidCarlisle I always do, but the OP tried something else ;-)

@Johannes_B I have no sympathy for such recklessness.

@Johannes_B You have spurious spaces. Try \addplot table [x=T, y=I1] {\EXPDATA}

@David How's it going?

Hello ;)

@UlrikeFischer spurious spaces ... Thank you.

I must write an answer on this site one of these days!

@yo' @DavidCarlisle Thanks. It's kind of counterintuitive.

Is there a simple way to change the font of minipage footnotes in basic article class? It is easy to change the font of the minipage, but the footnotes still come out in the default rmfamily.

@AndrewCashner footnotes do \reset@font so if you have a sans serif minipage don't just do \sffamily do \let\familydefault\sffamily\normalfont

@AlanMunn not if you have the correct intuition

@kan fighting luatex (again:-)

@DavidCarlisle Thanks! I had to do \let\familydefault\sfdefault\normalfont

@AndrewCashner Which is the correct code.

@DavidCarlisle :)

@AndrewCashner yes, you are supposed to do what I meant, not what I wrote.

@DavidCarlisle Teach a man to fish with a slightly broken fishing pole, and after he learns to fix it...

@AndrewCashner yes I'm sure that was my intention, an educational device, not just incompetence when answering.

I am writing a tutorial on citation style for my undergraduates, but thanks to biblatex-chicago, all I need to remember is \autocite. I wonder, which would be harder, teaching them Chicago-style citations, or LaTeX?

@DavidCarlisle oops, good luck!

@barbarabeeton you could try my nag.co.uk address?