@HaraldHanche-Olsen If you want a more intimidating insect avatar, maybe you could go for this: sciencefriday.com/articles/…

@UlrikeFischer It might be out of date, sorry. I've had a lot of trouble making tests that show this sort of thing consistently. You wouldn't expect things to change but...

tikz duck with india flag

@WillRobertson could you run a check on this test if it still passes for you?

@AlanMunn Yes, I know about the tarantula hawk. My avatar comes from my own photo, though, and I don't intend ever to get close enough to a tarantula hawk to photograph it. So I guess that idea is out. Actually, all my life I have suffered some phobia against stinging insects, but I knew this one was harmless, so no problem.

Strange behavior today with MikTeX. I load a file into TeXworks, and it won't compile, saying it's improperly configured. I note that the little window where it indicates what engine you are using is totally blank and the TeXworks Preferences tool, under the typesetting tab, shows zero processing tools available. Do I need to reinstall MikTeX?

@StevenB.Segletes Probably the configuration files have got zapped

@StevenB.Segletes I'd just re-build them: what does TeXworks tell you under Help -> Settings and Resources ...

Is it a normal quirk that amsmath symbols and text don't align perfectly in an align* environment?

@yo' Hi!

@JosephWright I just had my IT guy reinstall. Oh well. But I think I discovered the issue, and that is that our organization customized OS, in the past few days, will now not let any executables be run out of user directories...everything has to be in Program Files. It is an insane policy. What about the output of compilers, for example?

@PauloCereda hello!

@JosephWright (I'll be making some "picky" comments on the sys stuff) Just a thought, might make sense to restrict rm by forcing the argument (in unix) to starts with ./

@yo' o/

@PauloCereda sorry I have my phone landscape for faster typing and the ENTER gets in the way...

@yo' no worries. :) Have fun today, pal!

@JosephWright same thing with mkdir. And there could be the unrestricted versions, but under a longer name. In linux this still allows accessing the parent via ./../this-is-the-parent, I'm not sure how it works in windows though

@PauloCereda have to find an open Correios, the closest one opens at 930 it seems :-/

@StevenB.Segletes We have the same ... but can get 'special permission' to get round it ;) Our computing department are entirely exempt!

@yo' Oh I think using ./../../foo.bar is allowed, so it might be tricky to restrict folders... :(

@yo' How would that work with absolute paths?

@nathdwek No. Do you have an example you could show us?

@yo' Ah yes, it's the standard opening time (I thought it would be 9:00), since Correios is a federal institution.

@yo' That works on Windows too, but I'm not sure I see what you are aiming to avoid: probably one for some examples on GitHub

@yo' oh you just mentioned that. :)

Sorry.

@Ha

@HaraldHanche-Olsen: mmmh I ended up right aligning and It's ok like that

@yo' I did say 'first pass': one issue I have is that from TeX my room to manoeuvrer is limited, at least without a lot of shell scripting: compare the l3build functions, where error level and file type are both easily accessible. (I could arrange to return error level, but it seems rather over-the-top.)

Other question though, in a sans serif document, how do I write a roman italic math symbol?

mathrm give roman upright

and I can't combine with math it

@JosephWright well, it would not, that's the point. For absolute paths, you would have another ones.

@nathdwek Oh, you meant horizontal alignment? There are subtleties there: After the alignment tab &, the environment adds an empty group {}. Its purpose is to ensure that if a relation symbol comes after it (as is usually the case), it adds the usual space that relation symbols should have between them and ordinary math atoms. I guess that is what you ran into.

@PauloCereda @JosephWright I don't want to restrict folder access, but I wanna be on a safe site, thining e.g. of rm -f /etc instead of rm -f project/etc (which would easily happen if some tl wasn't set up properly and was empty) but thr thing is that as long as you don't call rm -rf, it's ok

@HaraldHanche-Olsen Ok just for completeness sake: i.imgur.com/9i9fYiM.png

The code is &\min_{\alpha_c, \alpha_d}&&\sum_{j=1}^{n}

@yo' oh I see!

&\text{s.t.}&&\text{M

(I dropped the end of each lign as it does not pertain to the questiono

@yo' if it's not much trouble, I poked you in Hangouts. :)

@PauloCereda @JosephWright with ./project/etc, you end up with .//etc = ./etc and you only mess up the working directory

@yo' huummm

@nathdwek The \min and \sum constructs are operators. When they come after an ordinary math atom, including the empty group {}, a thin space is inserted in between. If you add \ , before each \text (i.e., write &\,\text{…} the alignment should be better.

@HaraldHanche-Olsen Thanks! I'll remember for next time

for the moment I actually prefer it aligned on the other side

@nathdwek Sure, that probably makes sense in many cases.

@yo' I wouldn't like that loading a package that aims to provide support for cleaning the working directory leaves me with an empty directory because of some bug in the package.

@egreg nobody would I think :) but it's better than being left with empty hard drive :)

It's just an additional level of safety measures. Anyway I'm out

cya

@yo' I guess my issue is we could have an 'explosion' of functions: in the end, these are all meant to make people's life a bit easier. If they get too complex, that doesn't work. For example, there was quite a bit of discussion for l3build about absolute vs relative paths, and we ended up going for absolute ones.

@yo', @PauloCereda If there is a split, I guess I'd make it between rm -f (equivalent` and just rm: that's much less tricky than relative vs absolute path

@MdAyquassar -- this will take some mulling. i shall mull and report later.

@nathdwek -- an explicit example would be very useful.

@nathdwek -- okay, @HaraldHanche-Olsen has explained why the sum is set a bit to the right of the text below it. i assume you don't have any problem with the fact that the subscripts in the limit below \min aren't on the same baseline.

@barbarabeeton @nathdwek And if you would like to correct it: \min_{\alpha_{c\vphantom{d}},\alpha_{d}} will deal with it. It makes the first subscript seem as tall as the second one, pushing it down to the same level.

@JosephWright My opinion is that this is something for a package, rather than for the kernel.

@egreg The entire idea?

@JosephWright I guess so: the idea is supporting filesystem operation, isn't it?

@egreg Basic ones, yes: to me it fits naturally into the scope of the sys module

@JosephWright This introduces safety/security issues, whatever you do. I'm not sure we want to inflict them to all users.

@JosephWright Also calling gcc fits in the sys module, with this approach.

@egreg :)

@egreg You can only call the shell if you actively enable it, so 'out of the box' any such functions are no-ops

@JosephWright Why not embedding shell scripts or Python scripts? Maybe integrating PythonTeX… ;-)

@egreg I do take the point: there's a reason new stuff goes in as experimental

@JosephWright The shell is enabled for several other unrelated reasons (minted, for instance). On the other hand, it's quite easy to add programs to the “restricted shell”: just define the relevant variable in the environment.

@barbarabeeton No problem. I'll also need time to see whether the answer suggested by Circumscribed is free from adverse effects. (Other than that, my conversation with Mico is removed now, so, I'm afraid, you won't see old comments any more.)

@egreg The moment you allow unrestricted shell commands, one can have 'issues'

@egreg I've picked the set I have as the aim is to have 'platform-neutral' support for them: almost everything else is already of that form (e.g. running a script)

@egreg Perhaps one for the team list :)

@MdAyquassar -- i've already seen the comments, and they did make sense.

@egreg One could just go with the platform test, plus perhaps the 'turn / into \' business for Windows, and document how end users might use such functions (@yo': thoughts?)

@JosephWright @egreg I would like some "standard" command for pwd/ls (to replace e.g. the currfile package). But imho it should work in restricted mode. Couldn't one add a small script to texlive which allows to call both without (like cd) allowing also to change the directory and which could be added to the restricted list? Or extend pdftex by a new primitive to get this ;-).

@UlrikeFischer I've just commented on the team list that pwd is quite doable without shell escape: you can pick that data up if -recorder is active

@UlrikeFischer I'll probably add something to the file module for absolute path, and the extension of the top level file: both are done without shell escape by currfile

@UlrikeFischer ls/dir are I think pretty safe so might be OK on the restricted list

@UlrikeFischer I'll forward you a couple of messages

@JosephWright I know that's how currfile is doing it. I was only wondering about a version without this requirement.

@UlrikeFischer Doable, of course

@UlrikeFischer Something like @echo %cd% would do it, I guess (saved as rcd.bat)

@JosephWright Imho the main problem with --recorder is that you have to teach people to use it. Also it only works at the second compilation.

@UlrikeFischer Seems to work on the first past for me ...

@UlrikeFischer Hence my team list question: it would presumably be easy enough to get this enabled as-standard

@JosephWright yes, but it feels a bit like a work-around to read and analyze or more or less long file to get something like the current directory.

@UlrikeFischer Sure: it would certainly be easier if this data were exposed directly (also for the full file name of the top-level file). The question is whether adding a primitive is really worth it. Most of the time, the absolute path goes with doing other things which do need shell escape.

@UlrikeFischer You only need the first line for pwd, and finding the full top-level name is also relatively easy

@UlrikeFischer Perhaps I should have CC'd you on my team list mails :)

@DavidCarlisle I see Arthur is about at least to some extent

Everyone suffering from the European heat?

@JosephWright 25 and a bit cloudy here in Toronto. :)

@AlanMunn Ah, much more reasonable

@AlanMunn Rio was much more comfortable than Norwich has been this week!

@JosephWright Well it was the middle of winter. It's very good that the conference wasn't in December.

@AlanMunn If TUG2019 is in North America, are you likely to go?

@AlanMunn Yup :)

tug.org/tug2019

^^^ Got to plan ahead ;)

@JosephWright I don't know. The last one in N. Am. was in Toronto, but I still didn't go, because it was a busy time for me, even though Toronto is close and home (my mum still lives here). So it's unlikely I'd go to SF for it.

@AlanMunn Ah, shame, but I understand

@AlanMunn I'm not sure if I'll go to the West Coast ... depends on who else is going

@JosephWright If it were in Chicago, then yes. (3 hr drive from Michigan).

Do you think writing a package based on tex.stackexchange.com/a/434521/117050 is worth it/something people would need?

@JosephWright 29C here

@JosephWright it has become pleasant. In the afternoon it was too warm.

@Skillmon Are you sure you can't simply use a tcolorbox for the spanning part and a tcbraster inside? It seems as if this is fully covered by tcolorbox…

@TeXnician except that the job is way simpler than what tcolorbox is good for and if one just looks for this simplicity my solution is way faster than loading TikZ and tcolorbox. Also: How do you make a tcolorbox span the whole page, with the contents being inside of the margins (never done that before)?

@TeXnician I'm talking about the last piece of code in that answer, not the stuff in the first two code boxes.

@Skillmon See vvv

\documentclass[]{article}

\usepackage[most]{tcolorbox}
\usepackage{graphicx}
\usepackage{duckuments}
\newcommand*\picandtext[2]{\begin{tabular}{@{}c@{}}#1\\#2\end{tabular}}

\begin{document}
\noindent
\begin{tcboxedraster}[raster columns = 3, raster equal height, enhanced, frame hidden,
	segmentation hidden, interior hidden, fontlower=\centering\color{white}\Huge\bfseries]{
	enhanced, spread inwards, spread outwards, colback=blue!25,
	frame hidden, boxrule=0pt, arc=0pt, left=20mm, right=20mm}
	\begin{tcolorbox}

@Skillmon Okay, maybe it's a sledgehammer to crack a nut, but definitely not bad…

Good evening to everybody.

I ask with such kindness if someone could help this new user. Thank you very much.

@TeXnician I never doubted tcolorbox is "not bad", although I seldom use it I still like it. But it still is slow compared a few lines of mixed plain TeX and LaTeX stuff :)

@Skillmon Yes, but of course once you learn to love it you tend to do any colorboxing with it… Of course it's up to you whether you want to write a package for that ;)

@TeXnician Writing the package is easy (since the code is already there, it would only be making some options accessible and writing a short documentation). The question is: would you consider it something somebody might use in the future?

@TeXnician as far as I understood the question, the contents should stay inside of the margins. You got my upvote nonetheless.

TUG2018 report written

@TeXnician there is a spread sidewards key in tcolorbox as a shorthand for spread inwards, spread outwards.

@PauloCereda ^^^^^ the bear and his new friend.