« first day (3023 days earlier)      last day (1926 days later) » 

1:39 AM
@UlrikeFischer @DavidCarlisle @egreg @JosephWright @WillRobertson Just to say thanks for xfp, it really works great!
 
 
2 hours later…
3:38 AM
@AlanMunn I'm getting really old, you beat me by 43 seconds.
 
@marmot Well it's winter. You're in hibernation mode aren't you? And non-TikZ questions surely don't have quite the same allure.
 
 
3 hours later…
6:29 AM
@CarLaTeX Could you please try
\documentclass{article}
\usepackage{tikz}
\begin{document}
\begin{tikzpicture}[overlay]
\node at (current page.center) {X};
\end{tikzpicture}
\begin{tikzpicture}[overlay,remember picture]
\node at (current page.center) {Y};
\end{tikzpicture}
\end{document}

and tell me if I am seeing ghosts? I do only see the Y in the middle of the page.
@CarLaTeX What I want to say is that lyl has a good point, and I should have been more careful.
 
@marmot But you have only one page there
 
@CarLaTeX What fo you mean? (The result seems to be different depending on whether or not one adds remember picture.)
 
@marmot Ah, ok, now I see, it seems a strange behavior, however
@marmot I think it's a bug
 
@CarLaTeX I had noted it before, and forgot about it when writing the answer. In the pgfmanual all examples with absolute page positioning use remember picture.
 
@marmot Yes, but there's no reason why
@marmot You could ask a question: why remember picture is needed, here?
 
6:38 AM
@CarLaTeX I am not sure if it is a bug. The examples of the manual use it, so at least there seems not to be a bug in the sense that a statement in the manual is wrong. (And I can see why one should add remember picture is needed. It is also needed when one wants to use nodes from another picture, and that's what we are doing here: we use the anchors of the current page node.)
 
yo'
@CarLaTeX because TikZ uses a two-pass process to determine the position of the tikzpicture w.r.t. the current page. That's why you need remember picture.
 
@yo' Ah, ok, there's a reason!
 
yo'
@marmot even without it, your X is somewhere, just not at the right place...
 
@CarLaTeX The perhaps more interesting question is why you do not need to use it if you hook the picture in the shipout.
@yo' Yes, I know. The funny thing is that if you do not use remember picture and put the picture in the shipout, the stuff appears at the desired position.
 
yo'
@marmot because even at the moment of shipout, TikZ has to coordinate itself with the page, and it uses somenof the pdftex primitives to di this.
 
6:42 AM
The Stack Overflow developer survey is on. Even if you don't care about Stack Overflow, I hope they extrapolate the answers to the whole Stack Exchange network. Please participate and answer this question in a similar way:
 
yo'
@marmot I wouldn't ever rely on this :)
 
@yo' Yes, something like this. And I agree, which is why I added remember picture in my answer.
@HenriMenke Do you have a link for that?
 
@marmot Here is the link under the banner on askubuntu: askubuntu.com/dev-survey/…
@marmot Actually the banner hasn't appeared on Stack Overflow yet.
 
@HenriMenke OK, then I wait till it is official. (I just wrote some stuff where I let xfp do the computation in TikZ in order to avoid the dimension too large errors. Are there any plans to make something like this the standard parser/computational tool, or is this what l3draw is meant to be for.)
 
6:49 AM
@marmot PGF will never include expl3 and I don't plan to write a similar math engine. I might try to get a tighter integration of the luamath library though. That will also solve all dimension too large problem and is orders of magnitude faster than expl3.
 
@HenriMenke OK, great! Let's keep our fingers crossed that the arXiv supports lualatex at a given point....
 
@marmot I emailed them offering help. This is (part of) their response: “While it's likely that at some point we'll offer this as part of our compilation service, the timeline is unclear. I have forwarded your request to the development managers for eventual scheduling and deployment. If they are interested in deeper collaboration, they may reach out individually. Due to limited developer time, we cannot provide any timeline nor status updates for enhancements.”
@marmot They haven't gotten back to me yet.
 
@HenriMenke I guess that it is really a lot of work maintaining the arXiv.
 
@marmot Probably not, but I don't think anybody is working on the LaTeX installation full time. Most likely it's even volunteer work.
 
@HenriMenke Yes, that's what I meant. (It seems that there were some tensions among the moderators. One of them resigned.)
 
7:03 AM
@marmot I don't know anything about their internal structure.
 
Paul Henry Ginsparg (born January 1, 1955) is a physicist. He developed the arXiv.org e-print archive. == Education == He is a graduate of Syosset High School in Syosset, New York. He graduated from Harvard University with a Bachelor of Arts in physics and from Cornell University with a PhD in theoretical particle physics with a thesis titled Aspects of Symmetry Behavior in Quantum Field Theory. == Career in physics == Ginsparg was a junior fellow and taught in the physics department at Harvard University until 1990. The pre-print archive was developed while he was a member of staff o...
^^^^ he is very well known in my field and seems to be running the show.
 
@marmot Ah, high-energy physics. That's mostly unknown territory to me. I'm a condensed matter guy.
 
@HenriMenke Which is not too far away. He is doing lattice AFAIK. So in principle any type of QFT.
 
@marmot The problems I work on are neither relativistic nor conformal. Sometimes I even work on fluid mechanics which is purely classical.
 
@HenriMenke Lattice is not necessarily good for CFT. Anyway, he seems to be focusing on the arXiv now. In any case, I believe that lualatex may become the standard only after the arXiv supports it.
 
 
1 hour later…
8:33 AM
@marmot you shouldn't thank any of those people (Bruno did it all really:-)
 
8:49 AM
@DavidCarlisle OK, thanks, didn't know that.
 
@CarLaTeX It must be for your artistic skill in drawing that arrow. Exquisite!
 
@HaraldHanche-Olsen Surely that, lol!
 
 
1 hour later…
10:05 AM
@CarLaTeX Εάν ήσασταν εξειδικευμένος σε γλώσσες, όπως εγώ, θα γνωρίζατε την απάντηση από την ελληνική
 
@DavidCarlisle Non ho fatto gli studi classici, sono ragioniera
 
@CarLaTeX questo spiega perché non sai niente della pizza
 
@DavidCarlisle La pizza è un piatto povero, non servono studi alti per farla buona
 
@CarLaTeX naturalmente, basta prendere un po 'di pane povero che non è riuscito a salire e coprirlo con prosciutto e ananas
 
@DavidCarlisle giammai!
 
10:11 AM
@CarLaTeX avresti dovuto studiare i classici
 
@DavidCarlisle Il greco e il latino sono lingue morte, perché mai dovrei studiarle, quando tutto il mondo parla la lingua semplice delle popolazioni barbare?
 
@CarLaTeX La lingua di Shakespeare non è barbara, le tribù barbariche erano da sud di qui
My cat has come to check my Italian, although I'm sure it's all good.
 
@DavidCarlisle Good enough to be understood by an Italian :)
 
Is there a convenience macro to do \prop_put:Nnn \g_msg_module_type_prop { <module> } { <type> }? So something like \msg_set_module_type:nn? Would be a useful addition, to make things clearer, imho. It could as well be added to \ProvidesExpl...`...
So that \ProvidesExplClass automagically also does \prop_gput:Nnn \g_msg_module_type_prop { <class> } { Class }.
 
@Skillmon wouldn't you define a module-specific version in that module so \<module>_... rather than \msg_...
 
10:27 AM
@DavidCarlisle as far as I understood the property list is l3msg specific, to set whether to print "Package (<module>} Warning" or "Class (<module>) Warning".
 
@DavidCarlisle ooh a cat
 
@Skillmon oh yes sorry
 
@DavidCarlisle Sono inglese di Souldern
 
@DavidCarlisle so I think it would be a useful addition to add this to \ProvidesExplClass/\ProvidesExplPackage to also set this property. (since l3 often isn't about pure performance, the additional overhead for modules that don't print messages shouldn't be that bad, imho)
 
@PauloCereda and I thought you were a duck from Brazil:(
@Skillmon ask @JosephWright, I clearly know nothing:-)
5
 
10:30 AM
@DavidCarlisle you know everything!
 
@Skillmon that is google, not me
2
 
@DavidCarlisle I am a duck from Brazil. :)
 
@DavidCarlisle ohh
 
@Skillmon ooh
 
@PauloCereda ohh
 
10:31 AM
@Skillmon ooh
 
@PauloCereda Sono Non-Inglese di Non-Gerrard's Cross ;-)
 
@PauloCereda we should stop.
 
@Skillmon ooh a fixed point, so recursion stopped. :)
@ChristianHupfer I see what you did there. :)
 
@PauloCereda ohh
@JosephWright would you consider this a good idea?
 
@PauloCereda @Skillmon \end{ooh}\let\ooh\relax
 
10:33 AM
@Skillmon ooh
@ChristianHupfer oh :)
 
@PauloCereda sigh...
 
@Skillmon Could you mail LaTeX-L or the team? I think there migh tbe issues where the naming doesn't 'line up': @UlrikeFischer has one of those cases
 
@ChristianHupfer :)
@ChristianHupfer My laptop keyboard has a damaged key. Thankfully I can remap it. :)
 
@PauloCereda The windows start button is unnecessary, though ;-)
4
 
@ChristianHupfer ohh
@JosephWright haven't thought of this. In this case, it might be better to just not mail it :)
 
10:37 AM
@ChristianHupfer :)
@JosephWright Dear LaTeX-L, I am not dictating WHAT
From Skillmon dictates a letter sketch :)
 
@ChristianHupfer on the keyboard? No! This one is really important, as it is Super!
@PauloCereda ?
 
@Skillmon Monty Python....
 
@Skillmon it's a Monty Python sketch. :)
 
@Skillmon I won't defile the concept of a Super Key by assigning it to Windows Start Key -- that's blasphemy
 
@ChristianHupfer @PauloCereda didn't know that one.
 
10:46 AM
@Skillmon It's from the last season of MP, the one without John Cleese ... most of it is bad, in my point of view
 
@ChristianHupfer how unfortunate.
@ChristianHupfer I have to find the time to watch flying circus, etc.
 
@Skillmon It's on Netflix, for example ... and YouTube anyway... totally legal ;-)
 
@ChristianHupfer don't have Netflix. But maybe the latter suggestion will do :)
 
@ChristianHupfer Did you grow up using one of these?
 
@HaraldHanche-Olsen ooh there are Playstation keys
 
10:54 AM
user image
3
@Skillmon: That's CatFlix ;-)
 
@ChristianHupfer do they catflix and chill?
 
@Skillmon Yes...
@HaraldHanche-Olsen Nope
 
@ChristianHupfer LOL
 
@PauloCereda Note the Apple computer with pink unicorn stickers /facepalm
 
11:12 AM
@ChristianHupfer oh
 
 
2 hours later…
1:14 PM
@ChristianHupfer @PauloCereda Pink unicorns? The letter A in the title and subtitle of the book is more awful. Low centre of gravity and so not nice kerning.
Apropos unicorn: my daughter misses such a tikzling.
 
@StefanKottwitz Oh, what a pity @samcarter is not here!
 
@StefanKottwitz There are two X360 controllers!
 
1:30 PM
I got the next palindrome:
user image
5
 
Current status
2
 
@PauloCereda Impressive aT kerning.
 
@StefanKottwitz :)
 
user280247
2:14 PM
Hi guys, should I ask on the main site about problems installing vanilla texlive on ubuntu 18?
 
2:39 PM
@santimirandarp sure
 
2:53 PM
@StefanKottwitz Huh? Which book?
 
Hi. How can I draw this?
Feeling stuck...
 
@ChristianHupfer The one on the photo.
 
@StefanKottwitz You seem to have plenty of time to scale such photos in order to detect weird kernings etc. ;-)
 
@ChristianHupfer It's Saturday. Though the day of the week doesn't matter for procrastination.
 
@JouleV Part VI of the TikZ manual seems like a good start.
 
3:02 PM
@StefanKottwitz I don't know what procrastination is...
 
@ChristianHupfer You don't know what you're missing.
 
3:29 PM
@JouleV scatter plot in matplotlib (if you want to use Python and not directly TikZ)
 
3:56 PM
@JouleV I am sure such plots are on the main site. I try to find one.
@JouleV This is one example: https://tex.stackexchange.com/a/348418/121799
According to my vague memory there might have been an earlier plot without lualatex.
 
@JosephWright I think I found a real deficiency regarding the primitive \pdflastlink.
 
4:19 PM
@UlrikeFischer ?
 
@DavidCarlisle I'm trying (for tagging) to get the object numbers of the annotation objects of a link. But if a link is broken then there is more than one object, and \pdflastlink gives back only the first.
 
@UlrikeFischer well.. since it's an integer it can't do a lot more I guess... but some way to find the numbers would be good, are the object numbers for a broken link predicatble (if you can see it is broken?)
 
@DavidCarlisle no the follow up numbers are not predicable. As far as I can see the first number (from \pdflastlink) is reserved directly, but then e.g. font objects and the page object is created. So I have e.g. 1, 6,7,.... The numbers are later on in the Annots entry of the page but one can't see which belong to which link.
@DavidCarlisle I'm just writing a question on the luatex list (it is the same problem there).
 
4:52 PM
Is it possible to insert new line in \PackageError ?
 
@Sigur yes
@Sigur \MessageBreak (it prefixes the newline with the package name)
 
@DavidCarlisle, nice. Thanks.
@DavidCarlisle, I am trying to format it also. Is there some command for blank char?
 
@Sigur yes: x x or if you are in expl3 mode ~ of course, or\@spaces is 4 spaces, and \space is one
 
@DavidCarlisle, good
 
\def\@font@info#1{%
         \GenericInfo{(Font)\@spaces\@spaces\@spaces\space\space}%
                     {LaTeX Font Info: \space\space\space#1}}%
@Sigur ^ typical example (actually not such a good example (font) just needed here as there is no font package so it's using the lower level generic infoo, but \packgezzzz is same really
 
5:12 PM
@DavidCarlisle, \@spaces was useful to me. Many thanks.
 
5:22 PM
@DavidCarlisle I'm holding my breath if we broke luaotfload or not ;-)
 
@UlrikeFischer did you see the "question" earlier about luaotfload?
 
@DavidCarlisle yes, that was what I meant. But without more info it is impossible to decide if there is a real problem or not.
 
6:08 PM
@JouleV And if you do not want to use lualatex:
\documentclass[tikz,border=3.14mm]{standalone}
\begin{document}
\begin{tikzpicture}
\draw[clip] (0,0) rectangle (8,4);
\foreach \X in {1,...,1000}
{\pgfmathsetmacro{\myx}{rnd*8}
\pgfmathtruncatemacro{\itest}{ifthenelse(2*rnd<1+sin(\myx*300/pi),1,0)}
\ifnum\itest=1
\fill(\myx,4*rnd) circle[radius=1pt];
\fi
}
\end{tikzpicture}
\end{document}
 
6:40 PM
@UlrikeFischer I notice that you did ask the question, plus the related one on unmatched \Uleft or \Uright delimiters, on the luatex list. Thanks! I look forward to seeing the response.
 
 
2 hours later…
8:50 PM
@Circumscribe Your smuggling really works great! (You do not know by chance how to confine a definition to a given environment, like e.g. the tikzpicture I am currently in, right?)
 
9:14 PM
@marmot: Shouldn't something like \pgfkeys{foo/.store in=\foo,foo/.initial=blue} result in a predefined \foo expanding to blue? tikz reports \foo to be unknown... (I don't think this is an group issue)
 
@marmot Good to hear!
 
@ChristianHupfer No, use \pgfkeys{foo/.store in=\foo,foo=blue} instead (if you really want to go that way).
 
I suppose you could maybe keep on smuggling until \@currenvir changes… (wouldn't work if an environment is nested in another one of the same type)
 
@marmot That's what I am doing. What is the purpose of .initial then? Just for .get handling?
 
@Circumscribe OK, thanks! Seems a bit too complicated. (Sometimes one has to guess the right order... )
@ChristianHupfer You could do \pgfkeys{foo/.initial=blue} and then access this with \pgfkeysvalueof{foo}.
 
9:18 PM
@marmot Or hook into \end; not sure if that's better or worse :).
 
@marmot Ah... that's nice... pgfkeys ist ein Buch mit sieben Siegeln für mich... (noch ;-)) Danke
 
@Circumscribe Nahh. I do not think it is worthwhile. If there was a standard LaTeX way, then that would be a different story.) Thanks!
@ChristianHupfer I would agree that the discussion of the pgfmanual on this is not yet optimal, and I am not sure what I am saying is correct or the intended usage, but I think it works...
 
@marmot Sadly, it does not work or I am doing stupid mistakes... something like foostyle/.style={foo} (with foo as suggested by you above), is too optimistic to work, isn't it? (Compilation fails)
@marmot: Ah, got it: Since my foo key is from /foobar family, I've to use foostyle/.style={/foobar/foo}...
 
Or
\documentclass{article}
\usepackage{tikz}
\begin{document}
\begin{tikzpicture}
\pgfkeys{/tikz/foo/.initial=blue}
\tikzset{foostyle/.style={\pgfkeysvalueof{/tikz/foo}}}
\draw[foostyle] (0,0) -- (2,1);
\end{tikzpicture}
\end{document}
@ChristianHupfer ^^^
@ChristianHupfer But you're right, \tikzset{foostyle/.style={/tikz/foo}} is better.
 
@marmot Well, now I have to teach the /tikz keys to accept my /foobar path as well, without prepending /foobar/ each time
 
9:33 PM
@ChristianHupfer @ChristianHupfer search also?
 
@marmot \tikzset{/tikz/.search also=/foobar} then?
 
@ChristianHupfer The manual says \pgfkeys{/main path/.search also={/secondary path}}
@ChristianHupfer I didn't try this out so far, but if you have a key family this might work. From what I read this is what ".search also" is made for.
@ChristianHupfer Since @Circumscribe is around (at least the profile text says so ;-) : I am also wondering if one could smuggle pgfkeys. The only way out I know is to make the key global with \globaldefs1 (see e.g. here) but this can have severe side effects.
 
9:50 PM
@marmot Making an observation may actually change my aroundness.
 
@Circumscribe Are you a cat?
 
@marmot I wouldn't know directly, but I think these are just stored in pgfk@<path>
 
@Circumscribe Yes, that's what the error messages suggest that I often get. ;-)
 
@marmot Just a boring human I'm afraid :(.
 
@Circumscribe But at least you seem to share a crucial property with Schrödinger's cat.
 
9:56 PM
@marmot I'm part of the system!
 
@Circumscribe Not an observer?
 
@marmot You're the only observer; everything else is part of the system :).
 
@Circumscribe That makes sense. ;-)
 
@marmot I do believe I'm part of the system though, and that all of reality as I currently see and remember it is just a single observation.
 
@Circumscribe Yes, this is consistent with your earlier statement. ;-)
 
10:26 PM
@marmot not in front of the newcomers:-)
 
@DavidCarlisle OK, sorry, deleted it. (I was just amused that you got asked if you could recommend some packages for tables... ;-)
 
@marmot oh i don't mind really, but it may be confusing following the line of reasoning in this chat is confusing enough if you are english, i have no idea how ducks and marmots manage
 
@DavidCarlisle Reasoning? What reasoning? ;-)
 
Anyone here using pandoc to convert to MSWord? I have a question.
 
@marmot anyway you could have referred him to this recommendation:
If you use any of the mentioned packages other than longtable then any bug or problem reports don't go to me. That is a definite advantage of using those packages. — David Carlisle May 9 '12 at 18:49
 
10:33 PM
@DavidCarlisle ^^^^ really?
@AlanMunn Others have been excommunicated for less. ;-)
 
@marmot the question is asking about longtable v supertab as far as I recall
 
@marmot Yeah, well I have to deal with a system that requires me to upload Word documents, and I'm try to outwit it with pandoc. I don't think that counts as an excommunicationable offence.
 
@AlanMunn Die Wege des Herrn....
 
@DavidCarlisle hm, most of the comments in the old question seem to recommend (long)tabu ;-)
 
10:36 PM
@marmot oh yes the current question, tell him to wait unto Joseph gets round to doing an l3table module
@UlrikeFischer you'd better fix tabu then:-)
 
@DavidCarlisle So I shouldn't have deleted my comment, then? ;-)
 
@marmot Yes, Our Father, IRB, hallowed be thy name.
 
@AlanMunn The implication is that it is not the offender who can decide whether or not an offense is excommunicationable....
 
@UlrikeFischer no reply to my fr.comp.text.tex post :(
 
@DavidCarlisle well Jean-François had to wait more than two weeks, so there is still hope.
 

« first day (3023 days earlier)      last day (1926 days later) »