« first day (3093 days earlier)      last day (1921 days later) » 

1:57 AM
This user literally knows how to start my new day.
2:17 AM
@JouleV Please don't talk about users in chat.
@JouleV Quoting @AlanMunn / @DavidCarlisle "Behaviours are only annoying if you get annoyed by them."
@marmot Sure, but that's not really the point.
@AlanMunn It sort of is because if you do not get annoyed then there is no reason to mention those anywhere like the chat.
@marmot Something tells me you haven't quite bought into this idea though.
@AlanMunn I think I have but I do not have to be annoyed when I contradict someone, do I? (The someone is no one from the chat, of course.)
2:26 AM
@marmot This has too many levels of indirection for me to follow. :)
@AlanMunn Remember: we live in complicated burrows. ;-)
@marmot Indeed. :)
3 hours later…
5:34 AM
@AlanMunn Even if this could not be said in chat, I was wondering why he was suspended. We have trolls free to do whatever they want, whereas this user seems only a serial just-do-it-for-me asker to me (@JouleV, @marmot).
@CarLaTeX I do not think this user got suspended.
@marmot He said so in a comment
@CarLaTeX Yes, I know, but I do not know if that ever happened.
@marmot Nor do I. That's why I'm curious
@CarLaTeX Let's say if he was then only for a very short time. Not 24h.
5:45 AM
@marmot Maybe he has still another account...
@CarLaTeX You mean on top of the one mentioned here? Maybe. (I think this is not our problem. That is, the mods will most likely be able to find out and take actions, if needed.)
@marmot Yes
6:01 AM
@CarLaTeX @marmot It seems to me that he is not suspended at all. A new account is deliberate though. It must be for another purpose
@JouleV It's also possible he thinks to have been suspended, whereas he was not
@CarLaTeX I agree (@JouleV) ... a lot of attention for this user....
2 hours later…
8:08 AM
@JouleV it is not at all uncommon for new users to end up with multiple accounts until they realise how the site login works. That's why theer is a page telling people how to ask that the accounts are merged
8:36 AM
@DavidCarlisle Well, it is not good to expand this conversation anymore. However, I'm sure the user knows how to login to SE sites. He had access to his first account right after posting that comment, and he even accepted one answer using the first account.
@JouleV perhaps (actually I saw his comment that he thought he was suspended after I made the comment above) either way as you say best to let it drop:-)
8:54 AM
@JosephWright @DavidCarlisle Is it intentional that \int_div_truncate:nn does not do the same thing as a/b in conventional programming languages?
@HenriMenke not sure there is a lot of consistency between languages, this rounds towards zero (which is certainly better then \numexpr / which rounds to nearest integer)
@HenriMenke what would you expect, always round down?
@DavidCarlisle Yes, I would have expected to round towards -inf (which is what Knuth mandates in TAOCP). Another problem is that because \int_div_mod:nn is implemented in terms of \int_div_truncate:nn it disagrees with a%b as well.
@HenriMenke C integer division % rounds towards zero for example, that surely is "a conventional programming language"
@HenriMenke fortran rounds towards zero, python rounds down....
@HenriMenke Example? We've tried a few different versions, I think (all predate me)
@JosephWright Lua
@JosephWright That is probably the most problematic candidate.
@DavidCarlisle I see. I wasn't aware that there exists such a fragmentation.
9:10 AM
@HenriMenke i meant example of the disgreement
@DavidCarlisle I looked at TAOCP and Knuth defines x mod y = x - y * floor(x/y).
@HenriMenke he defines lots of things:-)

\int_div_truncate:nn {-11} {3}


@JosephWright ^^^^
@JosephWright Needs Lua 5.3, i.e. LuaTeX 1.10.0
@JosephWright Or use math.floor(-11/3) instead of -11//3.
@HenriMenke page describing why python decided not to follow C here python-history.blogspot.com/2010/08/…
@DavidCarlisle nice reading, thanks!
1 hour later…
10:45 AM
\cs_set:Npn \__int_div_truncate:NwNw #1#2; #3#4;
    \if_meaning:w 0 #1
        0 \if_meaning:w - #1 \if_meaning:w #3 - \else: 1 \fi: \fi:
        0 \if_meaning:w - #3 \if_meaning:w #1 - \else: 1 \fi: \fi:
        > 0
       \__int_div_truncate:NwNwNN #1#2; #3#4; - +
       \__int_div_truncate:NwNwNN #1#2; #3#4; + -
    / #3#4
\cs_new:Npn \__int_div_truncate:NwNwNN #1#2; #3#4; #5#6
      \if_meaning:w - #1 #5 \else: #6 \fi:
@HenriMenke, @DavidCarlisle ^^^
@HenriMenke, @DavidCarlisle One for the team list I think
@JosephWright wouldn't it be better to have an additional \int_div_floor:nn?
@UlrikeFischer I was thinking the same, you can't really call it _truncate if it rounds down as that refers to just losing the digits after the decimal point (ie, round towards zero)
@DavidCarlisle \___int_div_magic
@PauloCereda gives always 🦆?
@PauloCereda \integer_multiplication_does_not_have_an_inverse_so_there_is_no_division:D
10:59 AM
@DavidCarlisle the colon is missing ;-)
@UlrikeFischer cricket terminology: round towards a duck.
@DavidCarlisle \int_div_round_towards_a_duck:nn would be a nice function too ;-)
@UlrikeFischer yes we like plump round ducks....
11:26 AM
@JosephWright Why so insanely complicated and not just
\cs_set:Npn \__int_div_truncate:NwNw #1#2; #3#4;
    \if_meaning:w 0 #1 ~
      (#1#2 -
        (\if_meaning:w - #3 ~ - \fi:
          #3#4 - 1) / 2
    / #3#4
@UlrikeFischer ooh
@DavidCarlisle If we remove _ , we get instant German :)
Please vote to reopen the following question, I want to answer ...
Q: Error emerged when i run the following latex code, please help me. Thanks

Ghulam Hafeez\documentclass[energies,article,accept,moreauthors,pdftex,10pt,a4paper]{mdpi} \usepackage{booktabs} \usepackage{array} \newcommand{\PreserveBackslash}[1]{\let\temp=\\#1\let\\=\temp} \newcolumntype{C}[1]{>{\PreserveBackslash\centering}m{#1}} \newcolumntype{R}[1]{>{\PreserveBackslash\raggedleft}m{#...

@Kurt done!
11:47 AM
@DavidCarlisle ooh
@DavidCarlisle oh no
Jun 29 '17 at 16:15, by Paulo Cereda
@DavidCarlisle you are not mean :)
@DavidCarlisle you are... oh wait
@CarLaTeX Thanks, added answer just now.
@Kurt just upvoted :)
@CarLaTeX Thanks, one more out of unanswered questions
11:57 AM
@Kurt 👍
@HenriMenke Fails if the demonimator is negative but the numerator is positive
@UlrikeFischer Maybe, use

\cs_new:Npn \int_div_floor:nn #1#2
    \int_value:w \__int_eval:w
      \exp_after:wN \__int_div_floor:NwNw
      \int_value:w \__int_eval:w #1 \exp_after:wN ;
      \int_value:w \__int_eval:w #2 ;
\cs_new:Npn \__int_div_floor:NwNw #1#2; #3#4;
    \if_meaning:w 0 #1
        0 \if_meaning:w - #1 \if_meaning:w #3 - \else: 1 \fi: \fi:
        0 \if_meaning:w - #3 \if_meaning:w #1 - \else: 1 \fi: \fi:
@HenriMenke ^^^ (@UlrikeFischer, @DavidCarlisle)
@JosephWright yes, still a depressing amount of code to correct for a bad etex definition of / :(
@PauloCereda @CarLaTeX ^^^ we got greetings from Merano
12:18 PM
@UlrikeFischer Oooh <3<3<3 Btw I love Meran I have been there many times
@UlrikeFischer ooh
12:38 PM
@JosephWright looks good (the floor variant). Do we need a mod-variant too? (@HenriMenke mentioned it)


expl3 mod

$\int_mod:nn {11} {3}$ ~
$\int_mod:nn {-11} {3}$ ~
$\int_mod:nn {11} {-3}$ ~
$\int_mod:nn {-11} {-3}$

lua fmod


lua mod

2 hours later…
2:13 PM
user image
I just got a palindrome
and a "successive" badge number ;-)
2:33 PM
@UlrikeFischer I think they got it wrong. Shouldn't there be a bunny?
@marmot A quick query, how do I add a dot in front of/after an alphabet? ( not like \ddot which is above)
@GermanShepherd $A\cdot B$?
I want it at the top left/top right..
the dot cant be at the center..
@GermanShepherd I really do not know what you're after so I can only guess. Does $\dot\phantom{a}A\dot\phantom{a}$ go in the right direction?
2:58 PM
hi there, anyone got some time to help out a latex newbie..?
@GermanShepherd is this math mode or text?
@dot_Sp0T that is the purpose of the site, but the best way is if you ask a question on the main question and answer part of the site.
Anyone has recommendations for profiling Lua code in LuaTeX, I found few things for profiling Lua code but either they don’t work with Lua 5.3 or don’t give very helpful output.
@DavidCarlisle Asking questions on the main is what i usually do, that is if I am sure I understand the subjectmatter enough to actually be able to ask a question - there's just so many terms and such that I am not sure I could ask correctly at the moment
But I'll try anyway then
@dot_Sp0T you can ask here but if it involves much code or work it's better to ask a question as more people see it then and there is better formatting possibilities
it's probably easier to drop a question on the main. I am trying to understand macro expansion I think - but I might just as well be hitting limitations of the packages I am using (tikz and xstring in this case)
3:02 PM
@KhaledHosny not really, I usually just end up adding print statements...
@dot_Sp0T xstring and expansion do not go well together:-)
@DavidCarlisle i feared so...
But it's the nicest thing i found for string operations
@dot_Sp0T hmm
@DavidCarlisle Not sure how print can help in finding the slow code.
@DavidCarlisle I'll type up a question with my boiled down code(?)
@KhaledHosny well that's true, my Lua is at the level that getting it to work is good, speed is rather secondary:-)
@dot_Sp0T if you want to use xstring usually you need to use the final optional argument so it puts the result in a macro then you can use that result macro in an expansion context.
3:08 PM
@DavidCarlisle yeah, but the there created macro doesn't seem to want to work together with other commands (in my case the argument to tikz \includegraphics )
@dot_Sp0T then you did something wrong:-) (\includegraphics isn't tikz)
@DavidCarlisle that is good to know, but it seems to be used with tikz (is it pgf?)
Also I am pretty positive I am doing something wrong, otherwise it'd work ;)
@dot_Sp0T no it's me:-) the graphics package, part of the core latex release.
@DavidCarlisle Right. I wouldn’t have worried about speed except that it is almost three times smaller than LuaTeX and XeTeX, so apparently that my Lua code is very inefficient.
3:42 PM
@DavidCarlisle mathmode
@marmot I can draw an image of what I want it to look like... but I guess chats dont support images.
@DavidCarlisle pushed the question, thanks for making me
There's a fine line between a numerator and a denominator. Only a fraction of people will find this funny.
3:59 PM
user image
before the jump ....
@UlrikeFischer PARTY
@UlrikeFischer Now the big question is how on earth you get 3 additional points.
@GermanShepherd You already have an answer. ;-)
@marmot yep :) I didn't know it was a bullet :P
Thanks @JouleV
@dot_Sp0T ok I'm about to go out but someone will answer, note it is better to show error messages as text from the log (in a code block) not as an image, in particular the real error message shows which command is undefined, which that image does not.
@UlrikeFischer I'm hoping to pick up 120 with 0 (or at least an even number) of accepts...
@GermanShepherd You are welcome!
4:10 PM
@DavidCarlisle ? don't you need 105? or are you after 500005?
@UlrikeFischer palindromes more important than multiples of 10, apparently
@marmot Literally no one can downvote our Bär, so she can't get 3 reputations ;-)
@UlrikeFischer Now you need 1093 points more
@DavidCarlisle You could ask a question at 500000, then you can have both ;-)
@DavidCarlisle totally forgot about the text logs
4:29 PM
@dot_Sp0T well you got lost in the expandafters. The first should be \expandafter\gappto\expandafter\matrixcontents\expandafter{\expandafter\consimg\expandafter{\templeft} \&}. But I would suggest that you use expl3 to build the command - much more readable.
@UlrikeFischer :(
@UlrikeFischer I have only learned about expansion (as well as read of expl3) today. So I chose to use the one that was easier for me to understand and replicate from tex.se posts :) Would you mind writing an answer explaining how to do it with expl3?
@dot_Sp0T why are you building the matrixcontent inside a node? looks quite unneeded.
@UlrikeFischer I have the image of the fish in a node above, and want to put the matrix content as a whole straight below the fish image. so to say as additional information. It seemed the best solution to me to do it like that. The whole construct of fish and matrix is then put in another node which I then use as part of a complexer graph
I've started working with latex the day before yesterday out of a mood. Started with images that i wanted to connect, learned about tikz, used tikz node to realize graphs of images connected by arrow. Expanded from there to nodes with additional information (this now)
@marmot thanks for the comments!!
1 hour later…
5:56 PM
@UlrikeFischer Hey, that's 200K!
user image
@egreg ;-) Yes, I jumped ...
@egreg oooooh
Q: Engine problem? Creating dynamic diacritics using the Brill typeface

Ruixi ZhangI’m trying to recreate the following image from John Hudson’s (principal designer of the Brill typeface) presentation slides: Source: The Brill Types, p. 22. The typeface is free for non-commercial use, please see the official Brill website. I got stuck in the image on page 21. Edit aft...

This is the kind of question/answer pair that make the site truly great.
6:11 PM
@AlanMunn someone hit the head on the keyboard :)
6:49 PM
@PauloCereda -- groan.
@PauloCereda Hahaha, something like that was my thought when I saw that question.
7:06 PM
user image
I got mail ^^^
@UlrikeFischer That's fast!
@UlrikeFischer oh no I go out for a bit and I've got +15 :(
@DavidCarlisle Hm. I could ask how to make a easter egg in picture mode in obfuscated code and then accept your answer.
@UlrikeFischer unfortunately I can only do clear code I'm not sure if I could write anything obfuscated
@DavidCarlisle an easter egg in picture mode in very clear code would be fine too ;-)
@marmot they obviously prepared the surprise in advance ...
7:25 PM
@UlrikeFischer \put(0,0){\oval(60,20){\makebox[0pt]{Happy Easter}}}
@DavidCarlisle an perhaps I even get a downvote as a "do it for me question" and then I would be in a palindrome cycle again ;-)
8:05 PM
@UlrikeFischer ooh you found a combofont question:-)
8:31 PM
@DavidCarlisle ;-) let's hope that not some other of your question answer gets accepted in the wrong moment.
@UlrikeFischer, @DavidCarlisle I see work on PDF features goes with issues in HO bundle ... I'm thinking about BDC: needs some of the page/global stuff
8:50 PM
@JosephWright yes in the long run, handling pdf resources should solve this too, but imho the current problem is simply that the test in transparent for \pgfutil@addpdfresource@extgs doesn't fit in some case - the pgf-file organisation is a bit complicated.
@UlrikeFischer Right, yes
@UlrikeFischer This is probably the point where we need notes, not direct code additions. I really do need to collect stuff up
@UlrikeFischer Perhaps I should do an expl3 release?
@JosephWright yes, I also think that transparency is a quite good target for a first try how to manage such a resources. There are only two package imho who try to set /ExtGState, and there are not many values.
@UlrikeFischer transparency and pgf?
9:06 PM
@JosephWright yes. I can do a search to check if someone else is involved, but I think not.
C:\texlive\2019\texmf-dist\tex>findstr /s /m "ExtGState" *.tex *.sty *.def
@UlrikeFischer ^^^
@JosephWright oh, a few more ;-).
@UlrikeFischer I'm reviewing them now ...
9:40 PM
@JosephWright animate and hypdvips seem to set only xobject resources, so can be ignored for the moment, lecturer is plain and the rest must be handled.
@JosephWright the code is already deleted. And my implementation couldn't deal with parenthesis and wasn't lazy.
@Skillmon It's tricky, isn't it :0
@JosephWright yep, I think I'll just stick with what Bruno's superior brain can think of :)
@Skillmon Like I say, mainly Morten I think ...
@DavidCarlisle ^^^ ???
@DavidCarlisle Or was that while you were 'away'?
10:05 PM
@UlrikeFischer I guess we want to cover all of Table 3.30 in the PDF spec
user image
@DavidCarlisle ^^^^^
@UlrikeFischer I always knew @DavidCarlisle likes ducks.
@JosephWright yes, mostly, ProcSet is deprecated, and Font is probably handled by the engines but the other should be covered.
10:24 PM
@UlrikeFischer So all dicts
@UlrikeFischer Looking over the pgf code: the issue is in having 'current page' and 'all pages', I think. One way or another one needs a per-page hook
@JosephWright well ExtGstate is "all pages". They reserve an object number and fill it at the end with all the collected stuff, and in all page resources they simply add the object reference. That looks quite sensible.
@UlrikeFischer Yes, true
@UlrikeFischer For dvipdfmx we need a hook for adding to every page
Wow! Two big achievements the same day! Congratulations to both @UlrikeFischer and @DavidCarlisle (notice the order)
@UlrikeFischer There are places we only want stuff on one page (see the business with removing properties)
@UlrikeFischer For example /Properties needs to be addressable on a per-page basis
@UlrikeFischer Bottom line: I think we need atbegshi in the kernel, at least the basics, plus a stub version for expl3 generic mode. I'll raise on team list tomorrow.
@JosephWright yes /Properties is more complicated, that's why I said we should start with ExtGstate ;-).
@JosephWright pgf seems to use \pgfutil@insertatbegineverypage, but it refers also to a comman object.: /ExtGState @pgfextgs. Color and patterns are similar.
10:59 PM
@egreg In many academic disciplines the most senior contributor is listed last. Your ordering suggests that math is one of them.
@marmot Well, no. ;-)
@egreg Maybe you just didn't notice? ;-)

« first day (3093 days earlier)      last day (1921 days later) »