« first day (844 days earlier)      last day (2799 days later) » 

12:25 AM
@Speravir <3
@PauloCereda :-)
@egreg +1 for not using tikz to insert a tabular (and bothering to answer the question when the OP couldn't be bothered to use less than half a million packages)
12:52 AM
hello all! :-)
don't notice he's gone
1:28 AM
Ooh, Looking for “^” symbol is suggested to be a duplicate of How to typset the symbol “^” (caret/circumflex), what looks reasonable first. But this is itself closed as a duplicate … And now?
You can vote to close with the other duplicate and the one that gets the most close votes of the 5 wins, I think. But the best duplicate to link to is the one that seems like the best answer overall, I think.
@leo Paulo Psmith is too unreliable. I read mail-archive.com/ctan-ann@dante.de or actually have the RSS feed of it subscribed.
@ Speravir I think the duplicate link is actually the better one in this case, since it does give a direct answer, and then links people to how to look up any symbol, which in our collective naïveté we think people will pay attention to. :)
@AlanMunn Ha, I just was going to asking something similar. @tohecz should add his comment then, though, into that answer or as an own one.
(Own answer will not work …)
4 hours later…
5:58 AM
Does anyone know much about this? It still seems to be a problem (I am using \hat{c} and getting this error).
Q: Internal error: bad native font flag (XeLaTeX, fontspec, newtxmath, libertine, \dot)

QrrbrbirlbelI'm trying to use \dot{P} in a math environment. MWE: \documentclass{scrartcl} \usepackage[no-math]{fontspec} \usepackage{libertineotf} \usepackage[libertine]{newtxmath} \begin{document} \begin{equation} \dot{P}(t) + 1234567890 \end{equation} \end{document} Without the command...

2 hours later…
7:28 AM
@Speravir This is of course not a duplicate, since for symbols like ^, more explanation is needed IMHO:
Q: How to typset the symbol “^” (caret/circumflex)

Costi Possible Duplicate: How to look up a symbol? I need to display the symbol '^' d <- dist(fascores, method = "euclidean")^2 How do I do that?

7:59 AM
Is there a symbol in LaTeX for a set of matrices eg. Mat_(m,n)(F) ?
8:41 AM
@dexterity Not a predefined one; you can look for \DeclareMathOperator on the site; essentially, you can do \DeclareMathOperator{\Mat}{Mat} and then using \Mat as you would use \sin. Requires amsmath
9:04 AM
Anyone have an opinion on whether or not I should delete my answer in tex.stackexchange.com/a/99054/86 ? When I answered the question seemed to be about tracking whether or not a command had been used but it would appear that the issue is more complicated so I don't think my answer is really relevant and could be said to be just adding noise.
@AndrewStacey It's the fault of the qner for being unclear. I'd leave it, with the comment you have made, as an advertisement to the nuisance of having effectively communicated requirements change in this way.
Wrt that question: I have the leanest implementation of the updated requirements and have no upvotes...
Maybe I should add a linje \uspackage{expl3} to make it more bloated
@CharlesStewart did you test it?
No. I can't on my netbook, it just has Contet minimals. When i get to my desktop.
Contet is the lean way of writing ConTeXt Mk. IV
@CharlesStewart in my understanding of ifx the conditional as you wrote it should always be in the false block
Might well be. I'll see soon enough.
You mean it should be \if, I guess. That sounds right, now I think about it.
9:21 AM
no i mean you can't compare a macro a with a macro b that holds macro a. From tex by topic
\def\a{z} \def\b{z} \def\c1{z} \def\d{\a}
\ifx\a\b %is true \ifx\a\c %is false \ifx\a\d %is false
While \if expands until there are two tokens and compares those
Which sounds like the behaviour I want, but the cs's I define won't contain any tokens, so they need to be populated.
I'm not going to think about this until i have my code in an emacs buffer
@CharlesStewart two unexpandable tokens yes but \def\a{} \def\b{aabb} \if\a\b 1 \else 2 \fi makes bb1 as the two a compare true, which isn't usually what you want
9:42 AM
(surprised that &apos; is allowed in XML, but not in HTML)
\def\analyze #1{\begingroup
%\expandafter\def\csname #1 \endcsname{#1}% if comment always false
\expandafter\ifx\csname #1 \endcsname\analyzename %
\textsuperscript{\textcolor{black}{ #1}}\textcolor{red}{\sout{#2}}%
ifx does not complain about undefiend macros
and destroing the \analyzename is not necessary since the defdoes not survive the group.
and of course it should be \def\endanalyze{\endgroup}
w/o #1
@topskip poor people those web coders ;)
@bloodworks I like it much more than TeX coding :)
9:58 AM
@topskip i hate it because i don't like multi engine coding
@bloodworks you mean something like \ifluatex ... \fi? :)
@topskip if IE...
@bloodworks sure, but that's much the same as with TeX. Multiple engines that behave differently.
but at least the tex stuff should behave similar in luatex pdftex etc. So the IE guys didn't even try to catch up with the standard (in the past... when i did more html than today. I was told that even IE got better)
but for html good news came form scandinavia. At least one hasn't to care about presto any more ;)
@bloodworks But the same with TeX: there are libraries that hide the engine differences
Using jquery for example helps a lot.
10:06 AM
@topskip oh yes i never work with but i remember a article praising it somewhere
speaking of html how is texfragen.de doing?
@topskip It is allowed in html5 and so in all current browsers (for some version of all)
@bloodworks We'll have a FAQ session on the DANTE meeting - hopefully we get more q's and a's.
@DavidCarlisle Interesting. I've only found w3.org/TR/html4/sgml/entities.html - but non for HTML 5 yet (only a page listing visual representations - and there it is. dev.w3.org/html5/html-author/charref)
10:28 AM
Now that PS4 is announced, I should try to install TeX into my PS3. :)
@topskip Hixie grabs that list from my sources anyway so just use this file (which is also the file firefox uses internally as well, or at least I think it is I had long enough conversations with their lawyers confirming they could use it)
I've found a small error there: "Please report any errors to David Carlisle
via the public W3C list www-math@w3.org." should be "Please report any errors to David Carlisle
via the tex.stackexchange chat" :)
@PauloCereda isn't a ps3 a hex core?
@DavidCarlisle Thanks for the pointer!
@bloodworks I'm not sure.
10:34 AM
@bloodworks too bad TeX only uses one
@topskip can lua distribute tasks?
@bloodworks Yes (coroutines). Not sure if the implementation in LuateX allows more cores.
@topskip could be useful... my forecast is that if luatex will have a certain performance it will become #1 plotting tool in the whole round world. (maybe even in the univers)
11:05 AM
@bloodworks Thanks for the feedback, I've fixed the answer. It's leaner and more readable just to \edef the \csnames, rather than use \expandafter, though it does cost a line of code. See tex.stackexchange.com/a/99159/175
@bloodworks Ah, a bug: there's a spurious "analyze" at the end. So how come gobbling an argument with \def\endanalyze#1 causes junk to appear? I guess I've gobbled a gobbler cs.
And it is now not interestingly different from Peter Grill's answer: it uses \ifx to compare csnames rather than \IfStrEq on macro contents. Is it easier to understand? If not, should I keep it anyway?
11:24 AM
@CharlesStewart The code is wrong: \edef\analyzename{\csname#1\endcsname} with #1 equal to eddy, just defines \analyzename to expand to \relax (unless you give a definition to \eddy). And two tokens defined in this way will always be equivalent.
If you do \def\analyzename{#1} then the thing is different.
@CharlesStewart Look at this
  \ifx\test\analyzename % Check name (\1) matches \analyzename
    \textsuperscript{\textcolor{black}{ #1}}%
\title{Latex Document}
As part of this project, we intend to study the complex structures of Latex.
\deleteline{eddy}{This line will be stroked through.}
11:47 AM
@egreg @egreg if the names passed in correspond to let equivalent macros the test will be true but if they are undefined they wont \edef\a{\csname zzzz\endscsname} and \edef\b{\csname xxxxx\endscsname} aren't ifx equivalent even though both the inner tokens are let to relax. (The code's wrong anyway but for different reasons:-)
@DavidCarlisle Oh, yes! There's an \else that does exactly the contrary (and a spurious spaces that I didn't see. But using \csname seems very risky for no real gain.
@egreg well yes, just because the code is wrong, doesn't mean you are right though:-) So I'm still feeling rotten but can cheer myself up knowing someone's bound to star a comment saying I found an egreg bug.
@egreg I can't remember if we considered making hspace use \setlength. I know I spent quite some time searching the sources to find all primitive length register assignments and deciding if they should use setlength, but of course that isn't an assignment so might have missed the net, or we might have decided it was too slow or risky a change
@DavidCarlisle There are better ways for this: \hphantom{text}; I don't think somebody really wants a \hspace{\widthof{text}} that disappears at page breaks.
@egreg well true but they may want \hspace{\textwidth-6cm} of course they can have that now anway with \dimexpr but in general I think we claim that user-facing length arguments in latex should take calc syntax
12:05 PM
@DavidCarlisle Doing an assignment with (calc's) \setlength every time one has to do \hspace seems quite inefficient.
@egreg Yes, I wonder if calc should do it (at least as an option). Probably not worth it given dimexpr...
@DavidCarlisle It might interest a couple of users.
@DavidCarlisle Think of the juicy rep points. :)
I'm trying to improve the input syntax of this matrix-ball answer. There's no reason why the location needs to be specified as an (x,y) coordinate instead of a single index from 1-9 corresponding to the cells of the matrix read left-right/top-bottom. But TikZ's picture origin is in the lower left corner, which makes the conversion a real pain. Is there some clever method I'm missing?
@AlanMunn Where's the backgammon package? :)
12:19 PM
@AlanMunn why is the indexing 0 based in one axis and 1 based in the other? As my day job involves translating between fortran 1 based and C 0 based arrays I'm used to either convention but having both in the same array hurts:-)
@AlanMunn can't 1,1 be top left and 3,3 be bottom right?
The code I'm working on (not the code in the answer) converts the numeric 1-9 index with row = mod(index,3)+1 and col = index-row/3.
@DavidCarlisle That's the way the TikZ coordinates worked out.
@DavidCarlisle And that's what I'm trying to fix.
@AlanMunn well if it's really a "matrix" having 2 indices is probably more natural but matrices are always indexed from 1 starting top left. If it's just an arbitrary data layout then of course whatever indexing scheme works works:-)
@AlanMunn well you just need the linear relation between documented index and whatever tikz need here at (#1.5,#2.5) but I don't know what constructs you can put at that point in tikz syntax. I could ask @egreg I suppose.
@DavidCarlisle Try random words in random places
12:27 PM
@DavidCarlisle From what I gather the matrix part is always 3x3 so it would make more sense just to use a linear index.
Right now I can do the linear index but it starts from the bottom left corner and goes bottom-top/left-right (like the coordinates.)
Psmith, the TeX bot, in fixed font mode: Let's take a look at the last fencing results:

- Oh dear, nothing happened so far.

Our fencing expert Alan might explain these results later on.
Another exciting day in fencing world.
So another way to pose the question is given an indexing of
1 2 3
4 5 6
7 8 9
how to I convert that into an indexing of
3 6 9
2 5 8
1 4 7
@AlanMunn ask on cs.SX or math.SX ;)
12:44 PM
@tohecz That's just sad. I'm in a room full of people who should be able to do this in their sleep. ;)
@AlanMunn My last discrete-mathematical problem to solve was a polynomial bijection between $N^k$ and $N$, for a fixed $k$ of course. Despite it was couple weeks ago, I'm still a bit tired from that ;)
@AlanMunn Do I understand that right, for a given number between 1 and 9 in the first scheme you want to get the equivalent number in the second scheme? Or the x and y indices in a grid?
l = [[1,2,3],[4,5,6],[7,8,9]]
print(list(zip(*[x[::-1] for x in l])))
Have fun with Python.
@Jake Well I want to convert a number in the first scheme to the indexes used in the matrix-ball answer. In its current state (because TikZ uses the lower left as an origin) the natural indexing is the second matrix. So if I could convert from one indexing to the other I think that would also work.
[[1, 2, 3], [4, 5, 6], [7, 8, 9]]
[(3, 6, 9), (2, 5, 8), (1, 4, 7)]
12:55 PM
@JosephWright Is it a good idea to nest the \si declarations with per-mode=fraction? Example; if I want to write Newton meters per rad/s, I've tried \si[per-mode=fraction]{\newton\meter\per\si[per-mode=fraction]{\radian\per\second}}. I don't know if it's a feature, but then I get the first fraction turned into a slash which is quite nice. Is that dependent on the size of the denominator?
Alternatively can I tell TikZ to assume (0,0) is in the top left corner and grow down and left.
@AlanMunn 3 - (n-1):3 + 3*((n-1) mod 3) where : is truncated division and n is the current index. It should be easy to get it with pfgmath
@AlanMunn You can use -\x and -\y depending on the direction.
@AlanMunn \pgfmathparse{(mod((\i-1),3)+1)*3-floor((\i-1)/3)}
(gives the same result as Paulo's Python Program)
@Jake Which is the same as mine suggestion
1:08 PM
@Jake Perfect. That's what I was missing.
@egreg Yes, but @Jake supplied the correct random words. :)
@percuße: sir get online. :)
@PauloCereda Don José is singing to Carmen about "la fleur que tu m'avais jetée".
Ok. Now I can update the answer with a nicer interface. Does anyone here use this matrix-ball stuff?
@egreg oooh! :)
@egreg Oh yes, indeed
1:10 PM
@PauloCereda Poor guy. Nothing to do against Escamillo.
!!/answer closed form of 3, 6, 9, 2, 5, 8, 1, 4, 7 ?
@DavidCarlisle Hold on, my browser crashed. :)
@PauloCereda If you will use IE
!!/answer closed form of 3, 6, 9, 2, 5, 8, 1, 4, 7 ?
@PauloCereda Psmith, the TeX bot: 3, 6, 9, 2, 5, 8, 1, 4, 7, -270, -1887, -7554, -22951, -58588, -132295, ...
@DavidCarlisle (-480 + 1822 n - 1055 n^2 + 241 n^3 - 25 n^4 + n^5)/(12 (23 - 10 n + n^2)), according to Wolfram Alpha =)
1:13 PM
@DavidCarlisle oeis.org/A131579
@Jake yes I know I checked the form of question that WA would accept first, but PSmith dropped the ball again:(
@DavidCarlisle Hehe
@egreg yes I was testing PSmith not looking for an answer:)
Now a TeX question. The commands are going to be of the form \single(index){value} or \double(index){value}{value}. I can implement this with () to delimit the index or I could use (or abuse) the optional argument, so that the syntax would be \single[index]{value}. The latter uses `\newcommand` instead of a delimited argument, but the index isn't really optional, so I feel like I'm abusing the syntax.
@AlanMunn If it's mandatory, use braces
1:15 PM
@AlanMunn if you'd used 2 indices you could arguably have use (,)syntax but as it's a mandatory value why not {} why make it optional?
@egreg But I'd like to separate the index from the content in the actual command syntax. So I don't like \single{index}{value} as a syntax (although obviously that would work.)
@DavidCarlisle I'm trying to separate visually the index from the value(s).
@AlanMunn if you want to visually distinguish you could use tikz syntax something like \single[/some key/.extra dot/.index/value=4]{value}
@DavidCarlisle Or I could use picture mode. :P
@AlanMunn \def\single#1#{\@single{#1}}\def\@single#1#2#3{...}
@AlanMunn Yet another one
\begin{tikzpicture}[ball node/.style={draw,circle,inner sep=.1em,minimum size=2.5ex}]
\draw[style=help lines] (-1,-1) grid[step=1.01] (2,2);
\foreach \x[count=\xi from 1] in {-1,0,1}{
\foreach \y[count=\yi from 0] in {-1,0,1}{
\node[ball node] at (0.5+\y,0.5+\x) {\pgfmathparse{int(\xi+3*\yi)}\pgfmathresult};
1:23 PM
@percuße Thanks, although the problem isn't drawing the diagram, but providing an interface to the commands that uses a natural index.
@AlanMunn Sure and you can name the nodes instead of drawing them with the resulting number
@percuße Yes, sorry, that's true.
@AndrewStacey Hmm. That's cute.
What do others think of @Andrew's suggestion?
\single1{value} \single2{value} etc.
@AlanMunn Hmm. Non standard syntax.
@egreg Very. But it achieves my "visual typing" requirement.
1:29 PM
(by 'typing' I mean in the CS sense).
@AlanMunn Oh, hang on. I missed the fact that there's only 9 squares so you could just do \def\single#1#2{...} (also got confused between \single and \double in number of arguments) and call it as \single 9 {value}.
@AndrewStacey Yes, I got that. :)
Maybe I should just go for the standard \single{index}{value} and \double{index}{value}{value}. But I like the idea of separating the two types of argument in the markup.
1:53 PM
@tohecz -- Legs are doing good!
@kan glad to hear that!
@PauloCereda I need to read up this bit about using zip. I am not yet comfortable. :)
@kan List comprehension. :)
@PauloCereda This is where the manual describes it? Not yet googled -- figuring out a rule for derivative of the kth iterate of a function.
@kan I'm not sure.
1:59 PM
@PauloCereda Okay! :)
I'll look it up though.
@kan: in other news, there's a nice indian chant (sorry for the possible misuse of the word) playing in my office. :)
@PauloCereda Hmm, any links so that I can listen to it? :)
@kan Let me see, a friend of mine sent me the link. :)
!!/texdef -t latex bibliography
Psmith, the TeX bot, in fixed font mode: Here's the output from texdef:

macro:#1->\if@filesw \immediate \write \@auxout {\string \bibdata {#1}}\fi \@input@ {\jobname .bbl}
2:01 PM
And, I know an Italian prayer to say when I am hungry:
Scroll down on this page: nh7.in/indiecision/2010/12/21/…
@PauloCereda First up, you used the word in a right way, IMHO. :) And, thanks for sharing, sounds interesting. I have come across quite a few recorded by a guy in US on lord Krishna.
@kan I'm back to my Yoga meditation. :) After my T'ai chi ch'uan, of course. :)
@PauloCereda :)
Well, BBL going for dinner. Hopefully you enjoyed the hungry prayer. :)
@kan It works, I'm hungry. :)
2:22 PM
rm -R /Users/tobig/Library/Application\ Support/Steam hello disksspace galore
@bloodworks o.O
Oh no!
@egreg "using \csname seems very risky for no real gain." - Oh, quite, but a most pernicious excuse to waste time. I've run into a quite surprising error hacking csnames.
Anyone seen the following error:
1. .aux file has contents: 00000000 5c 72 65 6c 61 78 20 0a |\relax .|
2. On loading the .aux file, pdflatex hangs with:
[Loading MPS to PDF converter (version 2006.09.02).]
Context mkii file? From pdflatex? Why?
@CharlesStewart It's for intepreting MPS files (the product of Metapost).
This is an error I get from the latest revision (not working yet, with risks \csnames) of my answer to tex.stackexchange.com/questions/99030/…
Should be no MPS files.
@CharlesStewart because Hans had already written a metapost interpreter in tex and I couldn't be bothered to write another:-)
@CharlesStewart yes but it loads it anyway (it's not that big)
2:37 PM
Hans has bigger fish to fry
Must go
3:14 PM
Hey @DavidCarlisle
@N3buchadnezzar hi
I know you are a table man, heh. Mind looking at a picture of a table and give me a few hints on how to make it look good?
Not asking about latex, but rather aesthetics. Can not for the life of me decide on a style to use.
@N3buchadnezzar just ask, there're more people with opinions ;)
@tohecz Heh, sure! I just thought nobody else was on, my bad!
@N3buchadnezzar a timetable?
3:22 PM
longtable man, longtable man
Doing the things longtable can
What's he like? it's not important
longtable man

Does the package have a \dot, or even a spec?
If there's a conflict, is longtable a threat?
Or should we remove the package instead?
Nobody knows, longtable man
Almost, a list of the subjects I have taken. Red is mathematics, blue is physics, green is computer science, and orange is pedagogy.
@PauloCereda so, when will the "First book of TeXpoems" be published?
@tohecz We will come up with a TUGBoat article. :)
@N3buchadnezzar So the columns don't have really a significant meaning?
I would say it's good. If it was a timetable, some more vertial "leaders" would be probably ok (timetable is not really a table). And I would not be afraid of making some information bold
My problem is that when I have two subjects next to each other, they blend in a tad too much.
But I guess that does not matter as much.
3:26 PM
@N3buchadnezzar That's why I was asking whether it's a timetable. because in that case, I would add all the vertical lines and finito
Vertical lines over horizontal lines? Wut
(maybe first time in a long while when I defend vertical lines in a table)
@N3buchadnezzar in a timetable? yes, why not? but only with full cells, not between whitespace
well, thin white vertical lines would "seperate the cells", and they would not really be vertical lines
Do all the things! If I can I will try to add white lines, hmm
@N3buchadnezzar Well, this is certainly wrong. I'm not sure how to make white lines tho
I tried changing the colorsep part, but alas nothing.
3:56 PM
Please, close (see last comment)
Q: enumitem: newlist with description and bullet

navititiousI've seen quite a few questions here on how to use enumitem with descriptions and bullets. I couldn't however find a solution where they created a newlist and that's where I have my problem: I defined in the preamble \usepackage{enumitem} \newlist{codescription}{description}{1} \setlist[codescr...

4:10 PM
@PauloCereda hmmmm
@DavidCarlisle <3
@N3buchadnezzar \arrayrulecolor{white}
In the preamble or after \table ... ?
@PauloCereda Awesome Poem! @David You never FTBA.
@N3buchadnezzar get as many crossing horiz and vertical lines, preferably each of a different colour and post the result to any question with
@N3buchadnezzar well anywhere depending on what scope you want
@kan had to look that up, I assume you meant this
4:18 PM
@DavidCarlisle LOL
A clipped version of DFTBA -- Don't forget to be awesome.
@kan yes so the urban dictionary said but I'd have had to look up the long version as well.
I am increasingly suspecting that David has an acronym matching strategy that takes in context and churns out funny stuff.
Q: Is there any solution to include Java Applets in presentations/beamers?

EliasI am preparing a presentation on the compiler Texmaker "documentclass" in "beamer". I was fascinated with the package "multimedia" which is a solution to include movies in the presentation. Question: Is there any solution to include Java Applets in presentations/beamers or a link to webpage? Th...

I'm pretty sure this is a duplicate.
@PauloCereda and / am pretty sure this is a bad idea
4:35 PM
Getting there =)
@Andrew I have been meaning to ask you, how do you remap Caps lock key to, say tab? I did it with the xmodmap tool and wrote the output so that it is saved for all sessions. It still does not seem to have changed.
The behaviour gets mixed. The first tap is taken for a tab but also mistaken for Caps Lock and it gets toggled.
I remember a recent (one or two weeks) question which used lua to generate a latex table, if I recall correctly, but I can't locate it now. It would be interested in finding some examples which use lua to implement loops which generate tex output, in order to translate to lualatex my answer to the question: tex.stackexchange.com/questions/99256/…
Is this really a duplicate?
Q: List of mostly all quotation marks used

m0nhawkI'm seeking of a list of quotation marks used in different language. So far I found only this wiki-page. But there is only the marks itself in Unicode. I know about csquotes package. I just want to find a list of all this marks as a LaTeX "commands", like ,,, << etc.

it seems to me that he seeks for a (typography) big list of different quotation marks used in different languages.
5:04 PM
Is there any way to do something like \cmidrule[0.5em]{3-7} in latex?
@N3buchadnezzar What should that mean?
I know I can combine \bottomrule and \toprule with a specified width, but using it with \cmidrule messes things up.
It should mean, make a line in a table from row 3 to 7 with width 0.5 em.
@N3buchadnezzar Width=thickness?
Yes! I figured it out though
Seems the syntax is different for \cmidrule and \toprule
I need to do something like \cmdidrule(l{5pt}r{10pt}){3-7}
@kan Try this:
clear lock
keycode  66 = Control_L
add control = Control_L
@kan It sounds as though you might not be doing the clear lock.
5:09 PM
@AndrewStacey Well, is clear a flag for xmodmap? How do I do this? put this in the .Xmodmap?
@N3buchadnezzar \cmidrule[wd](trim){a–b}
@kan Yes.
@tohecz I tried that, but it then ignores the {a-b} part, care to look at the source?
therefore: \cmidrule[\heavyrulewidth](l{5pt}r{5pt}){3-4} should work
5:18 PM
@AndrewStacey Hah, OK.
@N3buchadnezzar what? no \cmidrulewidth exists, only \heavyrulewidth and \lightrulewidth
@tohecz This looks good, but it still ignores the {3-4} part
Oops! I am lost @Andrew. I am unable to see why ALL my attempts are failing. Do keycodes go in braces?
xmodmap -e clear (lock) ?
@tohecz pastebin.com/XrKLsa4i Take a quick look =)
@N3buchadnezzar they don't get ignored, try to change all white (3 times) to other colors (e.g. yellow, red, fuchsia) to see whta happens ;)
5:31 PM
@kan xmodmap -e 'clear lock'
@AndrewStacey Hah, thanks. So, let me write the keycode table into .Xmodmap now.
@tohecz I mean that when changing the width, it interprets {1-3} as {1-all} as in it does not only make the specific lines wider, it makes all the rows wider!
Pardon my english
@N3buchadnezzar no, it only leaves white space on the left and right from \cmidrule
I mean if i say make a line from coloum 2 to 3, it does 1 - 7 instead...
This is the result after using \cmidrule[4\heavyrulewidth](l{5pt}r{5pt}){3-4}
5:41 PM
Yeah, using yellow I see the very same thing. I also see that the entire row is shifted down the same amount.
Yes, there's a space left for the \cmidrule across the whole table
Well I do not want any space between the first two elements in each coloumn.
That what was what I was trying to say =)
I guess I could fill those with a different color.
@N3buchadnezzar if you can change \arrayrulecolor between the \cmidrules, and if \arrayrulecolor doesn't bring problem to \halign, \noalign and friends, yes, you can ;)
I tried this, did not work
\newcommand{\del}{ \arrayrulecolor{blue!7.5} \cmidrule[2\heavyrulewidth]{1-1} }
@tohecz This did the trick! thanks a bunch!
6:33 PM
My first LuaLaTeX experiment: tex.stackexchange.com/a/99264/12571 I guess that it won't be the last one... I had too much fun.
A: Let’s polish the Editors/IDEs question

doncherryEditor Sign Up List Texmaker: doncherry (Win7), hpesoj626 (linux) TeXworks: Joseph Wright, tohecz (linux) TeXnicCenter: Werner TeXStudio: hpesoj626 (linux) TeXShop: Alan Munn Vim: Psirus Emacs: Charles Stewart (OSX, linux), Tyler Smith Please add yourself with your favorite editor(s) here. In...

Don't forget to sign up!
I'm hoping @ulrikefischer will take WinEdt (I used to use it, so can if required)
Any other obvious people to take particular editors?
6:55 PM
@JLDiaz Nice!
@JLDiaz two comments: do you really need require 'table'? And table.getn() should not be used anymore, use #tablename instead.
@topskip Thanks, feel free of commenting any improvement you can see. As I said, this is my first attepmt at LuaLatex, and even at Lua, so it must be plenty of begginer's pitfalls
@JLDiaz No, not really
@topskip I had a hard time with tables being references. There is not a shortcut to do a table copy in a single asignment, like python's a=b[:] ?
Is a loop required to copy a table?
I'd also use \luaexec{ dofile('luafunctions.lua') } and just put the lua code from the tex file in there. That way you keep TeX from doing bad things to your lua code
@JLDiaz don't know what a=b[:]does, but you do need a loop to copy a table.
@topskip I learned about getn() at "Programmiing in Lua" site (lua.org/pil) Didn't know that it was deprecated
7:02 PM
@JLDiaz now you know :)
Lua has several difference in the language between minor versions (5.0 -> 5.1 -> 5.2)
@topskip : is the "slice" operator in python, you can get a copy of part of an array with: new_aray = old_array[start:end], but if you omit the start and end, it copies the whole array. I missed this feature in lua
With LuaTeX we're at 5.1 but will be switching to 5.2 soon
@topskip Must buy new Programming in Lua :-)
@JLDiaz There are many functions that you'll miss if you come from Python. Lua has such a small standard library that you have to code everything yourself (or take some 3rd party library)
@topskip Bit like TeX, then ;-)
7:08 PM
@topskip yes, of course, but I was refering to features in the core language, not to the "included batteries" :-) At least Lua has a very powerful table type which allows to implement almost any other datatipe required
@topskip I'm not looking forward to the changes to the way globals behave in 5.2
@JosephWright Yes, absolutely. The one thing I keep complaining about is that TeX does not have a decent (and solid) XML parser.
@topskip We should write an L3 one, I guess (perhaps @DavidCarlisle will oblige)
@topskip I thought that is or was an issue with LuaTeX too
@JosephWright You can embed quality XML parsers in Lua
7:10 PM
@JosephWright yes. But I am almost alone with my complaint, so I give up all hope :)
@topskip Ah, right
@CharlesStewart I meant 'out of the box'
@CharlesStewart Right (I've done that with libxml2), but as long as it's not in the standard distribution of LuaTeX, it's of no use
The biggest nuisance with Lua is Unicode
@CharlesStewart Unicode in terms of "UTF-8"?
@topskip Ok, I've implemented those changes, removed "table" requirement and used dofile to load the external functions. Apparently I didn't broke anything...
7:14 PM
@CharlesStewart I don't have any problems with the built in selene unicode library
(except for the part that some functions operate on bytes, others operate on utf8 characters - multiple bytes. It's easy to mix up byte positions and character positions)
7:26 PM
@topskip I've updated my answer to improve the code according with your suggestions, thanks!
@JLDiaz Great! Ive fixed a very minor "bug" :)
@topskip Thanks again!
8:01 PM
Observation: Whereas +3 °C is a good temperature for biking in short sleeves, -1 °C is not.
@topskip ah sob don't you like my parser:(
@DavidCarlisle :)
Does it work with LuaTeX?
@topskip if luatex has \def and \ifx then yes that's about all it uses:-) Although actually it would probably take some un-documented tweaks so it stopped trying to decode utf8 itself and let lua/xe-tex do it.
@DavidCarlisle I was thinking about UTF-8 compatibility
@DavidCarlisle The question I ask when I see an XML implementation (including my own ones): Does it handle utf16? xinclude? DTD's? (utf-8-)BOM? How fast is it? Does it handle 100mb XML easily without too much memory consumption?
@topskip well I think it would in luatex in classic tex it "handles" it by taking every other byte and hoping it is latin 1. Nothing else you can do as it assumes a delimited argument #1< looking for the next tag fins that not a random character with that as one of its bytes. but in luatex that would be handled before the nacro tokenizer so it should all work, I may set that up one day. The only other thing it fails on is that tex removes white space from ends of lines and you can't stop it.
it parses every construct allowed in the local subset of a dtd, and does full namespace processing. it doesn't really have any memory footpriint it just reports start/end although has hooks to grab element content as a macro so you just have to be careful what it grabs
@topskip but I'd never use it myself, just more convenient to use xslt at the front and just pass to tex something more easily digested by tex.
8:21 PM
@DavidCarlisle But that doesn't work when you want to read XML during TeX's run (unless you start expensive XSLT processors)
The point is, I'd like to have an xml parser that works even with things like UTF-16. And when you get UTF-16 and try to parse it by looking at 8bit bytes you're screwed when you read for example ȼ (which is 0x023C and < is 0x3C)
@topskip true, and for smalish xml fragments nid document it's fine. I've never found the need to process 100mb of xml mid document if I didn't know in advance that it needed to be parsed though. I've not actually run xmltex this century but some people use it
I've used xmltex as well and I really liked it! But it doesn't fit in a LuaTeX world easily - some existing C libraries would.
@topskip yep it uses utf8 internally and has hooks for any 8bit encoding, but it's a documented feature that utf16 is a lost cause. But as I say that shouldn't be a problem in luatex.
@topskip couldn't you write a reasonable parser in lua (it must be easier than writing it in TeX:-)
@DavidCarlisle "reasonable" can be stretched :) - I work now with a very small one (written myself), but I know it will explode if something unusual happens (like encoding != UTF-8)
@topskip oh do you have to handle the unicode yourself? I don't know lua but I'd have hoped it could give you a stream of Unicode data handling the external file encoding using some central libraries.
8:34 PM
@DavidCarlisle Well, there is the selene unicode library which does that. But still I must read the bom(if present) and the encoding="..." whatever char. I don't do that at the moment.
(Just found out that the unicode library is only utf-8/latin-1/ascii. Would need to convert utf16 to utf8 before parsing the file)
@topskip - Fully fledged utf16 is kind of rare - ucs2 is much easier to parse and from what I've heard pretty much covers the texts that are around.
@topskip Restricting an xml parser to handle only utf8 seems good to me: I should think you can always handle character class conversion with external tools.
@CharlesStewart I hoped that don't have to bother with the details and have a good library built in (LuaTeX) that does that conversion for me.
@topskip Why not build luatex/ texlua with libxml2 built-in?
8:49 PM
@CharlesStewart "rare" depends on your text content. It is not rare to go out of bmp if you use mathematics, or chinese.
@CharlesStewart because it makes documents unportable (unless that luatex is the default for texlive/miktex)
@topskip With dynamic loading, "build" only means having the library somewhere it can be found, so this hopefully isn't a huge problem.
Any chance of getting it into Texlive, I wonder?
@CharlesStewart It will be solved soon, I guess. The LuaTeX people are working on it.
@DavidCarlisle Without toks:
@egreg yes but I like my code to be clear and self documenting, not inscrutable and opaque:-)
@egreg that will fail I think if #2 is a counter as the \number will not scan ahead so the \xp chain won't reach #3
8:59 PM
@DavidCarlisle Yeah: [\number\numexpr#2\xp]
@egreg good plan a very under-rated feature of numexpr that it normalises the scan-ahead skip/notskip space process.
@DavidCarlisle Or just [\number\number#2\xp]
@egreg I think my edef version requires less explanation:-)
@DavidCarlisle Next, do it with \def only.
@egreg I'll just use pgfmath
9:52 PM
@tohecz Then vote for reopening. I had my doubts, too, BTW.
10:16 PM
@tohecz Hmm, he (?) wrote, he had found this wiki page. I understood, he had searched for (cite:) “a list of all this marks as a **LaTeX ‘commands’**” (markup by me). That’s why I voted as duplicate, see also my comment. There are, of course, the basic ligatures `` and '', and also babel shorthands.
@Speravir we might speak about another one. The new one is a duplicate. The old one is not.
and apologise me, I get up quite early tomorrow, so I gotta go.
@tohecz Sure? See, what message I refer in the second above.
@Speravir damn, I'm seriously confused
I'm not sure about the new one (IMHO it's off topic), but I'm sure about the old one.
@tohecz And now I’m confused. Which one do you refer now?
New: http://tex.stackexchange.com/questions/89051/list-of-mostly-all-quotation-marks-used
Old: http://tex.stackexchange.com/questions/77646/how-to-typset-the-symbol-caret-circumflex
New is closed and I'm fine with that (just the reason is IMHO "off-topic and belongs to GD", not "duplicate")
Old should be reopened, circumflex is a special symbol that can have different meanings and different ways how to typeset it in different contexts.
10:34 PM
@tohecz Now, I understand. Yes, the “old” question is phrased poorly, but Qrrbirbel’s (or so) answer is on- topic.
@tohecz That’s want I meant, with “vote for reopening then”.
@Speravir ok, good. And with this positive conclusion I can happily switch off the computer and catch a nice sleep :)
@tohecz Ha, good night and sleep well!
Just started to look at answering a question and found (via google) I wrote the package in the title, I'm sure I don't remember:-)
@DavidCarlisle pgfplot? ;-)
@DavidCarlisle scalefnt resides in a path tex/latex/carlisle. If you don’t remember, you had written it, then it was surely a relative of you.

« first day (844 days earlier)      last day (2799 days later) »