@Dynamic We have more than enough experts in here but absent at the moment.

@Dynamic what detail are you looking for? tug.org/whatis.html might be a start

@ClarkKent Actually I am pretty sure they added a plugin for Firefox

@TorbjørnT I can make a vim cheatsheet for far less then that: apt-get install emacs

@percusse: Have you posted your answer for sed question (I mean winedt one)? If not pl. post it.

@Dynamic used in what sense? and just TeX (the program/language) or the wider picture? For TeX directly, thoughts and pencil and paper .. there are quite some papers written by Don Knuth that are worth reading. For the wider range, the interviews of various TeX people on tug.org/interviews could be of interest (or the book version of it) as there you find thoughts by many people who have been involved in the making for a long time

@HarishKumar Done!

@percusse It was a perfectly reasonable question asking for a GUI editing tool , and gets high scoring replies for sed, awk and emacs, and then you want to spoil it by suggesting an editor that has menus and mice and things?

@percusse next you'll be suggesting checking the output with acrobat rather than looking at the \showoutput log in emacs...

@DavidCarlisle I represent the incompetent PC users in here. Somebody has to take care of my fellows.

@DavidCarlisle I'm deeply intimidated by the command line. It's been a while since I stopped after MS-DOS 6.2 and I remember people arguing over IBM-DOS 6.3 :). Moreover I don't know what sed,awk or emacs is but planning to switch to Linux after my thesis. Then I'll be able to hold an intellectual discussion

@percusse You don't have to use them (except emacs of course which is a mandatory life skill) (I don't use awk or TeX for example) You just have to sound as if you could use them so that you can bluff your way through discussions.

OP asks this closed as too localised tex.stackexchange.com/questions/56217/…

@DavidCarlisle voted

Need one vote: tex.stackexchange.com/questions/56219/…

@MarcoDaniel done

@DavidCarlisle too ;-)

Can I use the LaTeX loop with two conditionals

Something like this:

\@tempcnta=\z@
\@tempdim=\z@
\loop
  ... do something with dim
  \ifdim\@tempdim<20pt
 \ifdim\@tempcnta<100
\repeat

@MarcoDaniel yes but not like that your ifs and fis have to match up

@DavidCarlisle So I have to use:

 \ifnum\@tempcnta>100
    \let\iterate\relax
 \fi

to break the loop.

@DavidCarlisle that's tricky ;-)

The above being the TeX syntax for or (oh actually I think you wanted "and" which is the same but with the inner a and b switched

@MarcoDaniel you should look in the l3 sources one day:-)

@DavidCarlisle It's for splitting a box. I want to split until a special dimen is reached or the loop count is greater than 100

@DavidCarlisle ;-)

basically` \if a a` is "true" or \if a b is "false" but since \if expands you can have any number of nested if constructs to return either a or b to make up complex predicates. (that's basically what 2e's ifthenelse package does with \oORas well.

@DavidCarlisle interesting. I think I will switch my package to LaTeX3 if the issue with l3galley is solved.

@MarcoDaniel well you can do as there or if you want to make it understandable test your dimen first and if it is right just bump your loop counter up so the following test on that ends things

@MarcoDaniel "if the" singular????????

@DavidCarlisle if the -- I will do it after l3galley works well.

Or better I understand the implementation.

@DavidCarlisle There are some open questions (at least open for @MarcoDaniel :-)

@JosephWright It seems you remember my problem ;-)

@MarcoDaniel Vaguely: lots to think about

@JosephWright: The following fails if I load l3galley.

   \vbox_set_inline_begin:N \l_fullwidth_store_contents_one_box
   \group_begin:
   \dim_set:Nn \tex_hsize:D \l_fullwidth_resulted_environment_width_dim
   \dim_set:Nn \textwidth \tex_hsize:D
   \dim_set:Nn \columnwidth \tex_hsize:D

@MarcoDaniel \tex_hsize:D when we say D we mean DO NOT TOUCH THIS

I know that I must work with \l_galley_total_left_margin_dim if l3galley is loaded.

@DavidCarlisle ;-) Really ;-)

@DavidCarlisle There are some unresolved questions about widths, possibly flaws in the way LaTeX2e describes/implements things

@JosephWright It seems that the `caption label separator` template is not documented in the beamer manual. Perhaps you could include it in a next version?

@MarcoDaniel pretty much yes. It's like the inputenc question on site yesterday when someone had loaded inputenc (OK) and then changed the catcode of every letter (not so OK) and then defined latex-robust commands but used \write not \@protected@write If you define a higher level API you have to "force" users to use that, or they pull the rug out from under your feet.

@GonzaloMedina Log a bug :-)

@DavidCarlisle I will try to fix this in my L3-package and start a new test.

Thanks

@DavidCarlisle I'd love it if some experienced people would look over l3galley :-)

@MarcoDaniel If you use the l3 programming layer for loops and things that fits well with 2e, but the galley really needs to know about everything that on the .... galley .. so it only really works if you never put any space or boxes or anything on the main vertical list except through galley functions, and definitely don't change the width of the galley without telling the system by using hsize:D.

@JosephWright good point ... David?

oops didn't notice @FrankMittelbach was here:-)

@FrankMittelbach perhaps but it seems I take some time to recover from all night coding sessions getting w3.org/Math/draft-spec/mathml.html out the door:-(

@DavidCarlisle No shy ;-) -- I think after a new implementation I will upload the file ;-)

@DavidCarlisle thought you got a bunch of mails from me throughout today ... so yes I have window open into this chat for a while

@DavidCarlisle @Marco unfortunately what David says is that galley is one of those parts that really will work properly only if more or less everything is L3 (perhaps even only with a separate format, but at least only if one is very diciplined). This is a lot different to other stuff (and also stuff that goes far above the programming layer).

@DavidCarlisle Quite

In @FrankMittelbach's implementation there was more code to 'patch' LaTeX2e, but I think that will still break the moment you fiddle with \hsize or similar

And in this respect it is a bit unfortunate perhaps that we already got this more or less out of the door, as one can't really appreciate the power behind it (that I think it has :-) )

But then it was a fairly ready piece of code.

@FrankMittelbach The idea is I think right, which is what l3experimental is for (idea right, interface or interaction with other code perhaps less so)

@percusse: Nice one.

@JosephWright Of course.

@JosephWright sure - not complaining just trying to explain to @Marco why he may have too high expectations one one hand while actually not being able to appreciate the advantages

@FrankMittelbach: You remember correctly ;-)

Sorry, I'm late!

Now with new glasses! :)

@PauloCereda Can you focus at arara now? ;-)

@StefanKottwitz: Please see: tex.stackexchange.com/questions/55760/…

@MarcoDaniel Most certainly yes. :)

I almost hit a post while walking after putting the new glasses on.

@PauloCereda Is it the first time you're wearing glasses?

@egreg No. :) It's been 12 years now. :)

@PauloCereda So these are stronger than before.

But the values altered quite a bit this time. :(

@egreg Exactly. :)

@PauloCereda Been there, done that. :)

@egreg :)

I was wearing the old ones, so when I put the new ones, my eyes complained. :)

@PauloCereda It happened to me after I was in the army: a big decrease. Everybody says that myopia stops changing around 21-23: it isn't true.

@egreg Really? Mine was stable from 5 or 6 years in a row. I once asked my ophthalmologist if the values could also decrease; he said it's possible, though not so common. He also said the reduction could be around even around 1 or 2 points. :)

@PauloCereda Well, I'm in the "not so common case". :(

@egreg I'm there too. :) But for a short period. :)

@egreg: now I'm confused: I understand "decrease" as good (the value reduced). Is this the right interpretation? :)

In portuguese, when someone says, "o grau diminiu" means that the eyes are better. :)

@MarcoDaniel What's the best way for generating a build? Let's say, via a shell command?

@StefanKottwitz Maybe you can do the following:

mkdir temppgf && cd temppgf
cvs -d:pserver:anonymous@pgf.cvs.sourceforge.net:/cvsroot/pgf login
cvs -z3 -d:pserver:anonymous@pgf.cvs.sourceforge.net:/cvsroot/pgf co -P pgf
find * -name CVS  | xargs rm -r

Of course now you can create also a tar file.

And you can provide a Makefile. I love this one: github.com/latex3/svn-mirror/blob/master/l3experimental/xgalley/…

@egreg: Palmeiras has a new uniform for the next league: globoesporte.globo.com/futebol/times/palmeiras/noticia/2012/05/… :)

@Jake So close ... yet so far. In that I have absolutely no idea where the error is so it's going to be painful to track it down.

The red path is metapost's, the black path is generated by my code.

Hmmm beans. :)

@PauloCereda Nice. The bus driver did a nice joke to us: he arrived 7 minutes earlier, so I couldn't catch that bus and, of course, the train was in perfect schedule, so I missed it.

@egreg Oh. :( Murphy's law strikes again.

@MarcoDaniel Glad you like it :-)

@JosephWright It is fantastic. Daily I run:

#!/bin/bash
# small script to use the newest svn-version of LaTeX3
echo "Install LaTeX3"
cd $HOME
mkdir tmpinstallfolder-latex3
cd tmpinstallfolder-latex3
svn co latex-project.org/svnroot/experimental/trunk
cd trunk
make localinstall
cd $HOME
rmdir --ignore-fail-on-non-empty tmpinstallfolder-latex3

Can someone help me understanding the question?

@PatrickGundlach I don't think there is one: it's more of a statement of an idea. Might make a good answer to a suitable question :-)

I believe it's off topic and I am (again) the devil's advocate :)

@StefanKottwitz Check out this new user's answers:

@lockstep Nice answers :)

@PatrickGundlach Good for anyone wanting to become Deputy/Marshal. ;-)

And another answer, including a lot of "69"s ...

Spoke too soon. Turned out to be a confusing between 0 and 1.

Or even a "confusion".

There we go. Hobby's algorithm implemented almost completely within LaTeX3 with some PGF and TikZ wrapper code (the "almost" is because I needed the atan function from PGFMath, I'm told that's in the latest l3).

Exact duplicate, based on the OP's last comment:

@AndrewStacey Fantastic result. I hope I can find the time to debug my version. I have managed to pass it through TikZ natural parser now I need to store them in some sort of a key. The formulas are still wrong.

@percusse What parser did you use? I've ended up splitting the input at the TikZ level before sending it off to LaTeX3 for the actual algorithm. So I didn't use the \seq_whatever_it_was.

@AndrewStacey Actually you can intervene when it says Hey it's a controls syntax.

@percusse Ah, there's a hook? So how would a user specify that they wanted a smooth curve through a load of points in your code?

@PauloCereda Murphy at its full power! The next train (where I'm sitting now) is 20 minutes late. So those 7 minutes for the bus driver become almost 1 and a half hour for me.

@egreg Oh my! :( The bus driver is mean. :(

@AndrewStacey if there is no additional controls word you might be able to branch out to your algorithm. I don't want to speak to soon because I could only read the remaining part but in 2473 in tikz.code.tex file you can see that we have some space to squeeze in.

@egreg: I recommend an extra dose of Bach today. :)

Another off-topic post :)

@percusse You mean the mysterious \tikz@curveto@auto function?

@percusse Interesting idea. I can see that it would be possible to use that to gather the points together. The difficulty would be that at some point you need to call the curve generation code. This can only be done when the points have all been gathered, so one needs to say "when we stop getting .. (coordinate) then call this macro". I don't know if TikZ can do that.

Weird person bumps up a question with spam and I get two badges from collateral votes. Thanks weird person. :)

@AlanMunn LOL

@DavidCarlisle: No TikZ questions today? :)

@AndrewStacey Yes that's exactly the reason why I wanted to create temporary coordinates immediately in the picture with egreg's L3 help. As you say until we reach a point where a .. is not following the coordinate, one can ship it to the code.

@AndrewStacey, unless things have changed I think your choices for getting out of a group are same as always. The one you mentioned \gdef\tmp{}\egroup\let\foo\tmp or that old faithful \expandafter as in ` \def\tmp{} \expandafter\egroup\expandafter\def\expandafter{\tmp}`

@DavidCarlisle I tried such a way but it fails.

\documentclass{article}
\usepackage{expl3}
\begin{document}
\ExplSyntaxOn
\fp_new:N \l_hobby_veclena_fp
\fp_new:N \l_hobby_veclenb_fp

\fp_new:N \l_hobby_veclen_fp
\cs_new:Npx \tempa {}

\cs_new:Npn \fp_veclen:Nnn #1#2#3{
\group_begin:
  \fp_set:Nn \l_hobby_veclena_fp { #2 }
  \fp_set:Nn \l_hobby_veclenb_fp { #3 }
  \fp_mul:Nn \l_hobby_veclena_fp {\l_hobby_veclena_fp}
  \fp_mul:Nn \l_hobby_veclenb_fp {\l_hobby_veclenb_fp}
  \fp_add:Nn \l_hobby_veclena_fp {\l_hobby_veclenb_fp}
  \fp_pow:Nn \l_hobby_veclena_fp {.5}

@PauloCereda I could have implemented metapost in tikz, but I was too busy making mathml users happy by reformatting the mathml spec, apparently some people don't like html files with 57000 character lines

@DavidCarlisle LOL

` \exp_not:N \group_end:` is \noexpand\egroup if I recall correctly so that's not doing anything as \group_end isn't expandable so you can't stop it expanding. will look in a bit

@DavidCarlisle As a MathML user, I'm happy.

@percusse Okay, I have a method. When we encounter .. (coord) then we go into a smaller parser that keeps looking for .. and stops when it no longer gets one. Then it generates the curve and hands control back to the main TikZ parser. (Hmm, need to also exit if we encounter .. controls).

@DavidCarlisle Ah ok. I confused the expansion

@percusse \draw (-3,0) -- (0,0) .. (6,4) .. (4,9) .. (1,7) .. (3,5) -- ++(2,0); now works.

@AndrewStacey Hahah, nice!

I suspect that the global set a temporary is the easier one. using a primitive assignment it's not too bad using an expandafter chain to expand the right hand side but once you use a \set{\a}{\b} kind of syntax you need a whole load more to get past the set syntax as well

@DavidCarlisle I thought the following work -- but fails too

\tl_to_lowercase:n
   {
    \group_end:
      \fp_set_eq:NN #1  \l_hobby_veclena_fp
   }

@MarcoDaniel no there's nothing to lowercase there so it's a no-op (off to do some javascript, bye...)

@DavidCarlisle \group_end: is \endgroup. We're not using \bgroup ... \egroup or { ... }

@MarcoDaniel What's the plan?

@MarcoDaniel Of course it doesn't work. Is performing the assignment outside the group what you want to do?

@PauloCereda I'm at home (and had dinner, finally).

@MarcoDaniel \exp_args:NNNV \group_end: \fp_set:Nn #1 \l_hobby_veclena_fp

Why is everyone answering my question in comments or in chat? Isn't that what the "answer" box is for? Or is everyone scared someone will come along and flag as "Not an answer"?

@percusse, @Jake I've pushed my code to the TeX-SX launchpad site. Anyone who feels like testing it is welcome to do so. I need to document it properly but there are some examples in the file. Do pdflatex hobby.dtx to generate all the necessary files.

@JosephWright same difference it's not expandable either:-)

@AndrewStacey Acknowledged! I will do it as soon as possible also to get the correct values for the debugging. Thanks a lot.

@DavidCarlisle No, very true. I meant that we've picked one form of grouping for use in programming, so there are no odd { ... } pairs

@egreg ah yes V , I 'd forgotten that one, you should give that as the answer to andrew's question:-)

@AndrewStacey Not everyone!

@DavidCarlisle Just did. :)

@egreg Seen and voted. Will try it out in my code tomorrow.

@JosephWright oh I'm sure we'll need {\ifnum0=`}\fi before we're through:-)

@DavidCarlisle Oh yes, for special cases, but I meant in general

@JosephWright It's like learning french. The world consists entirely of irregular verbs, the regular case is a myth:-)

@egreg That will work?

@JosephWright It's the answer of a question.

@MarcoDaniel Yup, it's the officially-supported approach

@MarcoDaniel Well I've worked that out now :-)

@JosephWright Maybe a function \use_after_group_end would be nice

@MarcoDaniel We have vaguely discussed it, but you then hide the group closing, which is not ideal

@JosephWright I tried an approach of the form \edef\x{\endgroup\noexpand\foo{....

@MarcoDaniel Only works for stuff which is expandable. The new FPU is :-)

@JosephWright The question is really tricky. In relation to some questions here I think such an implementation is useful.

@JosephWright I have the new one ;-)

@JosephWright Is it as efficient as the old one? I mean, relatively efficient.

@egreg Yes, one of @brunolefloch's design constraints was 'roughly as efficient as the old one'

Depending on the operation, it may well be faster

@JosephWright :) Any news from Bruno?

@JosephWright I tried the following again but fails with the new one. What must I change to get the "old style"

 \cs_set:Npx\tempa
   {
      \group_end:
      \exp_not:N \fp_set_eq:NN #1 \l_hobby_veclena_fp
   }
\tempa

@egreg He's about

:4643602   \use:x
   {
      \group_end:
      \fp_set:Nn #1 { \l_hobby_veclena_fp }
   }

Two things to note: you can't set equal with this approach, and as the set operations are not expandable you don't need the \exp_not:N. Oh, and we have a pre-cooked \use:x.

@JosephWright That fails

\documentclass{article}
\usepackage{expl3}
\begin{document}
\ExplSyntaxOn
\fp_new:N \l_hobby_veclena_fp
\fp_new:N \l_hobby_veclenb_fp

\fp_new:N \l_hobby_veclen_fp
\cs_new:Npx \tempa {}

\cs_new:Npn \fp_veclen:Nnn #1#2#3{
\group_begin:
  \fp_set:Nn \l_hobby_veclena_fp { #2 }
  \fp_set:Nn \l_hobby_veclenb_fp { #3 }
  \fp_mul:Nn \l_hobby_veclena_fp {\l_hobby_veclena_fp}
  \fp_mul:Nn \l_hobby_veclenb_fp {\l_hobby_veclenb_fp}
  \fp_add:Nn \l_hobby_veclena_fp {\l_hobby_veclenb_fp}
  \fp_pow:Nn \l_hobby_veclena_fp {.5}

And a small hint. In you provided answer you wrote: \fp_veclen:NNN -- Should it be \fp_veclen:Nnn

@MarcoDaniel As written, the implication was that N-type arguments were being passed

@MarcoDaniel Oops, I forgot the \exp_not:N you do need:

  \use:x
   {
      \group_end:
      \fp_set:Nn \exp_not:N #1 { \l_hobby_veclena_fp }
   }

@JosephWright Great. -- @WillRobertson answered ;-) too

@JosephWright Ah ok.

I was a little bit irritate

@egreg: Flippity flop... :)

@Werner I studied \@dblarg for presenting it in my book. :)

@egreg The one you're translating to English? ;)