@WillRobertson Like, with a package which redefine them?

@leo Not sure what you mean — probably breqn would include an option to break automatically for things like \[...\], equation, etc.

@WillRobertson Exactly

I know that the question is very bad, but is it possible that there's a bug in interaction of W10 with LaTeX?

@yo' No: everything works fine with Win10

Anyone remember exactly how one use texindy and utf8? Just testing with texindy -L english -C utf8 file.idx keeps giving me Cannot locate xindy module for language english in codepage utf8. I have the latest updates for Linux.

@daleif That works fine here. Do you have the module files for english under texmf-dist/xindy/modules/lang/?

lang/english contains

as expected

@TorbjørnT. Which OS and LaTeX dist?

@daleif Hm. I'm not exactly an expert on xindy, so I don't have any other suggestions, I'm afraid. Kubuntu 14.04, TL 2015.

@daleif I seem to remember a similar recent question on the site.

@egreg I know, on Mac. tex.stackexchange.com/q/258339/3929

@daleif Exactly.

Ahh, it is this one: @styles = glob("$quotified_lang_dir/$variant$cp*-lang.xdy");, I do not know how glob works, but $quotified_lang_dir/$variant$cp-lang.xdy expands to "/opt/texlive/2015/texmf-dist/xindy/modules/lang/english"/utf8-lang.xdy and since " are valid in filenames that causes some issues. If I test with glob("$lang_dir/....) instead, xindy works fine.

@PauloCereda :) good choice!

@clemens <3

@TorbjørnT. What are your impressions on KDE 5 too? :)

@PauloCereda Haven't installed it. I'm still on the last LTS version of Kubuntu (14.04), and that won't get KDE5 in the repositories

@TorbjørnT. Ah got it. :)

Does anybody understand this? tex.stackexchange.com/questions/164164/add-lines-into-tex-file

@Johannes_B Well, it's clear what he want's to do, but I don't really see why he wants to do it.

@TorbjørnT. But why using LaTeX to open the file and insert stuff?

@Johannes_B Don't know.

@Johannes_B Or, I guess he wants to have an automatic process, so if the file is regenerated he doesn't have to edit it again.

@TorbjørnT. Possible. I bet there is something to do on the matlab part. Or use pgfplots directly, i.e. exporting plain data from matlab.

@Johannes_B You can do it Matlab, yes, see my comments.

@TorbjørnT. Let's hope it helps somebody. The question is quite old :-)

@egreg ^^^^^

@StefanKottwitz Where should we migrate the comments? It contains useful information for passers-by.

@PauloCereda Like my answer? tex.stackexchange.com/a/261052/37907

@Johannes_B Love it. :)

@PauloCereda <3

:-)

@Johannes_B :)

@ChristianHupfer Yummy sz-magazin.sueddeutsche.de/texte/anzeigen/43407/Roesler-Kuchen

@1010011010 Normally useful information should be moved to the answer. if it's lower level but useful, it can stay as comment, but I guess not many

@1010011010 Later readers would not need a chronology as in a long dialogue, but a summary

@1010011010 Also comments are not sorted, partially even hidden. So you may choose which should stay and delete the rest, I suggest. You can flag comments for deletion if they are then obsolete but by others

@StefanKottwitz In this case, an answer can't be formulated yet (there is no fix available), but the comments may help passers-by in debugging more general problems closely related to mine. What's your suggested course of action? Delete anyway and post the dialogue in summary as an answer?

@1010011010 at least the amount can be reduced, I did not check one by one but often comment threads are question answer question answer ... such as for further information, which could be updated to the question (or just personal q & a which can be deleted after reading)

@StefanKottwitz I've moved the conversation to the question. I'll write some kind of transcript from it when I have the time.

@StefanKottwitz It is quite sad that i cannot properly answer the question of a chinese guy, because the system blocks all chinese characters. Even soup and fortune. :-/

@Johannes_B you can presumably use ^^^^ notation, although it is not a good situation, egreg's bug report didn't get a change made it seems

湯

@DavidCarlisle I left a comment to the meta question and a short note in my answer tex.stackexchange.com/a/261042/37907 :-)

@Johannes_B seems like you can paste here so you could put the full code here and link to it. By the way probably worth saying in your answer it is xe/lua tex not pdftex

@DavidCarlisle Oh, right. Good idea. :-)

If i should ever get a tattoo, it'll just be I U+2665

hi,

i have question to latex(3) guru

I will start a new project

from scratch

I would need :

- separate private / public function

- dynamicly create function

- use key list

- use argument by key value

@Maïeul Sorry, I'm already lost.

My question is : should i start with latex3 or use standard latex2 modules (etoolbox, xkeyvalue etc.)

@Maïeul ah, well, IMHO you should use L3, but the problem is, it has no learning manual yet...

@Johannes_B: ^^ :)

yes, it sees

ok,

let's try with the handbook

@Maïeul you have basically 3 sources: xparse.pdf, expl3.pdf, interface3.pdf

ok

I have read expl3, and now I am lookin on interfaces3

it seems complex, but not so hard

xparse should be used for user-level commands. Then you have expl3 and all its "subpackages" that are used as a programming layer

ok

but xspace is not document in interface3 ?

@Maïeul XSPACE IS A HELL and has NOTHING to do with L3

(sorry for yelling)

sorry, and I would like say xparse

not xspace

@Maïeul ah, sorry then :-)

xparse doesn't officially belong to expl3, but it is a good company to it, for user-level commands.

(I am not agree with xspace judgement. the problem is not xspace, the problem is the way tex work with macro unedlimted)

ok.

thanks

As for interface3.pdf, you should read the first 6 parts, and only list through the rest.

@PauloCereda Hihi, that is cute :-)

@Maïeul No, the problem is that people use TeX the wrong way.

@yo : if people use the strong way, that means the normal way is not good ;-)

in my sens, code tool should be adapted to user, not inverse

When I write `\foo toto' there is no logical reason why space after foo is gooble if foo can't take any argument

the only logic is knuth logic

@Maïeul Well, you can design your own language, if you are that clever

so, anyway, that is not the question

@Maïeul no, indeed it is not.

I like latex a lot, I am not so much fan of tex, but I wont reinvente something, because no body will follow

@Maïeul well, I just hate people saying these shits. (And yes, you can flag this message)

you have easy hate

ok

stop talking here, we do not agrre

but I think it is a classical tex troll

Dancing turtle! :)

@JosephWright, @DavidCarlisle: ^^ There is something very wrong with this duck. :)

so if I understand well: \cs_new and similar when I want to create internal function and \NewDocumentCommand pour le reste ?

oh soryy, \NewDocumentCommand for user level command?

@Maïeul yes, \NewDocumentCommand for user lever commands, and \cs_new:Npn for internal commands

ok, many thanks @yo. This point was not clear. It is said only and the hand of expl3 handbook...

@Maïeul As I say, the whole L3 team knows that the missing learning manual is a burden, but, on the other hand, many features are still officially labelled "experimental"

yes, ok course, it was just a remark

ok, so, for now, i will start

@barbarabeeton That looks like a poorly printed penguin book. latex-community.org/forum/…

@Maïeul there are lots of logical reasons for that, how else would you enter \foo followed by toto ???

you mean something which produce expansion of foo glued to toto?

\foo\nospace toto for example

where is more time when you use macro without argument followed by space than adverse

@Johannes_B -- not about to disagree; don't have any penguin books handy to compare. i do notice that the "f"s are a bit tight (and the f-ligatures are present), and in line 3, first page, there is a "Th" ligature (or just overlap). not sure what the "Gamma old-style zero" (even pages; just the old-style zero on odd pages) in the running heads is supposed to be. on p.3, l.4, the "regular space after final cap before period" shows up. style of quotes isn't british. suppose it's what he wants.

@Maïeul and why in the world would you prefer that over not defining \foo at all? (Since all cases I have seen that had a need for xspace proved to be conceptually wrong)

@Maïeul and \nospace does what works by expansion but otherwise similar to `\ignorespaces? look at any tex macro code there are thousands of spaces in indented sources that rely on spaces after tokens being gobbled as part of the tokenization, it is rather rare to have commands with no arguments in inline text

for example : I have a command called \Bar

@Maïeul which does what?

which is expanded to \index[np]{Barnabas}

@Maïeul ah. Well, that's conceptually wrong.

why?

You shouldn't need the [np], but otherwise, how in the world is His name is \Bar more readable than His name is \index{Barnabas}?

@Maïeul all the white space in this bit of eledmac for example relies on this:

  \ifnumbering
    \led@err@NumberingStarted
    \endnumbering
  \fi
  \global\numberingtrue
  \global\advance\section@num \@ne
  \initnumbering@reg
  \message{Section \the\section@num }%
  \line@list@stuff{\jobname.\extensionchars\the\section@num}%
  \l@dend@stuff
  \setcounter{pstart}{1}
  \ifl@dpairing
    \global\l@dnumpstartsL \z@
    \global\pst@rtedLfalse
  \else

@Maïeul by the way you are missing a % after {1} which is the one place you do not benefit from this automatic gobbling of white space so have to avoid adding it.

it is nor more readable, but writing speedly

ok, @DavidCarlisle you are right on this point (but for this specific command, that does not matter)

and so yo I need [np] because I have not only one index

@Maïeul bullshit. Sorry, really. And when you have to re-read the text and it's full of \Bar and \The and \Rom and whatever else? You can't even find the right sentence in the code!

@Maïeul perhaps you are relying on being in vertical mode but I am sure I could find some horizontal mode commands in that file that similarly depend on no space tokens being generated after command names

@in general, re-read text directly in tex is not very pleaseant, especillay when you add footnote/citation reference

and as such command are for the main topic of my dissertation, I wrote them MANY time

and use \Bar instead of \index[np]{Barnabas}

prevent to make mistake and is shortler

@DavidCarlisle ok, you are right, but let recognize this is not so obvious

@Maïeul you have spell checkers to prevent you from mistakes.

hum...

And also, PLEASE, never use these shorthands when someone will read the code after you (like when you submit a paper)

in my domain, NO PEOPLE accept tex code

so I should submit or the .pdf version or worse an export to word

@Maïeul ah, well, that sucks

isn't it ?

@Maïeul sorry, what?

"that sucks"

@Maïeul ah you meant: n'est-ce pas?

yes, sorry for english

doesn't it will be better?

@Maïeul yep. Well, don't ask how's my French :-)

@GonzaloMedina Why did you delete your titlesec answer? It is almost correct (just take out the \clearpage from the [] argument in the \titleformat command.

@yo' Don't ask how my French is :p

@AlanMunn Don't ask how my English is :p

@yo' :)

@AlanMunn I hope you haven't listened to my talk.

@barbarabeeton penguin books are pretty cheap, cheap paper and the print usually looks horribly. Not even a static baseline, soometimes letters are dancing around.

@yo' I haven't, but given that I don't really speak any language other than English with any competence at all (despite knowing enough about a few to be dangerous) I have great respect for anyone who does.

@AlanMunn well, thanks :-)

@Johannes_B -- i guess i remember the "classic" penguin books. inexpensive they may have been, but well designed (ever hear of jan tschichold?). the ones i have in my library were probably all done when metal type was still in use.

hum, im not sure, but which in "general" language are "string variable" is conceptually "token list variable" in l3exp?

@yo' One of the side effects of speaking a language that everyone else speaks as a second language is you have very little direct incentive to learn other languages properly.

@barbarabeeton I might be confusing it and it is the Wordsworth Classic series that is badly printed?

@AlanMunn We no speak americano. :-)

@Johannes_B Me neither. :)

what if I need a local?

@Maïeul You don't :-)

@Johannes_B -- i'm not familiar with that series. but i also haven't looked closely at any really recent penguin books; i shall have to do so next time i'm in a bookstore. we all know that knuth threw up his hands in horror when he saw the first galleys, phototypeset, for the volume 2 revision of taocp, so standards can degrade with change of technology.

ok, I must misunderstood somehting

@Maïeul More specifically, we tried this out and it doesn't work that clearly as TeX grouping is explicit anyway

@barbarabeeton @Johannes_B I agree. I just bought one recently, set in Monotype Janson. Really nice.

sorry Joseph, I don't understand your last sentence

@barbarabeeton I really think the OP is looking for exactly that output that led to developing TeX.

in expl3.pdf , § 3.1, the naming schem is mentioning local variable

but what if all variable are created globally=

@Johannes_B -- you may be right. but then the discussion about what's different between metal and digital type isn't really relevant. what it should be about is "where phototypography went wrong". (metal type does have limitations, and that can be seen in, for example, a lot of scanned books from the 19th century -- or real ones, if your library still has them on the shelves.)

@Maïeul Variables are created globally in all cases but may be used locally

@Maïeul Variables are created by assigning a name with a register number, which can be done locally but not is by plain TeX, LaTeX or expl3

so the only way to be sur the variable won't be overriden by other function is to prefix them by the name of the function?

@Maïeul think of \int_new: as "reserving the name" and nothing more.

@Maïeul prefix them by your private prefix.

\int_new:c \l_maieul_whatever_int

that is the point: the private prefix is defined by package, of by function?

@Maïeul package. Or, if you have your personal functions, your "private prefix"

@barbarabeeton Asked for an example/scan of what he is trying to reproduce.

@Maïeul Same namespace approach TeX has always had, except we do have a list of the prefixes registered

@barbarabeeton btw, i made a decision today :-) chat.stackexchange.com/transcript/message/23380369#23380369

for instance I use ctuthesis inside the package ctuthesis.cls, but when I write my thesis, I use tohecz as the prefix (unregistered, it's truly private)

ok

hum

it seems very complex

@Maïeul do you design a package/class?

yes

a package

@Maïeul No more complex than any other programming: you have to have some kind of namespace approach

@Maïeul then use its name as a prefix. It should be a reasonable name :-) Then, register it with @Joseph :-)

@Johannes_B -- but that's an incomplete sentence.

@barbarabeeton If people started to use grammar and orthography when dealing with tattoos, the world would be a less funny place :-)

ok

in think In understand better, but it not so obvious

@Maïeul if you have a name for the package already, what is it?

reledform

@Maïeul Sounds good to me :-)

I have a general idea of the way I would like it works, and conceptual idea of internal function

@Maïeul well, just use reledform as your prefix then :-) Just ask @Joseph whether the prefix is still free :-)

but I would like just to be sure the l3exp system is not to more complex for a not programming people like me

I mean, an autodidact

@barbarabeeton I am wondering how many people would actually make sense of I U+2665

I am quite afraid

but I would try

@Maïeul it is not 100% easy, but you'll manage :-)

my problem is also because I want to have two part in my package

@yo' I don't think you can hide the fact that if you are programming you need some work

one should be the API and the other one should be a default implementation of the API

@Johannes_B -- true enough. i'm not likely to get a tattoo (too old-fashioned), but if i did, it wouldn't have even a single letter. re making sense of I U+2665 (would you use the symbol or the text string identifying it?), if you used the symbol, it would just seem like the first part of a lot of bumper stickers.

so the API should provides tools to user

so these tools won't use __

as it will be public

BUT the api will use it internally

ok, It's clearer in my mind now

@barbarabeeton The string. If a girl gets the pun, i have to marry her.

so @JosephWright no one have token reledform as prefix=

@Johannes_B -- good luck! (that'll be one geeky girl!)

@barbarabeeton :-D Yes

@Johannes_B -- but what if she's already married?

@barbarabeeton Well, she can get in line with the other girls i start to like, just to find out they are married.

@Johannes_B :)

to be sure

1) create list, private

2) \foo is use to set value of this list

Am I right =

?

@Maïeul Not exactly

@Maïeul \g__reledform_witnesses_clist or \l__reledform_witnesses_clist

@Maïeul Braces missing

\clist_gset:Nn \g__reledform_witnesses_clist {#1}

@Maïeul What are all those % doing in your code?

@Maïeul the percents are superfluous :-)

I think it is an "habitus" keep from php wordl to finish with %, in order to be sur to not add spurious space by mistake

@Maïeul looks good, just one little thing: Use indentation and spaces:

yo : it just a bad pastebi

\clist_new:N \g__reledform_witnesses_clist

% Starting command: user must declare the witnesses
\NewDocumentCommand { \SetWitnesses } { m } {
    \clist_gset:Nn \g__reledform_witnesses_clist { #1 }
}

I use two space as indent

(the spaces are "optional", I prefer slightly tighter code than this)

@Maïeul works good, it's a matter of taste

@yo' Well, we do have a style guide with an official position (which is a compromise between what the various members of the team like)

Hum, I prefer withou space around structure

but it just a question of convention

an other point : I know many people use docstrip

but I am not convinced

when I need to read code with comment, I prefer to read a .sty

because pdf search engine are slow

@JosephWright yeah I know. Sorry :-)

Should we react to this? mathoverflow.net/a/213268/43387

@Maïeul That's your call

Joseph?

@Maïeul What makes you think .dtx means reading PDF files :-)

@Maïeul Whether to have a source/unpack set up or write the package directly

sorry?

@Maïeul It is up to you whether you use the .dtx format or not: the team do not take a position on that

OK, nice;-)

@Maïeul On the spacing, we are trying to get everyone to stick to the same style: lots of other programming languages take the same approach (having a fixed style) as it makes it easier for collaboration, etc.

yes, and what is the convention ?

a question abouzt

convention name

\reledform_get_lemma:

is that the good convention name for a function without argument?

@Maïeul Yes

ok, and so a function with two argument will be, for example \reledform_get_lemma:nn

or \reledform_get_lemma:Nn

or \reledform_get_lemma:NN

depending of the nature of argument?

OK

many thanks

@Maïeul Depending on whether it's one or many tokens, yes

yes

@StefanKottwitz I am not succeeding in attaching an image, it is automatically inserted in the post. I don't want to resize that specific image, as the detail is needed. latex-community.org/forum/… Can you fix that please?

@StefanKottwitz The picture formerly was on ulrik.uio.no:4107/loeb.jpg

is this normal that \prop_gput:Nxn does not exist?

@Maïeul Yep, you can create a variant. But are you sure you don't want \prop_gput:NVn?

I could use it, but that implies to store result of a command in a variable

I explain : I would like to have a comamnd

@Maïeul then you're fine with x

ok. So I thinkg I start to understanding the logical

it is a great way to manage with all the problem of expansion caused by tex logic

It's just common that people use \prop_gput:Nxn \one { \two } { something } rather than the better \prop_gput:NVx \one \two { something }

som my suggestion of \prop_gput:Nxn is not so stupid?

@Maïeul but you can't get around the knowledge of expansion

@Maïeul if the second argument will be \processsomething {blabla} {foo} {bar} 12345, then it's correct

should be \clist_use:Nn no?

I would like the "return" of \reledform_get_local_witnesses: be the key

@Maïeul well, I'm lost now :)

I can understand

So i restart

whihc, basically, take a csv list, and then merge all item without anything between them

I I would like the result of such operation be a key of a prop

@Maïeul well, it doesn't take an argument

which "it" ?

\reledform_get_local_witnesses:

ah ok I see now what you mean, sorry

@barbarabeeton An update received. latex-community.org/forum/… I am now completely clueless what is going on, though.

so I should create \prop_gput:Nxn variante

with \cs_generate_variant:Nn ?

@Maïeul exactly

\cs_generate_variant:Nn \prop_gput:Nnn {Nx}

{Nxn} ?

@Maïeul you omit all last arguments' specs you don't change, it generates faster code.

ok

(page 28 of i3 interface3)

Ok, many thank

it is simplier with explanation that by reading handbook

@Maïeul you'll get into it after some time. I'm not on the L3 team nor an L3 expert. I've only written one quite complex class using the interface, about 70 kB of code, 1.8k lines

I should print a "shortguide" with the meaning of all key (x,N,V)

ok,

for now, I think I am quite prude of my test

I have the basis of my need

I havce understood some thing of l3expl syntaxis

now, let me complexefiy the code :p

many thank yo'

and also @Jose

@JosephWright

@Johannes_B What shall I do if you don't want to resize? I could only resize or crop or split in parts for the details.

@StefanKottwitz So there no way to just attach the image without display?

@Johannes_B For now, also attached images are automatically displayed :-(

@StefanKottwitz Oh, i guess i'll cut the image and display parts of it. But first a short break. Later :-)

shit, \prop_map_inline has no variant to insert something betwteen each item

@Johannes_B Great! parts would be very fine

hum

hum, property list are unordered

so there is now way to ordered them, except by maintening a parallel list l3seq ?

@Maïeul they indeed are unordered

what is a set? an "ensemble"?

all l3expl is documented in interface3.pdf, or there is other fragment?

@Johannes_B The template strikes again. tex.stackexchange.com/q/261070/2693

hum, there is now way to sort sequence? should I implement myselft this feature

@Maïeul Because they are accessed via their keys, and not their positional places. :)

Paulo : yes, it si logiccal, but in some programming language, you can sort the key

@Maïeul I think this is expected because what would be the criterion for sorting something? :)

@Paul : for example "alphabetic order"

but it could be any command which say "this is before this"

@Maïeul Yep, there should be a comparator. :)

yes, of course

Indeed, it was my mean hope when using latex3

I have such feature

because that is what I need for my projetc

@Maïeul yep, a mathematical set

yo : ok, it is nice that manipulating set are explained

@Maïeul oh sorry. There's no module l3set, is there?

no there is not, but there is some example of how to do it in the handbook