@alfred @DavidCarlisle: I didn't follow everything and has to go now but you shouldn't change language.dat in miktex -- it is created automatically -- but add a local languages.ini. Perhaps multiuser installation can do it through miktex settings (in single user installations the buttons are missing due to an old bug).

@UlrikeFischer yes (@alfred) same in texlive but I hoped the miktex one was like the texlive one that (a) says not to edit it and (b) has lots of comments saying what to do instead so I pointed at those comments

/wonders Why does the PDF spec not have a page colour concept?

@JosephWright Because they print on colored paper ;)

@TeXnician Could be

@TeXnician I suppose most designers deal with single pages that are coloured, so a 'full page box' is as straight-forward as colouring the page itself

@TeXnician That's the obvious answer. But it would have been nice to have: Then the page colour would show up in a pdf viewer (perhaps with an option to turn it off), but the printer would ignore it.

@JosephWright because it doesn't do page breaking so page colour is just a fill of the page bounding box

@DavidCarlisle Hmm

@alfred (@UlrikeFischer) to test new patterns in albanian/babel without having to keep building new formats you can use the following document in lualatex I just found some "sample albanian text" on the internet and made up some patterns that worked for this text but it might be a start) compare the hyphenation if you have \iftrue or \iffalse at the start.

\documentclass{article}

%\iffalse % use no patterns for albanian
\iftrue % use new patterns

\makeatletter
\newlanguage\l@albanian
\language\l@albanian
\lefthyphenmin2
\righthyphenmin2
\patterns{
ar5t
a5z
u5het
in5din
gj1e
jit1e
nj1e
er1ë
ë1z
z1it
}
\makeatother

\fi
\setlength\textwidth{6cm}
\usepackage[albanian]{babel}
\begin{document}

\tableofcontents
\section{Zzzz}
Të gjithë njerëzit lindin të lirë dhe të barabartë në dinjitet dhe në të drejta.
Ata kanë arsye dhe ndërgjegje dhe duhet të sillen ndaj njëri tjetrit me frymë vëllazërimi.

@PauloCereda ↑↑↑ The Adium mascot …

Looks a bit green … maybe he's watching @DavidCarlisle prepare dinner.

@alfred calligra seems to be a tfm font so you don't want to use fontspec (and easiest not to use xelatex either)

@HaraldHanche-Olsen ROFL

@HaraldHanche-Olsen oooooh

@HaraldHanche-Olsen David is mean :)

@HaraldHanche-Olsen oh

@DavidCarlisle ↑↑↑ beat you to it. A bit slow on the ball, lately?

I am hungry...

@DavidCarlisle oh no

who pays $35.00 to buy Gill Sans

@alfred Someone who likes the font?

Hi Peter Grill

murugans2k@gmail.com

getting "dimension too large" with dimensions of about 20cm in tikz. I give up. Until someone rewrites pgf properly using l3fp or whatever, as long as it's not prone to this stupidity

@alfred That's par for the course for a (professional, commercial) font file (so a few $100 for a full set of weights/shapes/_etc._)

@yo' Seems to be no problem in a simple file. What are you trying to achieve?

@yo' I have some low-level stuff in that direction (equivalent to pgf system layer): it's a question of getting round to speccing out a 'low level' 'picture mode' that uses this (@DavidCarlisle)

@JosephWright Indeed. I love my Lucida fonts. <3

@TeXnician well, it's random, I'm fine, very complicated picture, everything goes well. Then you change one -- to to[out=90,in=170] and it suddenly fails with no way to (1) debug (2) correct.

@yo' Bruno did write some code to link pgf to l3fp by swapping out the maths engine: whether that might help depends on where the issue is

@JosephWright I'm sure I'm not dividing by zero.

@yo' Sure: I mean whether the issue is in the bits of pgf that an be swapped out by changing the math unit, or the bits that can't

@DavidCarlisle do you know if it's possible to di1n8ji1t1et or is it just one per segment

@alfred yes of course. You really need to read the full specification of the pattern syntax before starting.

@JosephWright well, replacement of -- by to[out,in] should be about some computation of control points.

btw, probably I found a bug in either TikZ of fontspec (investigating)

@yo' Yes, seems likely

@JosephWright something to do with letterspacing in TikZ nodes

@alfred but most patterns are shorter than that, so they match more words

I saw people bathing in lake Garda. Probably some Britons who believed it was summertime, because the temperature was above 25 ℃.

@egreg if it was above 25 it must obviously be summer

@egreg Sounds about right

@DavidCarlisle Summer weather here I think

%\input{newutf8}
\documentclass{article}

\usepackage[utf8]{inputenc}
\begin{document}

äc

äcd

䪷ò
\end{document}

@JosephWright @egreg ^^ (save as latin 1) I'm contemplating making th eabove gov ebetter error messages (and not run off the end of the file) all non-leading bytes in utf-8 should have bits 11xxxxxx (and not be \end :-) so for example

    \def\UTFviii@four@octets##1##2##3##4{\csname u8:##1%
                           \string##2\string##3\string##4\endcsname}%

could be \long so doesn't give the primitive error on end of para and then check each of #2-#4 is a character in the expected range or if not give a malformed utf-8 error rather than as now a "character not set up" error for a spurious number. But to catch the case of running off the end of the document you'd have to take the arguments one at a a time not sure it's worth it high cost to good documents to catch an error...

@egreg bbc today has an article on the effect 25C has on the British public bbc.co.uk/news/uk-41620071

@DavidCarlisle Too hot!

@DavidCarlisle Yes, 25℃ is a temperature Britons are not used to.

Having fun reading xcolor carefully

@JosephWright That seems quite ambitious.

@HaraldHanche-Olsen you mean reading the doc? Just re-implementing the code is easy of course.

@DavidCarlisle :)

@DavidCarlisle Currently thinking about colour models (which are needed, forcing conversion, ...)

@DavidCarlisle Colour is a damned messy subject. Implementing code is indeed easy, at least in comparison to figuring out what the code should do in the first place.

@HaraldHanche-Olsen Oh, I'm largely aiming to pick up the ideas in xcolor and add to expl3

@DavidCarlisle We seem to be heading toward clarity :)

@PauloCereda Is Cuphead any good?

@DavidCarlisle i compiled a list of 170 patterns but can execute because of a ! Conflicting pattern ignored. l.185 }

@alfred well don't define conflicting patterns:-) you have specified patterns that assign two different numbers to the same pair of letters

from an earlier post here : This can happen after an update or when new languages are installed: LuaLaTeX gets confused about the numbers associated with the hyphenation patterns. Recreate the LuaLaTeX format!

I filtered out anything I think was similar..

@alfred I doubt that's an issue as you're not changing the patterns but adding a new one hard to know without any code

try with xetex--ini

it's your exact code, plus my new patterns

@alfred yes but naturally:-) it's complaining about your patterns not my code:-)

no doubt

@alfred stick the \patterns command here vvv

here is the code overleaf.com/11605254jfdfhhdgcrpt#/43904895

i2o, i2o looks like a duplicate to me

right, i deleted most of them

@alfred and 8jt jt1 is a duplicate for jt you want 8jt1 I guess (although having digits at the start or end is a bit uncommon I think)

I deleted both of them, still nothing

kë is also duplicated 1kë and kë1

still nothing, I also put on comment the babel package, it's just the patterns and the text now

checked with xetex

\catcode`\{=1
\catcode`\}=2
\catcode`\ë=11 \lccode`\ë=`\ë
\patterns{
1anit
1bër
1cë
1dik
1dor
1gji
1gor
1hej
1him
1hmë
1hrra
1je
1jën
1jës
1kën
1kës
1ku
1llit
1mas
1më
1min
1mit
1ndim
1ndin
1ndjes
1ne1
1nin
1nit
1no
1nor
1nte
1ntin
1on
1onte
1po
1pojë
1ra1
1ras
1rën
1rit
1rjet
1rrë
1rrin
1rrje
1shë
1shëm
1shtjes
1sia
1sit
1sko
1smë
1soi
1suar
1tani
1tër
1tët
1tit
1tris
1tuar
1tum
1ve
1yrje
8jt1
a2e
a2i
a2o
a2u
a2y
ba1
bes1
blio1
çmo1
ço1
d8h
dër1
di1
do1
dre1
e2a
e2i
e2o
e2u
e2y
fra1
fry1
fu1
g8j

I put on comment the first half to the end of 1xxx and than the inverse, the problem is between the two

it compiles with just one half

\documentclass is a latex command and you are using plain tex (or initex) but as I said before you need to preload them into a format for xetex, that's why the file I just suggested had the form it had it can be used with xetex ---ini zz.tex which doesn't do a lot useful but reports all the duplicate patterns in your original

@alfred you would get exactly the same error from tex or pdftex or luatex, you need xelatex or lualatex or pdflatex

I'll try that

@JosephWright yes, the art is very stylish and polished, and the game play seems fluent. You might need to remap the keys, as the default setup is not good IMHO. Beware, for that game is a cartoonish version of Dark Souls. You will die a lot. :)

@alfred you might want to put \bye at the end of the file otherwise you will just get to tex's * prompt

@DavidCarlisle I am in the mass, can I call you mean?

@PauloCereda no

@PauloCereda you can be impressed by knowledge of Albanian hyphenation though.

answer Sorry, but xetex did not succeed.

@alfred then you did something wrong, what error message did you get

@DavidCarlisle You are my hero. :)

@alfred if you put the commands I posted above into zp.tex then xetex should show

$ xetex --ini zp
This is XeTeX, Version 3.14159265-2.6-0.99998 (TeX Live 2017) (INITEX)
 restricted \write18 enabled.
(./zp.tex)
*
! Emergency stop.
<*> zp

which means the patterns were read without error

directions to a log but no error message. I saved a file zz.tex and opened a cmd, executed xetex ---ini zz.tex from the folder with zz.tex.

@PauloCereda guess what we had for dinner, it went with a nice plum sauce and roast potatoes

@DavidCarlisle oh no

@alfred did you really have three --- it should be one or two

@DavidCarlisle You're constantly violating the Cereda Convention for Duck Rights .... @PauloCereda :-(

@ChristianHupfer who cares?

@alfred put \end at the end of the file not \bye then it will end cleanly in initex

@DavidCarlisle The whacky Brazilian Bloke, I think ;-)

@ChristianHupfer ooh a convention

@DavidCarlisle you are mean

@ChristianHupfer ❤️

@Johannes_B can you answer a template question ...

@PauloCereda Yes, but not a convention for Star Trek Nerds ;-) ...

@Johannes_B vvv

@PauloCereda 'Citing' a song by The Animals: `Don't let you be misunderstood' :-P

@alfred see at the top where I made accented e's letters oh it's the c cedilla add a similar line for that, how come that didn't error for me, odd..

it doesnt recognize ç

@PauloCereda: I am pretty sure that @DavidCarlisle meant halibut with plum sauce and roast potatoes ;-)

@alfred yes add \catcode`\ç=11 \lccode`\ç=`\ç to match the line for e-umlaut but that's expected what isn't expected was that I don't see that error when I try it

uhf

works

the first code, not the xetex one, i deleted the two patterns with ç

wait, false flag

forgot about putting in comment the first half

@alfred yes that's good (if you try same with your earlier version you would have had the duplicate pattern error messages)

@DavidCarlisle oh

@DavidCarlisle too much TeX nonsense lately in this chat room.

:)

\let~\catcode~`76~`A13~`F1~`j00~`P2jdefA71F~`7113jdefPALLF
PA''FwPA;;FPAZZFLaLPA//71F71iPAHHFLPAzzFenPASSFthP;A$$FevP
A@@FfPARR717273F737271P;ADDFRgniPAWW71FPATTFvePA**FstRsamP
AGGFRruoPAqq71.72.F717271PAYY7172F727171PA??Fi*LmPA&&71jfi
Fjfi71PAVVFjbigskipRPWGAUU71727374 75,76Fjpar71727375Djifx
:76jelse&U76jfiPLAKK7172F71l7271PAXX71FVLnOSeL71SLRyadR@oL
RrhC?yLRurtKFeLPFovPgaTLtReRomL;PABB71 72,73:Fjif.73.jelse
B73:jfiXF71PU71 72,73:PWs;AMM71F71diPAJJFRdriPAQQFRsreLPAI
I71Fo71dPA!!FRgiePBt'el@ lTLqdrYmu.Q.,Ke;vz vzLqpip.Q.,tz;

@PauloCereda better? ^^

@DavidCarlisle much better. :)

Conflict with ti1 (filtered out through putting the rest in comment)

also si1 te1

I dont see a logic why

@alfred they were the ones I changed, you had 1si and si1 and I changed to 1si1

@alfred the version I post above has no duplicates I think

@alfred etex reports the exact duplicate, if I put back si1 I get:

$ xetex --ini zp
This is XeTeX, Version 3.14159265-2.6-0.99998 (TeX Live 2017) (INITEX)
 restricted \write18 enabled.
(./zp.tex
! Duplicate pattern.
l.139 si1

?

@alfred that's how I fixed the patterns that were on overleaf

ok, I thought i had the same set

it works