@UlrikeFischer What happens without luaotfload?

@MarcelKrüger then there is no kpathsea/mktexpk call.

@MarcelKrüger I'm going to bed now, perhaps we can find tomorrow some other tester.

Hey!

I want to install a new package on my Windows 10 machine (TikZ-UML package)

For do that, I am seeing: How can I manually install a package on MiKTeX (Windows)

Which addresses this other question: Create a local texmf tree in MiKTeX

Anyway, I can't find such texmf

I have the tikz-uml.sty file but I don't know how to continue

I am on MiKTeX Options, and then? :F

This is the sense of 'end' in 'end in itself'. @marmot Unfortunately, I'm stuck on the further translation into Welsh. (That is, I don't know which word to use for 'Zweck' or 'end'.) :(

I do not understand the term "tree" to which it refers or how to create it

@manooooh Do you need to install it manually? Doesn't MikTeX have that package?

@cfr thanks for your reply! Nope, MiKTeX does not have that package installed

@manooooh Then I guess (according to the answer to the question you linked), you have to create the directory structure in your file system and then tell MikTeX to use it.

@cfr yes, but I do not know the meaning of "creating a tree"

@manooooh You just create the directory structure in your file system and then tell MikTeX to use it. Did you create the directory structure?

@cfr Oh wow, this is really old text. I guess it says that "you should act in a way that you never either use yourself or another being (OK, Kant says person, but this was before it was known that marmots are to be treated with respect ;-) as a tool." Of course, I have no idea about Welsh (except that I might warn you that in German AFAIK welsch can either mean from Wales or from Wallis, a Swiss canton.)

@cfr no... a directory structure is... an empty folder where we have to fill it with a .sty file?

Can not we simply paste the tikz-uml.sty file into the folder where the other files that MiKTeX has are hosted?

@manooooh No, you need the hierarchy. What you need depends on what's in the package. You can't just put the .sty in the root of the new tree. For a .sty, you'd do something like <root of tree>\tex\latex. Then MikTeX should add another sub-directory when it installs the .sty file.

@manooooh You shouldn't do that, no. I don't know what MikTeX does in that case, but you shouldn't contaminate the distribution tree with manually installed packages.

@cfr hierarchy for what? Sorry but I do not know how to do that :( :/

@cfr ok

@manooooh That's the bit I then said. E.g. <root>\tex\latex or whatever. (I guess Windows uses `` for separations.)

@manooooh You just make the empty directories however you normally make new directories. Then you tell MikTeX about it. At least, that's what I got from that answer.

@cfr yes I see, but what does <root>\tex\latex mean? You are right with the format: Windows have \ to separate hierarchy

Thank you @cfr for your pattience! Let's see from the answer:

· Create a folder somewhere on your computer – with an important exception. It must not be a subfolder of your MiKTeX installation (it worked this way in older MiKTeX versions, but was changed for security reasons). Avoid as far as possible spaces in the path, it is a potential risk for malfunction. For use with MiKTeX portable the folder must be, of course, on the same drive as the MiKTeX installation, it will be saved as relative path. Here I will use c:\localtexmf.

<root> is wherever you want to put it. tex\latex are the extra directories you need there.

For example, the folder's name will be localtexmf right?

@cfr uhm

@manooooh So then, go wherever you want it to be, make the directory localtexmf. Then go inside that. Make tex. Then go inside that. Make latex. Then tell MikTeX to use localtexmf as a tree.

@cfr so the main folder will be tikz-uml and subfolders are text and inside it a new folder called latex, at first allocated in desktop, right?

@manooooh No. Don't make tikz-uml. And it needs to be tex and not text. The top one will be localtexmf, if that's what you want to use.

@cfr ok, a directory is a folder?

@manooooh Yes.

(Well, on Linux, a directory is a file, but let's ignore that.)

@cfr ok nice! So localtexmf's folder can be allocated at `C:` I think

@manooooh You're just making three folders - each inside the other.

@cfr yes yes

In C:\localtexmf\tex\latex?

@manooooh I wouldn't know. It sounds bad to me, but that's just because I wouldn't create it in / on Linux. I know zilch about Windows.

@cfr don't worry about \ or / :)

Oh, I will follow the instruction of given answers

@manooooh You don't. You get MikTeX to do it, I think. You just tell MikTeX about C:\localtexmf\tex\latex and then you tell it you want to install the package and it will put the .sty file in the right place. At least, I think so.

@cfr yes, @StefanKottwitz says that after creating this directory :)

Then I write you how it went

@manooooh Good luck. Bear in mind that, if it were my machine, I'd wipe Windows and install something else. So my advice is ... er ... not the best informed.

@cfr you are right. Windows is my best friend since I have a computer. Anyway, I hope you know that a Information Systems Engineer works (not yet in my case, I am only a student) with programming languages based on Linux systems... so I will have to change sooner or later

@cfr it worked 100%!! Thank you for your time!! :)

The only warning I am getting when compiling is Package pgf Warning: Your graphic driver pgfsys-dvips.def does not support fadings. This warning is given only once on input line 31.

I will investigate that

Lol:

That answer xD

It seems that the warning does not have a solution. I see forum.mathematex.net/latex-f6/… but it is in French :(. Does anybode know French?

@manooooh It's only a warning. If you aren't using a fading, don't worry about it.

@cfr ok. I suppose that any of us do not want to get a warning for no reason, so I asked

@manooooh Oh yes, there is a full country of people knowing French. ;-)

@marmot hahahaha marmot!! You like to talk technically... I like it

Commençons à parler en français

@manooooh You got the warning for a reason: you are using a driver which doesn't support everything in TikZ. That's important. It is just that you're not using the stuff it doesn't support, so it matters you know, but you don't need to change anything.

@marmot Only one?

@cfr The statement that there is a country full of French-speaking people does not entail an upper bound on the number of such countries. ;-) (BTW, if I am not mistaken, some French will insist that the French spoken in other countries is not real French. ;-)

@cfr ok. Anyway, in order to prevent the warning should I update/change the drivers of my computer?

@Sebastiano Thanks, I hope you are well too. I am not really back here...

@cfr ooh

@marmot ooh

@samcarter @marmot @UlrikeFischer ^^ :)

@PauloCereda Hmm, interesting. In the duck species I am familiar with, the males don't care one whit about their offspring. In eider ducks, on the other hand, females will happily adopt the ducklings of other females. Once I saw a single female with more than fifty ducklings trailing her.

@HaraldHanche-Olsen oh my! 50!

@PauloCereda Yup. They were really hard to count, as you may imagine. I suspect their number might have been closer to 60.

@HaraldHanche-Olsen ooh we need some sort of counter :)

@PauloCereda Do we have to worry about it?

@CarLaTeX No worries, TeX will never catch on. :)

@PauloCereda :)

@PauloCereda aww ;-) Could you compile the plain code here with luatex twice and tell me if you see the kpathsea: Running mktexpk --mfmode line on the terminal at the second compilation? github.com/u-fischer/luaotfload/issues/29

@UlrikeFischer It failed with me, perhaps I have an old binary.

l.55 \ifnum\luatexversion
                         <60 %

$ xetex --version
XeTeX 3.14159265-2.6-0.99999 (TeX Live 2018)

@PauloCereda luatex not xetex!

@UlrikeFischer Oh sorry! For some reason I read XeTeX!

I am an idiot.

kpathsea: Running mktexpk --mfmode / --bdpi 600 --mag 1+0/600 --dpi 600 bbm10

@UlrikeFischer ^^ got it at the second compilation

@PauloCereda no quite awake yet probably. At six in the morning I would be completly unusable.

@PauloCereda so you too. What is your OS?

@UlrikeFischer Linux (Fedora).

@MarcelKrüger ^^^^ What is your OS?

@UlrikeFischer (Arch) Linux.

@UlrikeFischer After recompiling everything (again) and deleting every last cache, I can finally reproduce it too!

@MarcelKrüger interesting. Which cache could change this?

@MarcelKrüger with the generic fontloader alone the font can't be loaded at all (probably because the afm is missing), with the "mode=plain" experiment I get the same message.

@UlrikeFischer Tracing the system calls it looks like $MAKETEX_MODE fails to be expanded in the path. But why?

@MarcelKrüger the docu in texmf.cnf says that If no mode is explicitly specified, kpse_prog_init sets MAKETEX_MODE to /,. Perhaps the library doesn't implement this part. When I add MAKETEX_MODE=/ to my texmf.cnf it works.

@MarcelKrüger it also works if I set an environment variable so as a work around luaotfload could check the value and set it if missing.

@UlrikeFischer The problem is kpse.lookup. kpse.lookup fixes the value of MAKETEX_MODE, but LuaTeX only sets it when the first page is written out. We can fix this by initializing kpahsea earler, but then pkmode/pkresolution would be fixed. Instead we can avoid kpse.lookup and use kpse.find_file instead. There is only one problem: kpse.find_file can only return one file, so after this change multiple luaotfload-blacklist.cnf files would no longer be supported.

@MarcelKrüger I would say multiple blacklist files must stay - it must be possible to have a global and a local cnf. Imho initializing kpathsea earlier is better. Is loaded luaotfload.conf early enough to allow a pkmode/pkresolution setting there if someone really cares arbout it?

@UlrikeFischer I realized that we can keep loading the blacklist through .lookup probably because the filetype is specified as tex. So only lookups with unspecified types seem to cause the problem.

@UlrikeFischer You got a PR.

Any font wizards around?

I'd like to identify this font

from the New SI style guide

bipm.org/utils/common/pdf/SI-Illustration-Guidelines.pdf

$ pdffonts.exe SI-Illustration-Guidelines.pdf
name                                 type              encoding         emb sub uni object ID
------------------------------------ ----------------- ---------------- --- --- --- ---------
GBUZMZ+MyriadPro-Regular             Type 1C           Custom           yes yes yes    242  0
GBUZMZ+SourceSansPro-Semibold        Type 1C           WinAnsi          yes yes yes     28  0
GBUZMZ+MyriadPro-Light               Type 1C           Custom           yes yes yes     35  0

@EmilioPisanty ^^

@DavidCarlisle SourseSans isn't it, and if I give this font-trier fonts.adobe.com/fonts/myriad the unicode Δν, it doesn't seem to be it either

it isn't input as text in the pdf - if I examine it in Inkscape, I can select all the other units, but this one is drawn as a path

=(

@EmilioPisanty so it's not a font then:-)

@DavidCarlisle ¬¬

for unhelpful definitions of "font", you're right, it isn't.

@EmilioPisanty or it was a font but they deliberately expanded out the paths to prevent re-use in a public document or ...

(pardon my annoyance. It is primarily directed at the BIPM and their shoddy incomplete style guidelines)

@DavidCarlisle I don't think it was deliberate. Hanlon's razor points to incompetence, I think.

@EmilioPisanty yet Occam's point to David's solution, as it's the simplest of the pack. :)

@PauloCereda I dunno. "They gave it to a clueless designer" is a pretty simple explanation. But in any case the "why" doesn't much concern me.

I'm re-doing the diagrams in this answer

to match the new style guides, at least for the colour

if I can get all the fonts to match then that would be helpful

for which I'd need an ID for a symbol font based on two characters

oh :)

and the BIPM has gone ahead and changed the task from easy to near-impossible, through (choose: malice or incompetence).

so, y'know. ¬¬.

damn, so close

@DavidCarlisle what did you have in the fancy restaurant yesterday?

@PauloCereda nutella and pineapple pizza, it was excellent.

@DavidCarlisle ooh

@PauloCereda actually American Hot

@DavidCarlisle ooh

@JosephWright failed last night o compile tl, I think I updated cygwin since I last did it and seem to be missing part of the toolchain to compile fontconfig, I added a few things to my cygwin setup and trying again..

@DavidCarlisle Cool: did you see Norbert's mail?

@DavidCarlisle Do you remember what the 'recipe' was?

@JosephWright yes but I'm trying to compile with it as checked in (ie after Akira added your first set) but I was getting errors about not having tools like fontconfig development headers (I just had the fontconfig runtime) not sure why an "update" to cygwin lost some of the optional modules but anyway I added a few more and its getting further. I'm just running ./build --without-x at the top level of a texlive-source git checkout

@DavidCarlisle Trying it now

@JosephWright after the first configure you should be able to recompile just xetex by cd /home/tlbuild/source/Work/texk/web2c then make

@DavidCarlisle I've spotted my latest mistake, I think ... I'll make sure it actually builds then circulate

@DavidCarlisle Ah, that was it, yes ... we should make some nots somewhere!

@DavidCarlisle Doing the same here

@DavidCarlisle Corrected diff (I hope):

diff --git a/source/texk/web2c/xetexdir/xetex.web b/source/texk/web2c/xetexdir/xetex.web
index 7844dca0..1a01db24 100644
--- a/source/texk/web2c/xetexdir/xetex.web
+++ b/source/texk/web2c/xetexdir/xetex.web
@@ -9878,9 +9878,11 @@ the codes for its extensions: |eTeX_version_code|, \dots\ .
 @d elapsed_time_code      = pdftex_first_rint_code + 10 {code for \.{\\elapsedtime}}
 @d pdf_shell_escape_code  = pdftex_first_rint_code + 11 {code for \.{\\shellescape}}
 @d random_seed_code       = pdftex_first_rint_code + 12 {code for \.{\\randomseed}}

@JosephWright build still grinding away I'll let it finish I think, it used to do a full texlive build so would like to get back to that even though typing make just for the web2c bits would be quicker

@DavidCarlisle @egreg look at that link: the list of worst pizzas ever seen

@CarLaTeX sadly no nutella or pineapple

@DavidCarlisle @CarLaTeX Nutella pizza is good

@PauloCereda Nutella is good, pizza is good, but not together :)

@CarLaTeX Nutella pizza is good :)

@DavidCarlisle Still building ...

@PauloCereda Nutella pizza is good if the pizza is a crepe

@DavidCarlisle Ooops

@DavidCarlisle I've stuck with the TL code I had earlier in the year: I'm hoping it all builds ...

@JosephWright I might try in the wsl

@DavidCarlisle That's what I'm doing

@CarLaTeX crepe pizza

@DavidCarlisle Bugger, I hit a typo ...

@JosephWright oh a swear word

@PauloCereda It's been building for like 30 minutes, had to start again ... all becuase I mixed up singular/plural

@JosephWright you could be using Simon's raspberry pi....

@DavidCarlisle :)

@DavidCarlisle That reminds me ... videos

@PauloCereda compiling on windows has that effect on people

@DavidCarlisle Be interesting to see who 'wins' here: I'm on my second pass ....

@DavidCarlisle Probably by the end of the day we'll have a working XeTeX with all the primitives: might be one for my blog (or the team blog)

Speaking of which, also need to finish blog move

@JosephWright I might give up since it failed somewhere, probably I should finish testing the option handler instead

@JosephWright the styling on your blog is still rather 1990s:-)

@DavidCarlisle Sounds good

@DavidCarlisle I know

@JosephWright I thought you might:-)

@DavidCarlisle ooh

@DavidCarlisle I've got a couple of weeks before I have to pay for the current hosting: by then I'll have ripped off enough of the L3 site to have something good enough. I do fancy getting help: I should ask UK-TUG for funds! ;)

@JosephWright do you have a transcript on the discussion about languages? I'd like to read. :)

@PauloCereda I'll see

@PauloCereda Just about to look at videos

@JosephWright cool

@DavidCarlisle Bugger again ... I'd left some experiments around

@JosephWright I just stop short before trying to benchmark the loading time of expl3 with l3benchmark ;-)

@UlrikeFischer :)

@UlrikeFischer Loading is slower with LuaTeX than with pdfTeX (where it's trivial), but I'm not sure we can do anything ... some of that will be Unicode data, for example

@JosephWright @PauloCereda see what I said about compiling:-

@JosephWright It really wouldn't hurt to preload the unicode data in the fomat...

@JosephWright luaotfload loads also quite a lot unicode character data. I wonder if it would make sense to unify this.

@UlrikeFischer We need it at the macro layer ... catcodes

@DavidCarlisle Perhaps, but it would need to be in the right place (I should check I'm right about the time it takes)

@JosephWright and you need for xetex.

@UlrikeFischer Yup, that too

@DavidCarlisle Looks like we are sorted

@DavidCarlisle Also, now I remember how to build stuff :)

@JosephWright ah good (I've been away from machine so not done the latex stuff:-)

Which tags should tex.stackexchange.com/questions/460598/… have? I can tell it's not [xetex] but am not sure what it should be.

@Skillmon used math-mode which is a bit generic but ...

@DavidCarlisle it's at least the right direction. There should be a code this for me tag...

/cat

:(

@PauloCereda I remember today, there is a very short, but very cool and somethimes usefull arara-rule, written by @Clemens. Do you like to update that for version 4? Shall I open a Thread?

!config
# Console rule for arara
# author: Clemens Niederberger
# requires arara 3.0+
identifier: console
name: Console
command: <arara> @{command}
arguments:
- identifier: command
  flag: <arara> @{parameters.command}

PS: I do not understand why ghostscript.yaml and console.yaml are not in the official list. :)

Can anybody tell me if 'Das Menschenbild' is the German translation of 'Leviathan'? Google says it means something I don't recognise, but I doubt that means much.

@cfr I doubt it very much. But which "Leviathan" do you mean?

@cfr I would translate Menschenbild with image of humanity or something along these lines, while Leviathan as a sea monster, something like Moby Dick (but if you are still translating Kant, who knows what he might be up to ...)

@cfr Leviathan is a terrible biblic demon, as to my knowledge.

*biblical

latex is choosing to split a footnote from one page to the next

this seems like quite an unusual decision?

i just set a big penalty on it, all good

@baxx Somewhat unusual, but not at all unheard of: A footnote of more than one line whose origin is near the bottom of the page, always risks such treatment. How could it be otherwise?

@cfr As far as I knew german "Menschenbild" is something like "idea of man" in english. In german we also know en.wikipedia.org/wiki/Leviathan (in German de.wikipedia.org/wiki/Leviathan_(Mythologie)) ...

@UlrikeFischer Hobbes. The students doesn't seem to have cited Leviathan and I was wondering if they'd cited a German translation instead.

@samcarter Na, I'm back to Hobbes now and on German/English rather than German/Welsh.

@baxx That isn't necessarily a good idea. Do you want a big gap in the text instead?

@baxx latex doesn't make the choice, you specified that the penalty for doing that was less than the penalty for doing any alternative..... (In some subjects people have a habit of writing more text in the footnotes than is in the main body and so splitting footnotes is almost inevitable)

@UlrikeFischer thanks, I suppose I need to push to ctan soon....

@DavidCarlisle I had one example with two nearly page length footnotes on the same line. In the footnote there were tabulars and lists - one wonder why they don't simply write everything in footnotes.

@DavidCarlisle but imho it should there only temporarly. A real solution for the page resources would be better.

Good evening everybody

@marmot Please can I disturbe you?

@Sebastiano Sure. What's up?

@marmot Thank you so much for your time. I added a response with a program that is a clone of Xfig but is more recent and more functional. Why did I get a downvote? Do I have to remove the answer or leave it?

@Sebastiano How would I know? I did not downvote. (I upvoted now.) Well, this might be just a coincidence, but some special user also posted an answer here, and I observed that his number of downvotes went up whenever you received a downvote. Note, however, that this might just be coincidence.

@marmot Have you read the daniel's answer?

@marmot I just copied the site's indications, but I didn't steal anything. I inserted a flag to the moderator. This user looks a lot like one who used to insult me.

@Sebastiano I cannot really say much. I never use these tools, nor do I have any idea who posted what on those. There are two logical possibilities IMHO (but I really don't know which is true): either daniel has a point, and you may want to include the source of some information, or he has not, in which case you may simply ask the moderators to take care of this comment.

@Sebastiano I believe you but I am not a moderator. (I also have two special friends on this site, who actually do shamelessly steal codes, but I guess the best way to go is to try to ignore them. The sociology of this site is tricky. ;-)

@UlrikeFischer sure yes but I don't think we can do that by the December release so the patch now is good.

@DavidCarlisle yes, that's why I added it (and I hope without typo ;-)).

@marmot I wrote to you because I think you're a friend. Now I've changed the answer, but accusing me of stealing is a very serious accusation. If I have inserted the link where to download the program, in my opinion it is logical that you go to the homepage. I am not malicious or bad. I always thank you very much for your availability. You are very good and very kind. Thank you very much.

@Sebastiano Really, you hurt those most by ignoring them. All they want is your attention and that you react, perhaps even overreact. Just stay calm, appreciate the fact that you live in a nice country in which people are very considerate (and do not put pineapple on a pizza ;-) and relax. ;-)

@Sebastiano but you really should not ping individuals here and ask why someone else downvoted

@marmot With all my heart, it's true that I'm not relaxed. I am very stressed but believe me I can not stand the free offenses especially from strangers. I always wish you well.

@DavidCarlisle Hello kind David, I'm currently angry. Then I pass me; already I have received so many offenses in this month from a user who I think has been expelled. You're right and I apologize. I am very instinctive and should calm down.

@Sebastiano if you get angry about these things you shouldn't use the site so much. This site downvotes far less (far far less) than most other sites in the network, but the internet is a big place with lots of people from lots of different cultures, so basically unless someone is threatening you with physical violence then just let bad comments (and downvotes) pass by without stressing about them.

@Sebastiano I have posted tens of thousands of posts to here and comp.text.tex and xsl-list etc over the last 30 years, and I think I've only been offended by no more than 3 responses in all that time.

@StefanKottwitz Hi and good evening.

@Sebastiano Hi!

@DavidCarlisle I see, but as you can see, I didn't even answer him. Now I remove my answer hoping that the moderator can remove everything. Thank you with all my heart.

@Sebastiano I recommend you should contact the StackExchange community support and ask that, if it's obvious, they should collect and summarize the accounts that obviously target you with downvotes and comments, and figure out what to do about this.

@DavidCarlisle It's not random, it's targeted over a long time now.

@StefanKottwitz I have removed my answer. This is not the first time that this user has brought me out of place comments. I flagged his comment. But his attitude reminds me very much of a PO that I think was recently expelled.

@StefanKottwitz @DavidCarlisle :-( Sorry to leave, I can't stay in the chat room. I wish you a peaceful rest and david thanks for understanding. I admit it when I'm wrong but you can not attack me continuously. Serene rest for everyone.

@Sebastiano Good night.

@Sebastiano SE can see statistics and conclude if it's bullying or a big, big, repeated happening of coincidences. I conclude the first, but I'm just a mod.

@StefanKottwitz you have more information than me, so OK. But still pinging marmot (or me or TeXnician and asking why someone else unknown downvotes isn't that useful)