2:42 AM
@marmot -- (coming late to the discussion.) i think the "do-it-for-me" questions that have raised the most annoyance are ones that were clearly homework questions from a student. in the past there have been times when multiple instances of exactly the same image with the question "how to get this result?" have appeared within a few hours of one another. usually these have been challenged, but since the challenge is in a comment, they aren't easy to find. but the annoyance lingers on.

3:28 AM
@barbarabeeton I fully understand that you find them annoying. After all they dilute the real questions. I do, however, not see why any one needs to justify oneself when answering them. (For a while I thought that answering them will encourage more users asking such questions. However, I do not think any more that is a real issue. Rather, it seems to me that these are mostly independent individuals. Of course, I do not have the possibility to check their IP addresses.)

4:10 AM
^^^^ finally I can cast proper shadows! I'm really relieved. No more cold and long winters! ;-)
3

2 hours later…
5:52 AM
@marmot I don't want you to justify yourself for answering, I would like to make the OP know that, even if s/he got an answer, it is not fair to post a just-do-it-for-me question, s/he got an answer only because is a new user, etc., not because the question is perfect and s/he can continue to ask such question in the future. Of course, this is my opinion, everyone is free to think differently, I don't want to establish new rules I only think it should be better doing it, also for the OPs.

6:19 AM
@marmot Further clarification added to my Meta post

2 hours later…
8:13 AM

8:33 AM
Morning all

@JosephWright Morning quack :)

9:09 AM
@ShreevatsaR Hi mr. velociraptor!

hello!

9:56 AM
Morning all
20

There have been a few questions here on MSE along with some on various child metas about how fixed/permanent the new site redesigns are. I'm posting a new question rather than answering so that we can have more real estate for dialogue (in answers) and because I was missing all of your pings from...

10:16 AM

@PauloCereda Probably not worth it

@JosephWright thought so

10:39 AM
@JosephWright But maybe we can lobby for getting back the fonts and some colors.

@egreg Fonts are explicitly listed as network-wide

@JosephWright Fonts are set network-wide with either a serif or sans-serif stack. Which fonts are in the stack may change as we're realizing the current stack isn't optimal for all of our sites, both technically and for design reasons.
@JosephWright We could see what serif font they're proposing.

11:09 AM
@egreg I guess
@egreg Hope you like the latest PRs :)
@DavidCarlisle Seen meta.stackexchange.com/questions/318910/…? Submit xii?
@UlrikeFischer, @DavidCarlisle Hope you are both enjoying my email rants :)
Waiting for my domain to move ... so I can finish of the move of texdev.net from WordPress to Jekyll ...
.... so while I wait, I'll do a beamer release :)

12:12 PM
@egreg comic sans would be good, cheer the site up a bit.
2

@CarLaTeX There is no reason for you to justify yourself either. ;-) I understand that this is only your opinion, agree with you that it is worthwhile to inform users about the best practice (and it is mostly the same users who ask newcomers for an MWE). However, I just don't want to have to justify not adding such a comment when I do not feel like adding one.

12:39 PM
@marmot What I'm telling is that I don't want to oblige anybody, it's only a suggestion, you can follow it or not as you like

3 hours later…
3:33 PM
@TeXnician As for your comment here: I fully agree. But doesn't that also apply to AboAmmar's "answer"? (... which got 3 upvotes even...)

@marmot it does but one answer was posted after the question was clarified and one before

@CarLaTeX Yes, I guess we more or less agree. In the end it is just a wording issue. I guess most users will agree that explaining newcomers why an MWE can be useful is a good idea. However, several users read in your post a suggested that everyone should do that. You got two answers that suggest that their authors read it this way, as well as a lot of comments.

@marmot and one other answer should be downvoted for gratuitous use of tikz

@DavidCarlisle Yes, but at this time the "answer" had one upvote, not three. (Please feel free to downvote... but if you at the same time argue that the answer that is not to the point deserves upvotes, I find that a bit hard to understand.)

@marmot oh I gave up trying to understand the voting logic ages ago, I can spend two days on a tex answer and get a couple of votes or 30 seconds stealing an image from wikipedia and get several hundred. people vote for what they vote for, Some people even voted for brexit apparently
6

3:46 PM
@DavidCarlisle Yes, sure, I know all that. I am just thinking that the user whose answer got deleted (and does not have the opportunity to see earlier versions of the question) may feel that there are double standards. IMHO the other incorrect answer should not get upvoted either.

@marmot Well, one problem with review is that you see the question and the answer and based on the question it did not answer the question. But I would agree with @DavidCarlisle that sometimes timing is everything.

@TeXnician So far I agree. But I do not see why the wrong answer should get upvoted after the OP clarified that this is not what they asked for. There are tons of questions which only get clear after edits.

@marmot Well, I do not understand that either, but adding another answer that basically even duplicates the content of another answer is definitely not quite useful. Nobody knows why some people vote on answers that do not answer the question, sometimes it's just the interesting approach (without wanting to be disrespectful it's probably not the case in that answer) and sometimes it's because it has a colorful picture (well, it has).

@marmot There are double standards. There are actually thousands of standards - every user has his/her own. And quite a lot probably upvote only because of a nice picture. It doesn't make much sense to care about it (my current goal is to get 200.000 points with less than 10 gold badges ...)

@TeXnician Well, probably I should not say this in public, but the OP of the "not-to-the-point" answer is IMHO the last person who should complain if somebody just copies their answer and posts it, because this is the main skill of this special OP.

3:56 PM
@DavidCarlisle I didn't really could look yet (just came home from a chess match). Is the pdfsavepos problem only with xetex?

@UlrikeFischer I could offer some bounties on questions you have answered. AFAIK you don't get gold badges from harvesting bounties, do you. (I am only concerned that this could be regarded as voter fraud.)

@UlrikeFischer I haven't checked but it will be pdftex as well, usual tex--xet issue that the box gets written backwards so non-text related whatsits get handled in the wrong order

@marmot no need. I think I will manage - I only will have to answer miktex questions only in comments and leave biblatex to moewe ;-).
@DavidCarlisle but luatex should be ok I guess?

@marmot I think I badly wrote that post, that's why I added the clarifications. As you know, I'm not native English speaker.

@UlrikeFischer it has a better chance of being correct:-)

4:01 PM
@CarLaTeX Which is good, so you do not put pineapple on a pizza. ;-)

@marmot Indeed I am always right only when I talk about pineapple pizza

@UlrikeFischer You can afford two more gold badges. (As long as they do not change the rules and give you 2 gold badges if your score reaches 2000 and so on.)
@CarLaTeX I agree. And I guess a meta post along the lines: "Why should one never put pineapple on a pizza?" will be very well received, at least by those who know what pizza is. ;-)

@marmot Of course, but it would be closed as off topic. Btw, I'm thinking of a seasonal challenge for Xmas...

@DavidCarlisle ;-) Do you have an idea how to get the commented line working (that's for github.com/ho-tex/hyperref/issues/63):
\documentclass{article}

\begin{document}
\let \mychar !

\ifnum ! <100 \number! \fi %fine

%\ifnum \mychar <100 \number\mychar \fi %error

\end{document}

@UlrikeFischer steal code from bm (uses \meaning) and dissects the result

4:17 PM
@DavidCarlisle I feared it would boil down to this. Injecting this in the middle of the hyperref processing could get challenging ;-(. With luatex I could use the token library (if I dared to use a function described as "experimental" in the manual) but this doesn't help with xetex.

\documentclass{article}

\begin{document}
\let \mychar !

\ifnum ! <100 \number! \fi %fine

\def\zzz#1 #2 #3#4\relax{#3}
\edef\zz{\expandafter\zzz\meaning\mychar {} {} {} \relax}
\ifnum \zz <100 \number\zz \fi %error

\end{document}
@UlrikeFischer ^^

1 hour later…
5:49 PM
@DavidCarlisle I think it doesn't work ;-(. Basically the whole title is fed to the converter with an \xdef and as the code is not expandable ...
\def\hy@tempa{some title with \mychar zzz}
\xdef\hy@tempa{\expandafter\HyPsd@DoConvert\hy@tempa\@empty\@empty\@empty}
@DavidCarlisle more or less like this ^^

6:36 PM
@UlrikeFischer a better solution really is not to do \let\mychar! and do \edef\mychar{\string!} instead. Are there lots of csnames \let to character tokens around in the wild?

@DavidCarlisle tons:
\documentclass{article}
\usepackage{unicode-math}

\begin{document}
\show\mitOmega
\show\mitnabla
\show\mbfitChi
\show\euro
\end{document}

@UlrikeFischer those ones don't count (as we know who to blame)

I’m looking for some big/complex LaTeX documents using fontspec that I can use for testing/benchmarking my font loader and shaper.

@DavidCarlisle Imho it is the largest batch (difficult to search for this time of command). I will ask Will if he can change it.

6:51 PM
It is funny how many years I spent hacking various TeX engines and packages and I never got to use it seriously, that I don’t even have documents to test with.

"time of command" ? I seem to be a bit tired ...

7:12 PM
@UlrikeFischer either change the default or if that's tricky hyperref could input unicode-math-table.tex while setting up the bookmark mapping and locally define them all (perhaps:-)
@UlrikeFischer I didn't even notice:-)

@DavidCarlisle I added an issue to unicode-math.
@DavidCarlisle ;-)

@KhaledHosny Will may have something in his fontspec tests, the main latex2e test suite is rather lacking tests in that area. However since the tex macro side of things isn't really affected, you might be better just to take "simple" text paragraphs as borrowed from web pages in whatever scripts to see if harfbuzz in luatex does as well as a browser?

7:39 PM
@KhaledHosny I have a number of tests in the luaotfload github (a lot stolen from fontspec/unicode math), but they are normally rather short as this is more suitable for l3build tests.

@DavidCarlisle I do that already, I’m trying to test real world files, you never no what you are missing.
know*

@KhaledHosny Oh OK (I don't use tex at all apart from small examples copied from this site:-)
@ShreevatsaR may have some examples

8:00 PM
@DavidCarlisle I just saw that ifluatex is in texlive in its own package. Did Karl do this?

@UlrikeFischer issue number 1 :-)

8:16 PM
@UlrikeFischer yes Karl had that set up before ho-tex was set up, and added that issue on day one, if I finally get round to splitting this up I'll finally get tp that issue

8:28 PM
@KhaledHosny As far as real-world examples go, here is something that someone used to typeset a 400+ page book with xelatex: github.com/shreevatsa/nmisra-arapv
(Note I'm just mirroring someone else's LaTeX sources; I don't know anything more about it, but I just now tried compiling it and it seems to work (with copious warnings), after fixing the font name.)

8:52 PM
@DavidCarlisle I’m in good company then :)
@ShreevatsaR Thanks!

@UlrikeFischer It doesn't work yet but... github.com/ho-tex/oberdiek/tree/split

@DavidCarlisle wow. I better not interfere. How do you plan to handle the hobsub problem? Will they pull from the parallel folders?
@DavidCarlisle I found at least a luatex solution which seems to work for the \mychar problem:
\documentclass{article}
\usepackage{luacode}
\usepackage[pdfencoding=unicode]{hyperref}

\begin{luacode}
function getcharactercode ()
local t=token.get_next()
tex.print(t.mode)
end
\end{luacode}

\makeatletter
\def\HyPsd@Char#1{%
\ifnum\directlua{getcharactercode ()}#1<128 %
\@backslashchar 000\Uchar\directlua{getcharactercode ()}#1%
\else
\ifnum\directlua{getcharactercode ()}#1<65536 %
\expandafter\HyPsd@CharTwoByte\directlua{getcharactercode ()}#1!%
\else`

@UlrikeFischer something like that, it's on a branch so I can easily junk it all if it doesn't work, I just decided to do it in last few minutes so it's not exactly a fully worked plan yet:-)

@PauloCereda ^^^^ choko ducks.

9:07 PM
@UlrikeFischer oh more edible ducks?

@DavidCarlisle a moral conflict for Paulo ;-).
5

1 hour later…
10:38 PM
@UlrikeFischer yay oh no but it's chocolate but it's a duck but it's chocolate but it's a duck oh no quacks in despair

@PauloCereda dinner and pudding

@DavidCarlisle oh no