12:04 AM
@DavidCarlisle Oh I see, thanks again!

@JosephWright @UlrikeFischer if I export TZ=CET before running l3build the luatex passes, so probably l3build should set GMT if running with a fixed epoch (I still get the xetx subset formt diffs)

2 hours later…
2:08 AM
@UlrikeFischer — regarding Harfbuzz, er, right, I've just added it back in as a synonym for HarfBuzz :) Using case fold for option handling would be good, although I might look to see whether it should be an addition to l3keys instead. WDYT?

2:48 AM
@yo' -- Our car was almost like that when we went out to dinner, except that it was really ice when we went to scrape it off the windshield. Winter is coming.

1 hour later…
4:08 AM
@barbarabeeton Hi Barbara.
@barbarabeeton You might know this.

@HenriMenke -- Know what?

@barbarabeeton Why is there a space between a \mathop{} and a \mathinner{}?
@barbarabeeton E.g. \sin\left(\alpha\right).
@barbarabeeton There will be a space between the mathop atom and the implicit mathinner atom of \left...\right.
@barbarabeeton Do you know why Don chose TeX to insert a thin space there?

@HenriMenke -- What's another example of mathinner? I think that \left...\right isn't the best illustration. For \sin x, "x" is mathord, so not relevant.
(I'm at a disadvantage at the moment -- all my books are in boxes, so I have few resources to check.)

@barbarabeeton The only other thing that's mathinner by default are the dots, i.e. \def\ldots{\mathinner{\ldotp\ldotp\ldotp}} etc.
@barbarabeeton The question could also be paraphrased as: Why is \left...\right implicitly converted into mathinner?
@barbarabeeton Related:
6

Which constructs in (La)TeX and (La)TeX packages have classification \mathinner? The TeXbook mentions: fractions \left...\right constructions \ldots, \cdots, \ddots Which fractions - all? Do other frequently used packages (e.g. AMS-LaTeX) define other constructs (e.g. I'm assuming all types ...

@barbarabeeton I think I'll ask a new question.

4:23 AM
@HenriMenke -- I'm not sure I'm going to be able to answer this. plain.tex defines only the dots as inner. The answer is going to have to come out of the TeXbook, and all I can get at easily is the file (via ctan).

@barbarabeeton You can also use the illegal copy from ctex.org
@barbarabeeton I just realized why there is space between op and inner. Take for example \sin{2 \over 3} (fractions are also inner).

@HenriMenke -- Yes, fractions definitely need a space after \sin. The situation with \left...\right isn't so clear, but I think it's assumed that a little extra space between \sin and an oversize open parenthesis is advisable. What's the space between mathop and mathopen? (That would be the \big[g]l situation.)

@barbarabeeton The space between op and open is 0, i.e. no space, compared to op and inner which is 1, i.e. thin space. Therefore \sin\left({1 \over 2}\right) will have different spacing than \sin\biggl({1 \over 2}\biggr).

4:41 AM
@HenriMenke -- Okay, I'm looking at the illegal ctex copy. (I don't know how many times I've asked for this to be taken down! It's the original version, not the current one.) There's an extended discussion about this, but it only says what does happen, not why. A question for DEK if you're ever at an "all questions answered" session.

@barbarabeeton I see that for large parentheses the effect is actually relatively small, but becomes more pronounced with square brackets or vertical bars. The main complication arises when people use \left...\right for each and every delimiter in their document.
@barbarabeeton When is the next one? I should write down all these questions, because when I met him at TUG2019 I didn't ask him anything.

@HenriMenke -- The use of \left...\right everywhere is just plain wrong. I'm sure that's stated clearly somewhere in the TeXbook. (But I'm not going to look it up.) People are just lazy, or, being a bit more generous, poorly taught.
@HenriMenke -- Unpredictable. If some university invites him to get another honorary degree and he accepts, that is usually accompanied by a q&a session. But I don't always learn about such events beforehand. I think he posts them on his web page at cs.stanford.edu, but perhaps without a lot of lead time.

@barbarabeeton When I write a letter to his office, do you reckon he'll read it or maybe even answer it?

4:57 AM
@HenriMenke -- It will be put into a pile for later attention. I can't predict when "later" might be. On the other hand, this seems worth trying to get an answer, so send email to Karl as if it were a bug; as subject, put "not a bugf, but explanation wanted". State the situation as succinctly as possible, with reference to mentions in the TeXbook (I think it's chapter 10), and say I suggested that you do it. That way, Karl's wrath will be directed at me, not at you.
The "reviewers" know a lot more about this than most mere mortals, and maybe they have some inkling of what was in Don's mind when he decided this. (I know who the "reviewers" are, but won't tell; Don has given them his blessing to do the preliminary checking before anything is sent to him.) The end of 2020 is the next scheduled DEK review of the collection.

@barbarabeeton Okay, thanks for the advice. I will write Karl, maybe he knows the answer.

5:39 AM

2 hours later…
7:33 AM
@yo' ooh
@DavidCarlisle oh no
@HaraldHanche-Olsen yes I also gave it a try! Nice software indeed!

8:10 AM

8:22 AM
@DavidCarlisle ^^
He left the teabag in... OH MY
Also, he thinks the most surprising thing about being PM is getting a curry delivered.

8:44 AM
@PauloCereda now I spent more time this morning on useless British politics than on my Master's thesis, thanks :(

@Skillmon oh no
@Skillmon You can do it, mr. rabbit! <3

@Skillmon the teabag comment wasn't sufficient warning to stay away?

@HaraldHanche-Olsen it's worse: he adds milk only 16 seconds after putting hot water in, and he left the teabag while pouring milk.
@JosephWright had tea during TUG 2018. :)
Also, he's very tall.

@PauloCereda όόη

@HaraldHanche-Olsen ooh

8:54 AM
@DavidCarlisle I ran a few tests with a hallo world. The string changes arbitrarly, but if I set set SOURCE_DATE_EPOCH=20161010 and then do xelatex --output-driver="xdvipdfmx -vv" the end shows always CQPECL+LMRoman10-Regular. So it is clearly related to the epoch setting. I need to check how l3build sets it.

@UlrikeFischer can you set the TZ environment variable to GMT and then try with the epoch settings?
@UlrikeFischer the epoch settings are enabled, and I always get the same string, just not the string you get I get the same as you if I set TZ=CET

@DavidCarlisle I'm just trying to find out if I can do this locally on a cmd line.
@DavidCarlisle I found tzutil, and changed the time zone (and my clock changed) but the string stays put. If I use a bash, how do I set the epoch env there?

9:20 AM
@DavidCarlisle if is not the epoch, it is perhaps the font(name). The message mentioned that the "BaseFont" is involved. Perhaps you are case sensitive or something like that. I think we need @JosephWright, he is on windows but in your time zone.

@UlrikeFischer What if you just precede the command with the env var?
SOURCE_DATE_EPOCH=20161010 xelatex...

@PauloCereda I used export VAR=, seems to work.

@UlrikeFischer cool! :) Preceding vars also works, just in case. :)

@PauloCereda what do you get, if you set SOURCE_DATE_EPOCH=20161010 and then run a hello-world with xelatex --output-driver="xdvipdfmx -vv" file. How is the string before the font name at the end?

@UlrikeFischer Let me try. A simple document?

9:28 AM
@PauloCereda yes, simply article and e.g. "hello"

@UlrikeFischer sorry it was late last night the time zone did not affect xetex/dvipdfmx it made the luatex one pass (that has a datetime in a string)

[paulo@cambridge ~] $SOURCE_DATE_EPOCH=20161010 xelatex --output-driver="xdvipdfmx -vv" hello.tex This is XeTeX, Version 3.14159265-2.6-0.999991 (TeX Live 2019) (preloaded format=xelatex) restricted \write18 enabled. entering extended mode (./hello.tex LaTeX2e <2019-10-01> patch level 3 (/usr/local/texlive/2019/texmf-dist/tex/latex/base/article.cls Document Class: article 2019/10/25 v1.4k Standard LaTeX document class (/usr/local/texlive/2019/texmf-dist/tex/latex/base/size10.clo)) No file hello.aux. LaTeX Font Info: Checking defaults for OML/cmm/m/it on input line 3. LaTeX Font Info: ... okay on input line 3. LaTeX Font Info: Checking defaults for T1/cmr/m/n on input line 3. LaTeX Font Info: ... okay on input line 3. LaTeX Font Info: Checking defaults for OT1/cmr/m/n on input line 3. LaTeX Font Info: ... okay on input line 3. LaTeX Font Info: Checking defaults for OMS/cmsy/m/n on input line 3. LaTeX Font Info: ... okay on input line 3. LaTeX Font Info: Checking defaults for TU/lmr/m/n on input line 3. @PauloCereda the important part are the messages from xdvipdfmx, they should come just before the "X bytes writen"- sorry just saw they are there @UlrikeFischer Ah @DavidCarlisle yes, I got this. Do you get like Paulo "MOOTDW+LMRoman10-Regular" with SOURCE_DATE_EPOCH=20161010? 9:50 AM @PauloCereda Ducks are human!!! Almost. @mickep ooh @PauloCereda Looks like we follow the same accounts on twitter. Or at least, the intersection is non-empty. @mickep ooh not a disjoint set We ducks are very good at naïve set theory <3 Zermelo–Fraenkel set theory is mean @PauloCereda You like the Peano–Duck axioms better? 9:58 AM @mickep ooh @mickep When I hear Dedekind, I cannot help but think of Kirby's King Dedede en.wikipedia.org/wiki/King_Dedede @PauloCereda Hahaha, from now on, I will too. Thanks! :) @mickep ooh penguins @UlrikeFischer no: RRXETV+LMRoman10-Regular-Identity-H CID Type 0C Identity-H yes yes yes 5 0 (@PauloCereda) $ xdvipdfmx --version
This is xdvipdfmx Version 20190503 by the DVIPDFMx project team,

@DavidCarlisle ;-(. I could ask on the dvipdfmx mailing list how this string is computed or someone will have to look in the sources ...
@DavidCarlisle but your font looks different, I get at the end
otf_cmap>> Creating ToUnicode CMap for "c:/texlive/2019/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf"...
(CID:c:/texlive/2019/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf[CQPECL+LMRoman10-Regular][CIDFontType0][6 glyphs][943 bytes])
3202 bytes written

10:16 AM
@DavidCarlisle oh

On other news:
C:\Users\Nililand-Surface\Documents\tests>lualatex-dev test-utf8
This is LuaHBTeX, Version 1.11.2 (TeX Live 2019/W32TeX)

@UlrikeFischer ooh

What is this HarfBuzz for?

@mickep Font shaping

@JosephWright Thanks! I'm reading some presentation from harfbuzz.org to try to understand more about what that really means.

10:25 AM
@mickep color full emoji: tex.stackexchange.com/a/497407/2388, and better script shaping, e.g. burmese tex.stackexchange.com/a/515934/2388
@mickep basically it does what the context fontloader is doing too, but there is much more man power to work on all the scripts and so harfbuzz gets much more scripts right.

@UlrikeFischer Ah, I was just looking into one of the context docs where there are smileys (which I do not so much see the point of having). But burmese and other "cool" languages would be good to have support for.

@UlrikeFischer Same here :)

But other than that, what has harfbuzz ever done for us?

10:40 AM
@mickep currently xetex (rather than luatex) is the system of choice for more or less all the indic scripts, mostly due to the use of HarfBuzz, so luahbtex changes that balance and allows people to use such scripts and access Lua....
@UlrikeFischer which options are you using to get that log?
@UlrikeFischer setting th eepoch and using xdvipdfmx -vvv I get
CID:/usr/local/texlive/2019/texmf-dist/fonts/opentype/public/lm/lmroman10-regular.otf[RRXETV+LMRoman10-Regular][CIDFontType0][6 glyphs][943 bytes]

10:58 AM
hi all
I want the default font of lstlisting (particularly for R language) to be textsf
is there a way to switch the default from ttfamily to sffamily ?

@DavidCarlisle like this it looks like my output (I use -vv) with the exception of the string. Should I ask or do you want to check the sources?
@adityaguharoy yes, listings has a key for the font. (basicfamily)

@UlrikeFischer I probably have no time today during the day

@UlrikeFischer @UlrikeFischer alternatively you can use ding
thanks friend. Cheers

11:03 AM
@adityaguharoy listings has a documentation, texdoc listings

@adityaguharoy oh not that ( I didn't read the previous comment) that would make the whole doc sans serif

11:13 AM
@PauloCereda barbaric!

@Skillmon it is!

@PauloCereda there is no excuse, how could a man like that ever be elected.

@Skillmon ooh

11:30 AM
@UlrikeFischer @PauloCereda as far as I can see the prefix comes from srand which is repeateable on a given machine but for example one stackoverflow answr describes as:
Let me also reiterate that the algorithm used by rand is not specified. Therefore, different compilers (more specifically different standard libraries) can and will produce different random numbers.

Finally, let me say again that you should not use rand if at all possible.
pdf_font_make_uniqueTag (char *tag)
{
int    i;
char   ch;
static char first = 1;

if (first) {
time_t current_time;
current_time = dpx_util_get_unique_time_if_given();
if (current_time == INVALID_EPOCH_VALUE)
current_time = time(NULL);
srand(current_time);
first = 0;
}

for (i = 0; i < 6; i++) {
ch = rand() % 26;
tag[i] = ch + 'A';
}
tag[6] = '\0';
}

@DavidCarlisle ouch, not reliable...

@DavidCarlisle I just sent the question and then you find the answer ;-). Well this obviously means that either they change the code or @JosephWright will have to normalize that Replace every "/".."six uppercase".."+" should probably work.

@UlrikeFischer Might be easiest ..

12:10 PM
@UlrikeFischer the person with strange texstudio that wants to update miktex has posted a screen shot of the messsage their keep getting. But it is in German, do you mind (I haven't keeps up with the German I lerned in school) ? tex.stackexchange.com/q/514361/3929

12:23 PM
@Rmano , @skillmon I put this tl-portable.shskript into ~/texlive/2019/ and I run it to set up the evironment in the current shell:
#!/bin/bash

####################################
# starts shell with TL environment #
####################################

TEXDIR=$(dirname$(realpath $0)) platform=$($TEXDIR/install-tl --print-arch) if test "$platform" = ""; then
echo 'Platform unsupported'
exit 1
fi

if test ! -d $TEXDIR/bin/$platform; then
echo 'Platform unsupported'
exit 1
fi

export PATH=$TEXDIR/bin/$platform:~/bin:$PATH export MANPATH=$TEXDIR/texmf-dist/doc/man:$MANPATH export INFOPATH=$TEXDIR/texmf-dist/doc/info:$INFOPATH 12:34 PM @daleif well that's the standard windows UAC-dialog. But it doesn't explain why texstudio triggers a miktex update or a miktex installation for the admin. 12:48 PM @DavidCarlisle I guess we might need to talk to them about what counts as a Unicode engine @HaraldHanche-Olsen, @mickep, @DavidCarlisle news.sky.com/story/… @PauloCereda Ohh, unforgivable! @JosephWright yes latex.ltx (and iftex) more or less arbitrarily take \Umathchar to mean "fontspec will work" \Uchar is getting close to that.... (\Uchar and \Ucharcat in e-(u)pTeX) @josephwright @davidcarlisle Searching through TeX Live, we found that all packages which detect Unicode engines (= XeTeX and LuaTeX) uses \Umathcode or some other Unicode math primitives; no package used \Uchar or \Ucharcat as a marker. @HironobuYAMASHITA Sure, but I mean what does it leave as the Unicode status of upTeX (I'm even more confused by pTeX): as I understand it, upTeX isn't a Unicode-compliant engine 12:58 PM @HironobuYAMASHITA has to be said \Uchar can be pretty useful, so it's a welcome addition, thanks! (just a bit scary I had to look at all the \ifx\U.. tests we have:-) @mickep you know what else is amusing in that report? They brought a tea expert. :) @JosephWright when did you give me +100 for that?:-) @PauloCereda It is good to take things seriously. :) 1:14 PM @UlrikeFischer thanks, I'm baffled as well. @mickep :) 1:47 PM ooh Macbook Pro 16" @PauloCereda WOW! @PauloCereda But does it come with LuaHBTeX? @JosephWright /hyperventilation @mickep ooh HB @JosephWright Is this brexit-related? @AlexG No 1:52 PM @JosephWright though I think upTeX is not Unicode-compliant in that it has some restrictions in handling multibyte characters, I'm not 100% sure what actually counts as a Unicode TeX engine. If you need some discussion with us, use this place: github.com/texjporg/ptex-manual/issues/7 @JosephWright I was afraid it would be, because it names geographical limits. @AlexG Hah! @HironobuYAMASHITA From my POV, I think of it as not a Unicode engine, as that really means XeTeX/LuaTeX with Unicode math support @JosephWright :-/ @AlexG I am probably kicked out then... :( (I am a UK-TUG member) 2:07 PM @PauloCereda Overlap is explicitly mentioned. So no worries. @AlexG woo @AlexG only "some" overlap ;-) @UlrikeFischer :-) @JosephWright clear enough, thanks @JosephWright So the distinction is Unicode Math? 2:16 PM @norbert Well, to-date we've had XeTeX/LuaTeX which have Unicode math, font loading, Unicode char support throughout, and other engines, which lack all of that; upTeX is currently treated by expl3 as an 8-bit-like engine as it's not Unicode-throughout ... it's complicated, I guess @HironobuYAMASHITA do you know the char range that pTeX and upTeX will have for \Uchar? Hmm, interesting way of thinking ;-) BTW, late comer here - what is going on UK-TUG? @norbert I think actually I'd go with the various tables: can the engine set \lccode, \catcode, etc. for all of UTF-8 @norbert See all the mails there ... @JosephWright how does uptex handle an input like ä does it see one char or two bytes? @UlrikeFischer Er, ask @HironobuYAMASHITA @UlrikeFischer One char, but a bit odd: \def\foo#1{\ifx\relax#1\relax\else\showtokens{#1}\expandafter\foo\fi} \foo é\relax This is e-upTeX, Version 3.14159265-p3.8.2-u1.25-190131-2.6 (utf8.uptex) (TeX Live 2019/W32TeX) (preloaded format=uplatex) restricted \write18 enabled. entering extended mode (./test.tex pLaTeX2e <2019-10-01u02>+1 (based on LaTeX2e <2019-10-01> patch level 3) > 辿. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi l.2 \foo 辿 \relax ? @JosephWright oh, UL-TUG, wow ... 2:27 PM @norbert Huh? @UlrikeFischer pTeX also treats as one byte ... Just read all the UK-TUG mails, seems you will have a funny meeting on Saturday! Enjoy! @norbert depressing reading @David @JosephWright are you sure? If I run this (with euptex) I get two show messages, so it looks as if it sees them byte by byte while with luatex I get only one. @DavidCarlisle Well, it was obviously to me for one specific reason ... 2:29 PM @JosephWright That strange behavior is exactly what we are going to change in the future (not completed yet). github.com/texjporg/tex-jp-build/issues/81 In current (u)pTeX, some byte sequences are misinterpreted into Japanese token in some case (\meaning, \show, etc.) However, changing the longstanding feature can cause a dangerous incompatiblity, so we should be very careful before changing this. @UlrikeFischer Yes, just double-cheked @norbert Oh, I am so looking forward to it @JosephWright I would happily participate, must be a great spectacle! @HironobuYAMASHITA Well yes: at the moment, for expl3 (where I do worry about this) the approach has been to stay away from anything outside of the classical ASCII range @norbert You could always join ... @JosephWright besides the flight to London ... @JosephWright on your test file I get: This is e-upTeX, Version 3.14159265-p3.8.2-u1.25-190131-2.6 (utf8.uptex) (TeX Live 2019) (preloaded format=euptex 2019.11.10) 13 NOV 2019 14:31 entering extended mode restricted \write18 enabled. %&-line parsing enabled. **cc321 (./cc321.tex > Ã. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi l.2 \foo Ã© \relax ? > ©. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi 2:32 PM C:\Users\joseph\Desktop\test>eptex test This is e-pTeX, Version 3.14159265-p3.8.2-190131-2.6 (utf8.sjis) (TeX Live 2019/W32TeX) (preloaded format=eptex) restricted \write18 enabled. entering extended mode (./test.tex(guessed encoding: EUC-JP = euc) > 辰. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi l.2 \foo 辰 \relax ? ) No pages of output. Transcript written on test.log. @DavidCarlisle ^^^ (I added \end) C:\Users\joseph\Desktop\test>euptex test This is e-upTeX, Version 3.14159265-p3.8.2-u1.25-190131-2.6 (utf8.uptex) (TeX Live 2019/W32TeX) (preloaded format=euptex) restricted \write18 enabled. entering extended mode (./test.tex(guessed encoding: EUC-JP = euc) > 辰. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi l.2 \foo 辰 \relax ? ) No pages of output. Transcript written on test.log. @JosephWright you get (guessed encoding: EUC-JP = euc) I do not, presumably after that everything is different (@HironobuYAMASHITA ?) @JosephWright you mean if you hit enter the compilation ends? That's odd. I get also a line about a binary encoding: (./test-utf8.tex(guessed encoding: BINARY = jis). @UlrikeFischer I've just added end: my demo is now: \def\foo#1{\ifx\relax#1\relax\else\showtokens{#1}\expandafter\foo\fi} \foo ä\relax \end @JosephWright with that I still get no guessed encoding and log file This is e-upTeX, Version 3.14159265-p3.8.2-u1.25-190131-2.6 (utf8.uptex) (TeX Live 2019) (preloaded format=euptex 2019.11.10) 13 NOV 2019 14:35 entering extended mode restricted \write18 enabled. %&-line parsing enabled. **cc322 (./cc322.tex > Ã. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi l.2 \foo Ã¤ \relax ? > ¤. \foo ...ifx \relax #1\relax \else \showtokens {#1} \expandafter \foo \fi @DavidCarlisle I'm booting my VM ... 2:36 PM @josephwright @davidcarlisle The misinterpretation depends on the internal kanji encoding (shown as (utf8.euc) -> euc, (utf8.sjis) -> sjis). See the last section of pTeX guide named "strange beast" @HironobuYAMASHITA thanks @JosephWright I made a new file with only the content and it worked like yours. And then I copied a few hundreds line of the mess in my test file (below the code) and it changed. @UlrikeFischer Ah .. pTeX and upTeX distinguishes Japanese characters from byte sequences by its internal kanji encoding; when some byte sequence matches its internal kanji encoding, the byte sequence is considered (or misinterpreted) as a single Japanese character. When no match, that byte sequence is interpreted similar to Knuthian TeX; that is why the misinterpretation depends on the internal kanji encoding. @HironobuYAMASHITA which document? I looked at github.com/texjporg/ptex-manual/blob/master/ptex-guide-en.pdf but I don't think that is the right thing? 2:45 PM @DavidCarlisle Linux VM gives the same: must be partly platform-dependent @DavidCarlisle Sory we don’t build pdf regularly, see github.com/texjporg/ptex-manual/blob/for-tl2020/… @JosephWright it is enough to add a ❤ ;-). @JosephWright On windows, internal encoding of eptex is sjis. Other systems should be using euc. The internal encoding of euptex is always uptex (thus utf-16be) @UlrikeFischer :) @UlrikeFischer It's great that you found that! @HironobuYAMASHITA Ah, right: so like I said, platform is important @HironobuYAMASHITA Thanks for taking the lead in documenting this stuff: it really is useful @HironobuYAMASHITA yes I looked at the tex as well, but wrong branch:-) 2:47 PM Good night. 2 @HironobuYAMASHITA "Therefore, the overall range of <character code> is \emph{not} continuous." you do make sure life is interesting:-) @norbert night. @norbert 'Night @JosephWright arabic works to (and probably also the other scripts currently running around in the test file ;-). But this makes it even more difficult to say if it is an unicode engine or not, if it sometimes reads in 8 bit blocks and sometimes not ;-(. @norbert Night, I hope you enjoy the luahbtex ... @UlrikeFischer Well yes, I think it has to be all-or-nothing @UlrikeFischer I certainly am @norbert night! 3:53 PM @DavidCarlisle @josephwright Another notice: in TL2020 (e-pTeX 190709 or later) you will be able to check "whether an integer is a valid Japanese character code or not" by using \iffontchar (still only described in Japanese github.com/texjporg/ptex-manual/blob/for-tl2020/…) @HironobuYAMASHITA I'll look forward to some English docs ... @JosephWright let's hope it will appear in this weekend ... yesterday, by David Carlisle @cis @UlrikeFischer will confirm that my German is excellent, strangely on a par with my Japanese. you can ask @DavidCarlisle to translate, ^^ @UlrikeFischer ooh @UlrikeFischer ooh @JosephWright ooh 4:21 PM @DavidCarlisle ooh @PauloCereda I realise my English prose style may be hard for a duck to follow @DavidCarlisle quack @PauloCereda Dinner! @DavidCarlisle oh no 4:26 PM @DavidCarlisle exported from the pdf? @UlrikeFischer yes all the links in all the pdf in oberdiek bundle (w3c link checker is checking it at the moment) @DavidCarlisle how did you extract them? I actually found a link this morning but then forgot to sent it: stackoverflow.com/questions/8100592/… @UlrikeFischer seemed I had a pdftohtml from somewhere on this machine then grep for lines with <a href.*http> and make a html file with just those, @UlrikeFischer oh that link is almost same path just using pdftotext rather than pdftohtml as stage 1 @UlrikeFischer useful exercise, showed up some errors... 4:42 PM @DavidCarlisle I'm sure of this, there is always a wrong link somewhere. I bet if you do in the latex documentation ... @PauloCereda, @egreg ^^^ @JosephWright I cannot afford it... :( @UlrikeFischer has adobe rearranged their site, once I have fixed structural ctan links, and http https redirects I get http://partners.adobe.com/public/developer/pdf/index_reference.html redirected to partners.adobe.com/public/developer/pdf/index_reference.html partners.adobe.com/public/developer/en/opentype/glyphlist.txt redirected to partners.adobe.com/public/developer/en/opentype/glyphlist.txt partners.adobe.com/public/developer/en/pdf/PDFReference16.pdf redirected to partners.adobe.com/public/developer/en/pdf/PDFReference16.pdf partners.adobe.com/public/developer/en/tiff/TIFF6.pdf redirected to partners.adobe.com/public/developer/en/tiff/TIFF6.pdf which all 404 after the redirect @DavidCarlisle apple?! @PauloCereda ? 4:57 PM @DavidCarlisle you were quick. :) @DavidCarlisle hmmm something in motion @DavidCarlisle Fixed @JosephWright Oh OK thanks, I changed the link to point to the new one as well. (the w3c link checker moans about redirects as well, just not so much as broken ones) @UlrikeFischer @JosephWright any thoughts on what to do about luatex package, I'd like to not distribute it but I'm not sure I can get away with that? 5:49 PM @DavidCarlisle If I remember correctly the discussion last year it is more or less dead code. But pdftexcmds loads luatex-loader.sty so would need to be adapted first. Then we could replace the luatex files locally by something that errors and run a few test. @DavidCarlisle Lets try to kill it @JosephWright ooh an assassination plot @PauloCereda remember the dragon ;-) @UlrikeFischer ooh the Deutschedragon :) @DavidCarlisle an the new list (pmboxdraw.dtx) github.com/adobe-type-tools/agl-aglfn/blob/master/aglfn.txt @DavidCarlisle for the pdf reference in attachfile2 one could update to 1.7 adobe.com/devnet/pdf/pdf_reference.html, the tiff pdf in bmpsize is here adobe.io/content/dam/udp/en/open/standards/tiff/TIFF6.pdf 6:12 PM @UlrikeFischer thanks I'll update in a bit @UlrikeFischer oh I thought I'd got rid of the dependencies. grr, let me look... @DavidCarlisle I would have done it but I wasn't sure if the github is up-to-date. @DavidCarlisle I'm actually not sure if it loads it really, I only found the name in the file. % \end{Version} % \begin{Version}{2017/01/29 v0.24} % \item % Avoid loading luatex-loader for current luatex. (Use % pdftexcmds.lua not oberdiek.pdftexcmds.lua to simplify file % search with standard require) % \end{Version} @UlrikeFischer but I need to check if that's true @UlrikeFischer only if luatexversion <80 @UlrikeFischer yes I have pushed all my changes @UlrikeFischer I hope ctan will forgive me but for these structural URL change I'm not changing version numbers/dates (as changing them here is painful) so plan is it will go to ctan with makerobust moved to a new repo and holtxdoc.sty with new date/revision with new ctan macros, but all the rest unchanged/ @DavidCarlisle oh yes, that would be a problem ;-). Let's hope. 6:27 PM @UlrikeFischer changing dates would be easy, I could make them all today but doing a global edit to bump the version numbers by 1 would be a bit more fiddly, and would mean the ctan maintainers would have to update every catalogue entry 6:39 PM @DavidCarlisle oh yes the catalog too. They are recalculating the question votes. It will push me over 222222 ;-(. 4 7:05 PM @UlrikeFischer I'm sure we could downvote you back if you want @DavidCarlisle oh no 7:19 PM @DavidCarlisle -- Advice: When you're getting ready to retire (and do it far enough ahead of time), make sure that your home computer is compatible with your work computer. It's a real bummer if it isn't. @barbarabeeton currently my work computer is my home computer... @UlrikeFischer oh I hadn't seem that.I think @egreg and I are safe 7:41 PM @JosephWright miktex will use tomorrow luahbtex for lualatex-dev ;-) Sara Chipps on November 13, 2019 In my very first blog post, I wrote about what a personal experience taught me about the Stack Overflow community. I said we were going to step back and re-evaluate how we deliver feedback, how we can improve content quality, and how we can reduce friction between people. I said that our goal is to have the question asking process be painless and beneficial for new people and Stack Overflow veterans alike. During this re-evaluation period, we noticed something in our reputation reward system. We give anyone who receives an upvote on an answer ten additional reputation points, but only give five reputation points to people who receive an upvote on a question. … @UlrikeFischer @JosephWright nothing in my texlive tree uses luatex-loader and just this one uses luatex /usr/local/texlive/2019/texmf-dist/tex/lualatex/luatodonotes/luatodonotes.sty: \RequirePackage{luatex} (hyperref passes option to luatex but no longer loads it, I'll fix that) @DavidCarlisle Kill it then @JosephWright So I'm earning more points than @egreg and @DavidCarlisle together today? 3 @yo' Looks that way 7:56 PM @JosephWright oh and even that package doesn't load it: % \changes{0.3}{2015/11/12}{Remove package \texttt{luatex} for current % versions of Lua\LaTeX (as it caused problems)} % \begin{macrocode} \@ifpackagelater{luatexbase}{2013/05/04}{}{ \RequirePackage{luatex} } % \end{macrocode} @JosephWright oh no 8:10 PM 5 @AlanMunn -- Thought you might have some reaction to this ... (The east coast analysis is pretty good, provided you're far enough north.) 8:45 PM Can I create this with a loop? \newfontfamily\LexendDeca{LexendDeca} \newfontfamily\LexendExa{LexendExa} \newfontfamily\LexendGiga{LexendGiga} \newfontfamily\LexendMega{LexendMega} \def\zz#1{\expandafter\newfontfamily\csname Lexend#1\endcsname{Lexend#1}}\zz{Deca}\zz{Exa}\zz{Giga}\zz{Mega} or you could use , list rather than \zz if you want I have a list that contains the full names already. \def\lexendVariants{ LexendDeca, LexendExa, LexendGiga, LexendMega } And since I'm doing this for a package, wouldn't it make sense to clean the helper command up? @DavidCarlisle How would I do that with a list? 9:06 PM \def\lexendVariants#1{\@for\tmp:=#1\do{\expandafter\newfontfamily\csname \tmp\endcsname{\tmp}} if you don't have spaces after , just a bit nore if you want to trim space How do I use that? Sorry for all those stupid questions, this is just stuff I never did before and I find it a bit hard to understand... Currently I have this: % Font list \def\lexendVariants{ LexendDeca,% LexendExa,% LexendGiga,% LexendMega,% LexendPeta,% LexendTera,% LexendZetta% } % Font aliases \def\lexendVariants#1{\@for\tmp:=#1\do{\expandafter\newfontfamily\csname \tmp\endcsname{\tmp}}} @BrainStone do you want the list for something else, or just for that, I was writing it as if you would distinctly execute the list rather than save it in a macro first? @DavidCarlisle I want it as a command so people can iterate over the variants @BrainStone that's what I mean, given the def I suggeset you could do: \lexendVariants{ LexendDeca,% LexendExa,% LexendGiga,% LexendMega,% LexendPeta,% LexendTera,% LexendZetta% } and it would do the newfontfamily (but not save the list) I don't know if you want to save the list? @DavidCarlisle Yes I'd like to save the list 9:14 PM @DavidCarlisle : Apparently, after updating grffile, multiple dots in filenames don't work anymore :(. Adding option [v=1] didn't help. @BrainStone ok then probably def\myfontfams#1{\expandafer\@for\expandafter\tmp\expandafter:\expandafter=#1\do{\expandafter\newfontfamily\csname \tmp\endcsname{\tmp}}} then \myfontfams{\lexendVariants{} @AlexG they work for me:-) @AlexG do you have latex 2019-10-01 patch level 3 ? @DavidCarlisle Perfect. Can I undefine it after? @BrainStone undefine what, the loop comand or the font family? @BrainStone Or, with a few extra _ and : \clist_map_inline:Nn \lexendVariants { \exp_args:Nc \newfontfamily { Lexend#1 } { Lexend#1 } } ;-) @DavidCarlisle Yes, patchlevel 3. TL updated today. 9:17 PM @PhelypeOleinik I was leaving that for @egreg but you can step in and be his apprentice @DavidCarlisle The loop command @DavidCarlisle That's what I got in the Team for ;-) @BrainStone you can, \let\myfontfams=\@undefined and in 1993 we would have done, but saving half a dozen bytes isn't usually worth the effort these days I see It's just to prevent exposing commands not meant for outside the sty file @DavidCarlisle And grffile.sty 2019/11/11 v2.1. 9:20 PM @DavidCarlisle That code doesn't work: @BrainStone well I'm confused now as you could call it a name with @ then it wouldn't be usable in the document (that is the whole point of the@ convention) but if it is just used once in the style file why not have a loop directly rather than make a command? ! Undefined control sequence. \myfontfams #1->\expandafer \@for \expandafter \tmp \expandafter :\expan... l.31 \myfontfams {\lexendVariants{}} @BrainStone well it works up to typing errors. \expandafter I don't care how it's done Looping directly is fine @BrainStone no that's what I did and you said you wanted a command.... 9:22 PM @DavidCarlisle My apologies I misunderstood your question I want the list \lexendVariants to stay @AlexG forget grffile the old code doesn't work anymore and the new code does nothing at all. @AlexG can you show an example? And how do you compile? @AlexG the question is why multiple dots don't work, best if you ask a question with a test file @DavidCarlisle I am putting together something. 9:34 PM @PhelypeOleinik Thanks for the code that works Having the expl3 package as a dependency in a package shouldn't be an issue, right? @BrainStone No. Besides fontspec is written in expl3, so you'll be loading nothing more. I see @DavidCarlisle, @UlrikeFischer, @PhelypeOleinik I'd like to acknowledge your great help. Would you like to be mentioned by name in the acknowledgements and if so is the name you use here your real name? @BrainStone Fine with me (this is my real name), and thanks :-) @BrainStone fine with me, and yes it is my real name too. 9:53 PM @DavidCarlisle , @UlrikeFischer Here we are: % Preparations % ln -s$(kpsewhich example-image.pdf) ./example.image.pdf
% ln -s \$(kpsewhich example-image.tex) ./example.image.tex

% pdflatex mwe
% this used to work until yesterday or this morning
%
% ! LaTeX Error: Unknown graphics extension: .image.
%
% See the LaTeX manual or LaTeX Companion for explanation.
% Type  H <return>  for immediate help.
%  ...
%
% l.10 \includegraphics[scale=0.5]{example.image}
%
\documentclass{article}
\usepackage{graphicx}
%\usepackage{grffile}

\begin{document}
(I keep the pdf and its source file together.)

@AlexG no problem here: <./example.image.pdf> is include. Can you show the log-file?

@UlrikeFischer You also need example.image.tex in the same directory. Then it fails. (What used to work until this morning.)

@AlexG oh, I see. @DavidCarlisle ^^^
@AlexG but adding the extension works: example.image.pdf.

@BrainStone no need to mention me, but this is my real name...
@AlexG let me see...

Alright. Will put you in ^^

10:10 PM
@AlexG @UlrikeFischer oh er well that's tricky (if that worked before it was by luck, did it really? I suppose it did if you say so...) the first thing it does is try \IfFileExists{example.image} if it exists, it tries to include that as an image, if it doesn't exist then it tries adding default extensions and so would find example.image.pdf but here \IfFileExists{example.image} is true as tex adds the .tex extension in its internals, and the macro layer can't see that.
@JosephWright ^^

@DavidCarlisle Ah, that, yes ....
@DavidCarlisle goodie :(

@DavidCarlisle why don't one get a similar problem if there is example.pdf and example.tex?

@UlrikeFischer don't you (with \includegraphics{example} I'll test but I would expect that you do..

@DavidCarlisle No, works fine: (./example.tex) <example.pdf, id=1,

@UlrikeFischer who wrote that code? I have no idea how it works.
@AlexG the only way to make that work I think is never to accept an "unknown" file extension and if it is unknown skip the file exists check and move straight to trying default extensions, so it would not attempt example.image but would move straight to trying example.image.pdf, but some people like using weird extensions (and saying they are all eps or whatever)
@UlrikeFischer ah yes in that case it can't guess a filetype as it has no extension at all, so it tries adding extensions. I wonder if I can unifiy them, althugh if you use example.png you really want it to do that not try example.png.pdf etc

10:24 PM
@DavidCarlisle but could it not first split the extension and if it doesn't know it, reassemble the file name and then try adding?

@UlrikeFischer just looking

10:43 PM
@AlexG @UlrikeFischer @JosephWright too late for real testing but
\documentclass{article}
\usepackage{graphicx}
\makeatletter
\catcode\*=11

\def\Ginclude@graphics#1{%
\ifx\detokenize\@undefined\else
\edef\Gin@extensions{\detokenize\expandafter{\Gin@extensions}}%
\fi
\begingroup
\let\input@path\Ginput@path
\set@curr@file{#1}%
\expandafter\filename@parse\expandafter{\@curr@file}%
\ifx\filename@ext\Gin@gzext
\expandafter\filename@parse\expandafter{\filename@base}%
\ifx\filename@ext\relax
\let\filename@ext\Gin@gzext
\else
\edef\Gin@ext{\Gin@ext\Gin@sepdefault\Gin@gzext}%

@DavidCarlisle This file name handling thing reminds me of that quote Knuth speaks of: "Err and err and err again but less and less and less" :-)

@PhelypeOleinik I think 8 uppercase letters a dot and 3 uppercase letters should be enough for anybody.

@DavidCarlisle seems to find the file fine. But with \includegraphics{example.blub} I get
LaTeX Warning: File example.blub' not found on input line 322.

! Missing \endcsname inserted.
\Gin@ext
l.322 \includegraphics{example.blub}

?

! LaTeX Error: Unknown graphics extension: \Gin@ext .

@UlrikeFischer it detects blub recognises that must be you so throws you a random error to keep you amused.

@DavidCarlisle it knows you too:
! Missing \endcsname inserted.
\Gin@ext
l.322 \includegraphics{example.zzz}

10:51 PM
@UlrikeFischer basically it could be made to work, but I have an ifx in the wrong place somewhere, too tired now, also I'm not sure if I want to fix this, it makes the code even more complicated and @AlexG could just add the .pdf on the filename so it works, I'll look again tomorrow.

@DavidCarlisle yes, while it is nice that one doesn't have to add the extension, if people start to confuse the code by adding random periods, they should do the extra work ...

@DavidCarlisle, @UlrikeFischer, @egreg, @PhelypeOleinik Thoughts on github.com/latex3/latex2e/commit/… welcome

@JosephWright Might be nitpicking, but the final example might be misleading for users who do \usepackage{expl3}`. Suggestion: "[...] and the loading of \pkg{expl3} in documents and packages can be done as usual."

@PhelypeOleinik It's a draft: please edit :)

@JosephWright Will do :-)

11:06 PM
@JosephWright don't you want to mention "by loading large parts of the code in the format" and that one doesn't have to load the package?

@UlrikeFischer Er, remember the Plan

@JosephWright ok, so we wait that someone else mentions it ;-)

@UlrikeFischer Yes

@JosephWright should there be also a section that lualatex-dev is based on luabhtex=

@UlrikeFischer Ah, yes: you take that?

11:14 PM
@JosephWright okay, but not today - it's getting late ...