@DavidCarlisle In the bundle of TUGBoats I got given, there is the TUG2000 proceedings. Reading an article on something called xmltex ;)

@JosephWright excellent system

@DavidCarlisle Seems so :)

@DavidCarlisle Article before is a description more-or-less of TeX Live, before it was called TeX Live ...

@DavidCarlisle 'Don't Read Past This Point' :)

@DavidCarlisle There's an article from Frank on xor :)

@JosephWright oh this shows gridtypesetting ...

@UlrikeFischer Not this particular article ...

@JosephWright ? I just looked at it, and it even had a page with lines to show the grid: tug.org/TUGboat/tb21-3/tb68mittel.pdf

@UlrikeFischer That's the final version: I've got the pre-print volume

@Kurt In Italy the summer holidays for the second grade school where I teach start, in particular, mid-July and after about a week after mid-August for the exams of debt recovery training for those who have not passed the discipline(s) in June. I am always the usual unfortunate that I am appointed in committee. I write it to you with all my heart I'm really very tired that my eyes close by themselves. A big hug.

I think nobody can write a PDF crawler in TeX (let's hope Bruno sees this)

@MarcelKrüger regarding your question about named arguments: in \context\tex\texmf-context\tex\generic\context\luatex there is a luatex-preprocessor and luatex-preprocessor-test file which imho shows how it is done.

@Skillmon luatex....

@DavidCarlisle in TeX!

@DavidCarlisle not in Lua.

@Skillmon luatex is tex if pdftex is:-)

@DavidCarlisle as soon as you need \directlua it is no longer TeX but Lua.

@DavidCarlisle I see you are sucking me in to MathML support ....

@JosephWright seems reasonable to me.

@Skillmon and if you write it all in expl3 but a few of the base pdf access commands are implemented internally using lua calls hidden behind \luadef, what then?

@DavidCarlisle We are going to have 'fun' there I suspect

@DavidCarlisle nah, eTeX is fine, but not \luadef. And go Bruno! :)

@Skillmon @DavidCarlisle's point is that \pdfstrcmp, for example, is not available in LuaTeX and has to be emulated in Lua

@Skillmon what is a pdf crawler?

@JosephWright did I miss something? But I think at sometime we will have to do something with mathml ...

@UlrikeFischer a script that essentially extracts information out of a bunch of PDFs. Google's search utilizes a web crawler for example.

@Skillmon well lua can do it in theory, it has built-in library to inspect a pdf, tex could probably do it if you at first uncompress the pdf. But if you want content that it is in the page stream you perhaps need quite some code to assemble this to something sensible.

@UlrikeFischer Chemistry stuff ...

@UlrikeFischer that's why we hope Bruno sees this :)

Quick question about writing % to file. I had a bug report for memoir with page notes (memoirs version of endnotes). The issue for the user is that a space is left at the end of each entry (actually it is a newline as all memoirs pages notes converts spaces to newlines, long story). I know Frank added % to all toc writes in the kernel at some point. In just wondering whether it is enough for me to just add \@percentchar at the end of my writes

@JosephWright do you need more? I have Kajas paper collection in storage.

@daleif Er, no thanks ;)

@daleif Paper archive here now has about 3 gaps

@daleif the toc writing of % was a little complicated due to the double-write as it's written to the aux fist then to the toc so aux gets \@writefile{toc}{\contentsline {section}{\numberline {I}Main}{1}\protected@file@percent } which then writes a % to the toc

@JosephWright I don't have them on display, but I seem to have more

@DavidCarlisle I was looking at that. Interestingly the user who complained is still using TL16 (debian). So I cannot rely on \protected@file@percent. But at the other hand the page notes to .ent printing is done directly in memoir, it is mostly done via \immediate\write#1{\detokenize{#3}} so adding \@percentchar or similar seems safe enough

@MarcelKrüger I just saw the intopdf package. As we are just working on the l3drivers it would be a nice proof-of-concept to check if the package could be written without "do not use command" like \pdftex_pdfobj:D but with ` \driver_pdf_object_new:nn` etc.

@daleif This is more-or-less a full run of TUGBoat ..

@UlrikeFischer I see it wants/needs l3str-convert ...

@JosephWright ;-). (And tagpdf needs it too ...)

@UlrikeFischer Yeah, I know

@UlrikeFischer I think by this time next week that should be sorted ;)

@JosephWright I hope. Which reminds me, will you or should I sent the notes around? (If you can it would be fine, I'm a bit in a hurry currently.)

@UlrikeFischer I'll do it

@UlrikeFischer I think most of it is covered by the drivers, right? Should then work with all engines

@JosephWright it looked like this, and as it is quite short, it is a nice test case.

@UlrikeFischer So it begins ...

@JosephWright saw it. It's my fault ;-) I will answer

@UlrikeFischer Remember, never admit it's your fault, blame someone else (@DavidCarlisle)

@UlrikeFischer any bets on who will be first to get a gold harftex badge?

@DavidCarlisle you could enter the race by describing how to compile in cygwin ;-)

@UlrikeFischer I created a tag wiki for you, compiling in cygwin is easy you just type ./Build.sh except some weeks it builds, some not, depending on er, something unknown, usually.

@PauloCereda Thank you very much!

@DavidCarlisle a tag wiki only for me ;-)

Hi

Do you know of a tool that can beautify the file containing the bibliography?

Right now, some of the entries are indented in a way and other entries in other ways

I am using texmaker (on a Mac)

Can latexindex do that? I have latexindex, but when I execute it, I get Can't locate Log/Log4perl.pm in @INC

I guess I need to install Log4perl

@UlrikeFischer I just tried converting it. Basically it works fine, but it felt more like fighting against l3drivers than using it: All l3drivers commands seem to explicitly avoid expanding their arguments, without exposing the expanding versions.

Yes, I can use \cs_generate_variant:Nn to get :x or :e type versions, but given that at least for pdfTeX/LuaTeX all this stuff is expanded by the backend by default, explicitly expanding seems redundant. Than every object needs a (globally unique) "name", but most objects I generate are only used once or twice, so there is no point in naming them. This just provides problems with when a command is used multiple times, so I ended up auto-generating unique names...

This seems especially odd because annotations do not even support names but only the more traditional _link_last: This seems inconsistent. Also: Is there any built-in support for escaping names/strings in PDF objects?

I am still getting this error even after doing cpan Log::Log4perl

@MarcelKrüger We could define the n-type variants to do expansion as part of their 'native' behaviour, as happens for example for \int_set:Nn and so on. However, usually for anything where there might be non-expandable content that's not really desirable.

@MarcelKrüger I'm not sure what you mean about unique names ...

@MarcelKrüger This is something we have but currently not in l3kernel: one open topic for the very near future is addressing that

@nbro depends on what you mean by beautify. I use Emacs for my editor and it is very good at rebreaking all the lines making the code much nicer to look at.

@MarcelKrüger The driver code is very much work-in-progress, and really is meant to be an abstraction to allow me/@UlrikeFischer/ perhaps you and @AlexG to define more general programming interfaces. It quite deliberately doesn't do anything that is 'common to all engines'.

I want my code to look consistent (e.g. the indentation)

@MarcelKrüger That's a similar argument to \write, though, and there I think everyone is happy with \iow_now:Nn and needing \iow_now:Nx for expansion

@UlrikeFischer I know you like to read documentation

@MarcelKrüger that's just a matter of naming though, if the primitive expands it can be named :x and a non expanding one be named :n, in core code you don't have to define the x variant in terms of the n one, it can short cut straight to the primitive.

@DavidCarlisle Sure, but at present I've defined \driver_foo:n #1 { \tex_primitivething:D { \unexpanded{#1} }, basically

@MarcelKrüger I'll mail later: we should discuss properly

@JosephWright agreed should take to mail, but what I meant was doing that doesn't stop you defining the x version directly to be the :D version, you don't have to define it using the exp args x variant on the n version

@JosephWright should we answer the question that's coming up:-)?

@cis it is bit hard to see the context in that image, google translate suggests Identification numbers which sounds about right (or perhaps "catalog numbers" if it is a known scheme you would just say "K one nine seven" (but that sounds like the scheme for Mozart....

@DavidCarlisle Maybe 'paragraph numbers' or 'parnumbers' is a good word.

@cis well as I say with the clipped image you can't really tell what is being numbered (especially if you can't read the text:-)

@cis I don't think parnumbers is a word (it's not German: you can't miss out all the spaces in a sentence and call it a word:-)

@DavidCarlisle Mmmh... complicated.

@cis if it's numbering the paragraphs then "paragraph number" (like line number or page number) on the other hand if it is a formal catalogue and it is numbering the thing described in those paragraphs, then a different name would be better,

@DavidCarlisle My problem wasn't that it is named :n, that makes sense. But I would like to see a predefined :x version, because I think that (is is used very often and) it is the more natural mapping to the primitives. So while I am fine with having :x names, it feels weird that I have to declare this "primitive" variants and it makes porting "harder": There is no predefined equivalent to the engine primitive.

@JosephWright Sure. You got my address?

@cis I don't know. I only know it's common in certain theological texts.

@MarcelKrüger yes all sounds reasonable, probably an x version will get defined ..... :-) (but I'm not really in the loop at present Joseph and Ulrike leading the way here)

@UlrikeFischer If you are interested, I pushed the first attempt of a l3drivers version on GitHub.

@MarcelKrüger It's in texdoc intopdf :)

@JosephWright I wasn't sure if you read documentation ;P

@yo' Yes, it is known from the book.

@MarcelKrüger I'm interested (after all I suggested that you try) but can't look today, I'm away soon. I fighted also a bit with expansion and I'm not quite sure if it is yet right (it is not always clear when/what the primitive expand), but that's why it is good if there are more tester. I wasn't quite sure about the named objects either at first but in the end I found it is quite nice, it is easier to separate the object creation and the use. @JosephWright

@Skillmon Can I ask you a question about Overleaf?

@cis from which "the book"?

@yo' "The Book", the book of all books, the holy book... :()

@cis well, there the verses are numbered, but not paragraphs. However, for example in the encyclical Laudato Sii and in similar texts, the paragraphs are numbered.

After Fedora 30 update biber in TL 2019 is unable to find libcrypt.so.1. Has anybody else encountered this? I can install biber from Fedora repositories -- could I override the texlive binary with a symlink say in my local texmf tree somewhere?

@musarithmia I remember I had to install some missing library in order to work, but it was straightfoward...

Sadly, my F30 laptop is far from me now...

@musarithmia from experience, only install the missing library from the Fedora repository. :)

@PauloCereda I've been trying to find the right library with no luck. It looks like Fedora has moved from libcrypt to libxcrypt.

@PauloCereda fedoraproject.org/wiki/Changes/…

@PauloCereda Okay, found it. The package is libxcrypt-compat. I'll post a Q&A on the main site.

@musarithmia you could raise on the texlive list. ([email protected])

@DavidCarlisle done

@musarithmia I don't see it yet, but if you are not subscribed I think it is moderated, so it will appear in a bit

@santimirandarp That's a 'piece of string' question: it depends on local requirements, conventions in your subject area, ...

@santimirandarp Not really: wider spacing is mainly about giving room for marking/notes/...

@santimirandarp if you are using \noindent on paragraphs, something is wrong somewhere,

@santimirandarp no, only increase the linespacing if your thesis rules demand it.

@santimirandarp And if you want non-indented paragraphs, then you must separate them with blank lines (load the parskip package to do this). But it's not recommended unless required.

@santimirandarp that is a reasonable style yes, best to use a class designed for that, or use parskip package but that is completely different to changing the linespacing.

@santimirandarp For a thesis it wouldn't be my preferred style, but if you like it and it's common in your language's typographical tradition, then it's obviously fine to use it.

@santimirandarp hmm dubious advice in places

@santimirandarp Wikibooks is notoriously unreliable for LaTeX information.

@santimirandarp mentioning to use \indent for example (typically it won't work in latex)

@santimirandarp The advice to use parskip is correct. So if you like the unindented style, it's fine. It's mainly a matter of personal preference or typographical tradition.

@santimirandarp I would say that reading the site here is a much more reliable source of information, although somewhat harder to navigate.

@santimirandarp by default inline math is designed to fit in the default line space, but yes if you have a lot of inline math you may want to increase the space a bit to maintain an even baseline

main site blocked with TLS errors?

@DavidCarlisle yes :)

@DavidCarlisle Not for me.

@AlanMunn it works in chrome and edge, but FF says there is an issue with the TLS layer and won't show the page

@DavidCarlisle Hmm. I'm using FF. And I have no idea what edge is. :) (well not really)

@AlanMunn Just think Internet explorer but slightly less horrible, and by end of the year it'll be Chrome with an internet explorer disguise

@DavidCarlisle Pretty much what I thought. (I've seen it in my Windows VM that I rarely use, but have never actually tried it).

@DavidCarlisle But now I'm getting the same error. Weird. Only happened once though.

twitter.com/StackStatus/status/1133787602027286529

@MarcelKrüger Taking a look at the driver code, the place that I can see expansion is suppressed is #2 of \driver_pdf_object_write:nn. Would a 'pre-defined' x-type variant (which would be the primitive in some cases) be OK?

@MarcelKrüger On the object business, are you after a 'one shot' function that does away with the need to have separate new and write steps?

@musarithmia mail just got distributed to tl list

@JosephWright Yes, a predefined x-type variant would be great. Right, my idea would be a one shoot function combining new and write. Objects created with such a function should probably be written immediately and not be delayed until referenced. Related to that: According to the documentation, "Objects are only written to the PDF when referenced." At least on pdfTeX/LuaTeX, that's not what's happening. They are written immediately by the write macro. Maybe that could be documented more clearly.

@MarcelKrüger OK, I'll look at that: it's all pretty easy to sort

@MarcelKrüger So something like \driver_pdf_object_now:n(n|x) {<type>} {<data>}?

@JosephWright Yes, exactly this!

@MarcelKrüger On the nx variant, although I see your point about the primitive, there are a few places that are a bit tricky to directly expose and keep things 'clean', so I probably will stick to a standard variant. With it being pre-defined, you won't have to worry I hope.

@MarcelKrüger Give me an hour or so: have to pop out. Assuming this is what you need, I guess a release later today?

@JosephWright That's exactly what I need :) But there is no need to hurry, it is not urgent.

@MarcelKrüger Well this is a 'hot' area, so getting stuff out is good: I'm keeping CTAN busy!

Right, need to rescue some TL2019 DVDs

bostonglobe.com/metro/2019/05/28/…

@PauloCereda @UlrikeFischer ^^^

@MartinSchröder Clearly someone with an agenda. :)

Justified text is easier to read that's why I formatted the source code in xii.tex the way I did.

@AlanMunn good counting to 7, and you're not even a mathematician.

@DavidCarlisle But I play the Peano.

@AlanMunn a good axiom to follow

@CarLaTeX ooooooooooooooooooooooooooooh

@AlanMunn ooh

@DavidCarlisle ooh

@AlanMunn That's very (dede)kind of you to say. :)

@PauloCereda I must confess having to look that one up. :)

@AlanMunn <3

@AlanMunn it's one of those jokes in which the entire audience remains quiet, except three nerds from Trinity giggling nonstop. :)

After all, let's be Fraenkel: who gets those math jokes? :)

@PauloCereda They Wile(s) away the time.

@AlanMunn ooh it's getting better and better

@PauloCereda Hey Paulo, what's new about arara? ;)