TeX, LaTeX and Friends - 2019-07-12 (page 1 of 3)

« first day (3176 days earlier) ← previous day next day → last day (1749 days later) »

3:06 AM

@HenriMenke Thanks! My question was more why it stopped working. Were there changes somewhere else?

3:16 AM

@Skillmon How can one test what a code produces on overleaf without paying them? (I am really not excited about discussions like the one below tex.stackexchange.com/a/499577/121799, there is no way to debug this.)

2 hours later…

4:48 AM

@marmot It broke in this commit to fix this issue and was reverted in this commit.

5:19 AM

@marmot You can use Overleaf without paying them, I have done it for years. I think you can also use it without sign in, for temporary projects.

5:54 AM

@HenriMenke Thanks!

@CarLaTeX I get always asked to create an account, something I do not want to do.

@HenriMenke As one could have expected, all @UlrikeFischer's fault. ;-) (Just kidding!)

@marmot Now I have an account, but at the beginning I didn't. I don't know if the rules are changed.

@CarLaTeX I think they did.

@marmot Ulrike's next fault is ready for distribution ;) github.com/pgf-tikz/pgf/issues/702

Jun 27 at 5:53, by Henri Menke

@UlrikeFischer It seems as if I was just too eager with scoping commands. This fix seems to work https://github.com/hmenke/pgf/commit/8a45771e0c0f2abfb09421fbad69699703d021b5 (at least it fixes your example and it still compiles the manual). Should I merge it?

Jun 27 at 7:12, by Ulrike Fischer

@HenriMenke if I say yes, @DavidCarlisle will blame me if it breaks tikz ;-(.

@Nasser Could you report it?

6:19 AM

@HenriMenke She seems to be very efficient. ;-) (BTW, since \pgfmathsmuggleis now "official", i.e. there is a version without @ in, do you think it is worthwhile explain how that can be used to smuggle stuff out of groups/loops? AFAIK some had used the above-mentioned issue to retrieve the length of lists.

Would pgfkeys work with \NewDocumentCommand?

6:37 AM

@HenriMenke could perhaps also look at this tex.stackexchange.com/a/472276/2388. In my opinion pgfplots should simply not use color names with spaces as it doesn't work on the dvips route. But the discussions seems to imply that Christian waits for a pgf fix instead.

@HenriMenke If you want you could write an answer to tex.stackexchange.com/q/499475/121799, your fix works great. Then we could close future questions that run into the gcd problem as a duplicate thereof.

I've been looking at tex.stackexchange.com/a/34318/3406.

6:54 AM

@marmot I have submitted PGF 3.1.4 to CTAN today where this and other things are fixed.

4

It's helpful, but the answer would have been better if it had used a completely user-defined function insead of \parbox.

\parbox has 3 optional arguments. However \newcommand has only one optional argument. It's not clear from this argument how to make things work with \newcommand.

Or could I use \NewDocumentCommand with pgfkeys?

@FaheemMitha ? why do you want lots of positional arguments ?

@FaheemMitha yes of course

@DavidCarlisle I don't want "lots". I want a few, yes.

@DavidCarlisle Ok.

7:12 AM

@FaheemMitha one of the benefits of keyval arguments is that you don't need lots of optional arguments with some order I always forget.

@UlrikeFischer That should have been fixed in 3.1.2 already.

@UlrikeFischer For me the MWE from tex.stackexchange.com/questions/470077/… typesets without problems in DVI mode.

@DavidCarlisle Sorry, I now realise that I don't know what you meant. Perhaps what I was saying was unclear.

Ok, to be clear, I want to just use optional arguments with key values.

Not positional arguments.

What I was saying is that (based on Ryan's example) it's not clear how to make key-value optional arguments work with \newcommand, since he's using an example (\parbox), which has 3 optional arguments. But perhaps I should just use \NewDocumentCommand.

@UlrikeFischer I don't want to use positional arguments. That's why I'm trying to use key-value.

7:27 AM

@HenriMenke perhaps you did something to avoid the need for the prologue option in tikz. But the option is still not usable together with pgfplots because of the space in the internal color name which means that you can't use local named colors in postscript.

Ah, now this is a clear example: tex.stackexchange.com/a/400429/3406

I see that the idea is to pass the argument to \pgfkeys and let it unpack things.

@FaheemMitha it absolutly doesn't matter which command he internally use to demonstrate the use. You could use a picture environment or a list or something else. \newcommand or \NewDocumentCommand doesn't make a difference here.

@UlrikeFischer Well, like I said, using \parbox, which has 3 optional arguments, obscured things, because the user would tend to reach for \newcommand, which has one optional argument. That's all I was saying.

Obscured things in the sense of making it harder to mimic directly. Which is what I tend to do, at least to start with.

It seems like the general approach would be to pass the arguments to \pgfkeys or similar inside the macro definition, and let it unpack values into the individual macros, and then use those macros. But that's not immediately clear from the documentation.

Andrew's example is quite clear, as is the following one. Does the LateX 3 approach offer advantages over this? And I think one would have to use one optional argument, whether using \newcommand or \NewDocumentCommand, because the multiple optional arguments in \NewDocumentComand are positional.

@FaheemMitha l3keys has basically two commands (like the other keyval systems with the exception of pgfkeys): one to define the key and one to set them. I prefer this, imho the code is clearer. With pgfkeys I'm never quite sure if I'm currently creating a new key or using and old one.

@FaheemMitha you don't need any optional argument at all. You can pass keys through the mandatory if you want. See e.g. \sisetup{....}.

7:43 AM

@FaheemMitha the usual reason to switched to a named argument syntax is to avoid having positional optional arguments, so normally you only have one.

@FaheemMitha I don't understand that comment at all, whether you use newcommand or NewDocumentCommand makes no difference the content of the argument isn't used at all it is just passed on so if the underlying command uses keyval or pgfkeys or whatever is not relevant at all.

@marmot move the bear a bit below the baseline so that his head doesn't bang against the previous line, this hurts ...

2

@UlrikeFischer Hmmh, or clip? ;-)

@marmot that hurts even more and he will probably bite the marmot here ...

8:35 AM

@UlrikeFischer I see. Thank you.

@UlrikeFischer That's \sissetup in the siunitx package?

@DavidCarlisle Only one optional argument?

@FaheemMitha yes that is more or less the idea of using keyval syntax, instead of \foo[t][3cm][c][5cm]{hello world} and getting all the options in the right order, you use \foo[valign=t,halign=c,width=3cm,height=5cm]{hello world} so there is no need for multiple optional arguments

@DavidCarlisle Well, in Ryan's example, he was passing the macros generated by pgfkeys to the arguments of \parbox. Which made it possible to use key-value arguments with \parbox, if I understand correctly.

But my use case is not customizing existing macros.

@FaheemMitha it makes no difference. In either case you have an underlying command, either one that you have written or an existing latex command, and then you specify some optional argument syntax over that.

@DavidCarlisle Yes, I understand. Also, if you want to pass a different subset of the arguments on different occasions, it all becomes a big hassle trying to keep track.

@FaheemMitha so normally the single optional argument that you can define with newcommand is all you need

8:45 AM

@DavidCarlisle It does make the difference that with an existing macro you need to adapt the way you pass in the arguments.

E.g. Ryan did:

\parbox[\myparboxPosition][\myparboxHeight]
    [\myparboxInnerPos]{\myparboxWidth}{#2}

This isn't how you would do things with a user defined macro. That's all I'm saying.

@FaheemMitha no it makes no difference, the command being defined in that answer was the new command \myparbox

But I guess the difference is cosmetic.

@FaheemMitha you might, why not?

@DavidCarlisle I guess you could.

@DavidCarlisle Yes, it was.

@FaheemMitha the point is the concerns are entirely separate, you use \newcommand or \NewDocumentCommand to gather the arguments then you pass those arguments to the underlying command using whatever syntax that needs, the underlying command may be using [] or just mandatory {} or delimited arguments defined by \def but the outer declaration of the user syntax doesn't need to know the implementation details of the inner command

8:51 AM

@DavidCarlisle Ok.

Python, for example, supports named arguments directly in def. Is there some reason that the functions/macros that create macros (like \newcommand etc.), can't do the same?

@FaheemMitha you could do: you could inline the definition using newcommand or NewDocumentCommand, its just bad style, it's better to define a basic command with simple mandatory arguments and then wrap it with a user syntax declaration with optional arguments using NewDocumentCommand which allows a separation of layers and possible alternative syntaxes.

@DavidCarlisle I realise it's more flexible to keep it separate.

But some languages do provide this feature out the box.

@FaheemMitha yes but tex doesn't and while it costs nothing to use a fancy top level syntax parser for a document command like \section if you do the same in a tight inner loop in some calculation you'll find that you are spending far more time parsing and re-parsing arguments than you are doing actual work

Sadly I am more or less describing my day job:-)

@DavidCarlisle Time as in CPU time, or human/programmer time?

@FaheemMitha cpu

9:01 AM

Are the rules for Python different because it typically doesn't call functions many many times? Or for some other reason?

@DavidCarlisle Oh

I guess Python doesn't necessarily have to operate in real time. But if people don't get TeX to compile right away, they get antsey.

@FaheemMitha as you say the named argument syntax is built in to python it's essentially "always there" so there is essentially no cost to doing foo(x=2) instead of foo(2) but in tex doing \foo{2} is possibly one expansion step and an xparse defined \foo{x=2} is possibly several thousand, if you just do it once in a top level document command it doesn't really matter but if you do the same at every step of nested code then...

@DavidCarlisle Always there, as in compiled in?

I have a hard time remembering that TeX is very much not a compiled language.

@FaheemMitha yes just part of the language, it's not as if you can say you will never use named arguments and so opt in to a faster parser that doesn't look for them

@DavidCarlisle Ok.

9:27 AM

user image

@UlrikeFischer Problem solved. ^^^^

@marmot you'd better hope that @barbarabeeton never sees that, her proofreader's appreciation of white space usage will be somewhat assaulted....

@DavidCarlisle So back to clip?

When Andrew wrote in tex.stackexchange.com/a/400429/3406:

> Another option is to store the values inside the keys rather than in macros. I prefer this approach because the keys exist anyway so you might as well use them. This also makes it easy to set defaults -- and by calling \pgfkeys inside an environment you can make any key changes "local" to that environment.

Did he have a point? It seems like a slightly more awkward approach, and doesn't use the existing mechanisms.

@FaheemMitha Well the pgfkeys model is very structured

user image

2

9:33 AM

And I'm not sure what he means by "local". Isn't it local anyway?

@FaheemMitha Yes, they are

@marmot ^^

he's talking about scope, right?

@JosephWright So is there any reason to do it that way?

The answer below, by gigabytes, seems more conventional.

user image

2

@DavidCarlisle @UlrikeFischer ^^^ perfect spacing.

@DavidCarlisle Don't show this to @UlrikeFischer

I really hope that I'm not barging into or interrupting an ongoing discussion but I have an opinionated question for anyone who's familiar with the TeX internals (i.e. I don't think it would belong on the site proper)

9:39 AM

@marmot Barbara will complain about the kerning.

@texdr.aft you mean important discussions on how to adjust the typesetting of words avoiding a cartoon bear's head? :-)

@UlrikeFischer Ja mei.

@DavidCarlisle and cricket

This is a pretty interesting question and answers -> tex.stackexchange.com/q/45183/3406

@DavidCarlisle that's what I thought was going on, but with no context I didn't want to accidentally disrupt something significant

9:43 AM

But I don't understand what Frank says about not being concerned about backwards compatibility. Sure's that an issue even for LaTeX3, since people are using it?

(not that this isn't significant)

@marmot I'm puzzled. A bear, not a marmot?

@FaheemMitha it's a play on words.

@PauloCereda Yes, I see that. I guess "marmot with me" would not make much sense.

Regardless, I thought a marmot would be preferred.

@FaheemMitha no one is using a latex3 document level syntax

@texdr.aft so the question was???

9:47 AM

@DavidCarlisle ah yes. sorry

@DavidCarlisle Oh, he's talking about the top level formatting commands? Ok.

If anyone were to re-implement TeX today (in 2019), how closely should they adhere to the original

@texdr.aft that depends

@DavidCarlisle wackiness level?

Pascal is easy enough to directly translate to a lot of languages but if it's not good code...

some of TeX's core is very closely connected to its implementation

I suppose that there are two overall possibilities: (1) approaching it starting with tex.web/TeX: the program or (2) doing a "clean-room" implementation starting with the TeXBook

9:51 AM

@texdr.aft LuaTeX took pdfTeX, ran web2c on it and continued from there with refactoring.

@HenriMenke is LuaTeX in some way related to metatex.org/cxtex? I seem to remember it is, but I'm not sure.

@HenriMenke regardless, that's definitely a good (successful!) starting point

@texdr.aft JSBox is probably closest to what you are after: it's not public code, but the details we have suggest it implements using ideas from both

@texdr.aft it's not really an answerable question. of things that are "visibly tex-like" pdftex and xetex use the same web sources for the tex bits, luatex uses some C that was originally translated from that but hand maintained as cweb jsbox is (unseen, private) hand written C, but then other things use tex algorithms, MSWOrd's math layout inspired by tex, openoffixe (and some javascropt librraies) use a re-omplemented tex hyphenation algorithm..

@texdr.aft , sile implements some tex inspired algorithms in lua... An important first question is to define what you mean by "re-implement tex" do you mean make a new tex-inspired typesetting engine, or do you mean make something that can typeset 30 years worth of legacy tex documents with essentially no change?

@DavidCarlisle, @JosephWright meanwhile Norbert just updated 70MB of code for the cockpit GUI thingy. :)

@DavidCarlisle I mean the latter. something that would pass the trip test at least

10:00 AM

@PauloCereda people shouldn't use java for lightweight guis:-)

@DavidCarlisle Agreed. :) But this time it's the Java community's fault. If they kept FX in the correct place the first time, this should not have happened. :)

@texdr.aft well then the options are very limited, as you note the pascal source is pretty simple (and translates to C that's pretty ugly but brutally efficient at what it does) so if you re-implement by hand there is not a lot you can gain and lots of opportunities to add unintended differences.

@DavidCarlisle after all, it's Knuth. :)

@DavidCarlisle that's what I figured.

@PauloCereda as long as you have someone to blame, and it isn't you, than it's OK if the software breaks.

2

10:02 AM

@PauloCereda as I was reading through the source, that fact became more and more apparent

@DavidCarlisle ooh

@texdr.aft Yes, that's Taco's original conversions of pdfTeX to C.

@PauloCereda the first tex distribution I installed fitted on three 700K floppies :(

@DavidCarlisle ooh

@PauloCereda "installed" wasn't really a thing since the machine had no hard drive, so all you had to do was insert the floppy into the machine and it was done.

10:07 AM

@DavidCarlisle emacs?

@PauloCereda well it wasn't my machine I had a sun3 with emacsview the sunview gui version of emacs, but in the math department the machines were.. less advanced...

@DavidCarlisle hmmmmm

does LuaTeX have a string pool (in the WEB sense)?

10:23 AM

I think I asked this before already (and I can't remember if anyone answered), but does LuaTeX pass the TRIP test?

@UlrikeFischer I think I see what you mean now. It uses \pgfkeys for everything. It's confusing.

@texdr.aft Yes, it's not changed that (changes from TeX are detailed in the manual)

@FaheemMitha pgfkeys is object oriented

@JosephWright for a broader view of OO. :)

@JosephWright But TeX isn't.

@FaheemMitha No, but one can create OO structures

Now reading the manual yet again. Sigh.

@JosephWright Hurray.

10:28 AM

@PauloCereda As in everything is done as a property, hence no separate 'set' and 'define' functions

@JosephWright sure thing. :)

@FaheemMitha Till wrote a whole OO module for TikZ: it's detailed in the manual

@JosephWright What is it called?

@JosephWright thanks

@JosephWright The problem is that in usage that distinction is hidden. Not ideal, imo.

I like things to be explicit.

10:30 AM

@FaheemMitha huh?!

So I suppose what action is taken depends on the handler.

I dunno if I'd call this object-oriented, exactly.

@PauloCereda Hmm?

To quote the manual:

> • pgfkeys supports handlers. These are call-backs that are called when a key is not known. They are very flexible, in fact even defining keys in different ways is handled by, well, handlers.

here's the beginning of "pretty-printed" TeX source from TANGLE. I started with the ptop utility and fixed it up by hand; all c++-style '//' comments are my additions and are used when the original WEB source used a macro that TANGLE expanded:

gist.github.com/texdraft/cb5d4360f16ac3d7d48f204ef286aa01

parts that aren't done yet (along with, well, the rest of the program) are omitted

I tried not to change anything

(I only did this to see what it would look like and not to try to get around WEB or anything)

It should be possible to modify TANGLE to get it to do something like this by itself, but I haven't got it working with fpc and gpc is too difficult to install on Linux

@DavidCarlisle poor bear, beheaded!

Der Baerenkopf (deutsch Bärenkopf) ist ein Berg in den südlichen Vogesen, acht Kilometer südöstlich des Ballon d'Alsace. Er beschließt das Tal des Flüsschens Madeleine, das in 780 m Höhe an seiner Flanke entspringt. Er ist bewaldet und mit einem gut ausgebauten Wanderwegnetz ausgestattet. Den Gipfel des Berges teilen sich die Gemeinden Dolleren in der Region Grand Est und Lamadeleine-Val-des-Anges in Bourgogne-Franche-Comté.…

10:51 AM

@JosephWright think I missed \immediateassignment before or forgotten about it

11:02 AM

@DavidCarlisle Huh?

@JosephWright luatex primitive for expandable \def

@DavidCarlisle That one must be new (ish)

\edef\zz{\immediateassignment\def\zzb{??}xx}

\show\zz
\show\zzb

\bye

@JosephWright it is in the tl2018 manual, but not 2017

11:18 AM

The pgfkeysintro section doesn't say so explicitly, but I understand correctly, \pgfkeys works in two modes. One where a handler is specified, in which case something action is taken based on the handler. The second is that a handler is not specified, but a key=value pair is, in which case the key is executed.

And multiple "statements" within pgfkeys are comma separated. Presumably it's possible to mix these two modes together.

I really hope I got this straight. Looking at more examples now.

2 hours later…

12:57 PM

(removed)

(removed) <-_I can do this too

ooh

1:15 PM

@UlrikeFischer @PauloCereda you are both fakes, and poor ones at that.

@DavidCarlisle oh

ooh

1:40 PM

So apparently pgfkeys uses key value pairs, but the values can be missing. Hmm.

And for some reason replaces missing arguments with nothing. That's unexpected.

Not very TeX-like at all.

@FaheemMitha Quite common: same in keyval

@FaheemMitha That approach varies between different keyval packages

I suppose the keys can't be missing, though.

2:13 PM

@FaheemMitha I did that in keyval for boolean values, you can go \includegraphics[clip]{..} meaning same as \includegraphics[clip=true]{..}

2:37 PM

Hi everybody!

Is there a way to set a boolean variable to keep track in which column the text is on a twocolumn document class option?

plz explain

If I have \documentclass[twocolum]{book} and a bunch of text on the document. But let's say the text ends in the first column of the page or the second column of the page. How do I tell LaTeX that I want to know in which column the text ended?

2:53 PM

@DavidCarlisle BTW, the manual is dated 2014, but I suppose keyval must be much older.

In the case I'm thinking of, the value isn't a boolean, it's simply missing.

@Levy LaTeX has \if@firstcolumn but I'm not sure if it does what I think it does

@texdr.aft I'll check it.

@Levy it's defined in file K in source2e.pdf if you want to see it in the source.

3:20 PM

How do I set the max height my footnote can take into the layout?

@Levy Maybe this helps? tex.stackexchange.com/q/214806/2693

@Levy But I'm not optimistic...

17 hours ago, by Alan Munn

@PhelypeOleinik :) I'm nothing if not pragmatic.

is there something like github's gists for LaTeX?

(where the output is shown, of course)

@texdr.aft Scratch projects on OverLeaf?

3:37 PM

@JosephWright alright. that's what I figured

@JosephWright ew

@FaheemMitha Karl touched some spelling over all the core documentation around that time, the code is from '93

@DavidCarlisle Oh. I thought it was from the 80s.

And that was the first ever key value package?

@Levy when do you want to know (it is basically like asking what page something is on and the answer there is to use \pageref, but for columns it is trickier)

@texdr.aft it does, but it will probably give an unexpected answer unless you expect the right thing.

@DavidCarlisle that makes sense.

3:46 PM

@FaheemMitha it was the first one called keyval, the syntax was used by epsfig and before that pstricks (but the keyval parser was more friendly eg allowing white space around , and =, epsfig was rewritten to use it and pstricks switched to equivalent code as well

@DavidCarlisle I see.

@FaheemMitha pstricks did a lot of stuff first:-)

pgfkeys is much more elaborate than I expected. (Now that I have some idea what it is doing.) Clearly the author thought he should consider as wide a range of possibilities as possible.

@FaheemMitha well it's the core of a package with a thousand page manual elaborate shouldn't be that unexpected

@DavidCarlisle I have been just wondering about this too. Do you think using labels in a \marginpar or \marginnote would be reliable? something like this:

\documentclass[twocolumn]{article}
\usepackage{lipsum}
\usepackage{marginnote}
\makeatletter
\newcommand\refcolnum[1]{\ifcsname r@firstcol#1\endcsname 1\else 2\fi}
\newcommand\collabel[1]{\marginnote[\label{firstcol#1}]{\label{seccol#1}}}
\begin{document}
   AAA\collabel{a}\lipsum*[1-4]BBB\collabel{b}

   label a is in column \refcolnum{a}, label b is in column \refcolnum{b}.
\end{document}

3:49 PM

@UlrikeFischer I'd use \pdfsavepos and see what half of the page it is in

@DavidCarlisle that's the other possibility. But it means that you have to calculate the page layout.

@DavidCarlisle The core? I thought the graphics were the core.

@FaheemMitha Once you start one 'some handlers' or similar, there are quite a lot of things you might do. l3keys is the same; the important thing it to have a clear interface for creating keys. That works well in pgf (and l3keys); contrast xkeyval (which I feel is not that friendly as a programmer)

@FaheemMitha syntactically tikz is built by setting lots of keys for things, the fact that the bottom pgf commands in the stack draw curves is true as well of course

@FaheemMitha That's the core of the user experience, but not the core of the code of the package.

3:53 PM

Thank you for the feedback, folks. :-)

@JosephWright Does l3keys have a similar design, then? I got sidetracked, and haven't read your articles, though I might get to it later. Ryan's answer is really quite helpful, though it could do with some fleshing out. IMO, it should be added to the PGF/TikZ manual, perhaps at the beginning of the pgfkeys section.

I'm having a really bad time trying to make the footenotes nice and tide in my project. Although, I have tried many things and none seemed to resolve the bad layout I get. Can somebody help me to get it a better look? Please, the files are hosted in here: drive.google.com/open?id=1Vp8xPvendDUIPdDPtnOJpT7wU2qbk7cs the file to compile is Holy Bible.tex with LuaLaTeX.

@Levy it's completely irrelevant, but I really like the way your code looks.

The text got all messed up

And because of the footnotes the counters for the verses got wrong

@FaheemMitha Similar in the sense of using .property to create effects, but different in the sense it retains separate define and set functions

@Levy are you wanting the footnotes to end on the page on which they begin?

4:08 PM

@texdr.aft I want the text to occupy the gaps produced by the footnotes

@Levy spaces in filenames :(

@Levy I know, but in addition do you want footnotes to always end on they page on which they start, or would you accept them ending on the next page

I would accept them to end in the other pages

I want to get rid of the gaps

@Levy I don't have the font so it's not going to look right for me anyway

@DavidCarlisle In the code I sent, I commented the part with the font I used, so if you compile it it should appear with cmu to you

4:13 PM

I got this I just assumed the weird spacing was due to font issues, but perhaps not?

user image

@JosephWright It's probably better to use different functions.

@DavidCarlisle With the font I use the gaps are the same. Let me show with an image

user image

@Levy lots of missing % at ends of lines

@Levy unrelated to the spacing but `\if@firstcolumn
\gdef\markleft{\relax}
\else
\gdef\markleft{{\scshape\currbook} \thebiblechapter:\thebibleverse}
\fi` can not affect the headings in the way that you want, you need to use the \mark system

@DavidCarlisle I think there's an issue with utf8andspace on Windows ... probably an OS think

@JosephWright with the space bits or utf8 bits?

4:23 PM

@DavidCarlisle I just tried lualatex-dev as I wanted to check how it handle whitespaces, and it fails for me for every graphic: `! error: (file "example-image".pdf) (pdf backend): cannot find image file '"example-image".pdf'

@DavidCarlisle UTF-8

@JosephWright do you mean what I just remarked?

@DavidCarlisle If I put and % at the end of the lines of Genesis.tex file, It doesn't create the paragraphs for each verse.

@UlrikeFischer I get a test failure with the file with Euro signs in the name (required/graphics). Clearly works on Unix.

@UlrikeFischer web2c file handling explicitly handles % but I wonder about luatex's direct access it was working for me but....

4:24 PM

\openin1 foo€€€.eps %
\ifeof1 %
\T
\else
\F
\fi

@JosephWright is the utf8andspace code already in develop or do I need something more?

@DavidCarlisle, @UlrikeFischer ^^ Try that with a suitable file available

@UlrikeFischer In dev

@Levy not the document, your code files, in your definintion of \verse for example, \refstepcounter{bibleverse} adds a space to the output as well as incrementing the counter, as does \gdef\markleft{\relax}

@JosephWright er but \openin syntax gives no possibility to make the utf8 safe does it? wouldn't you need a \detokenize there?

@DavidCarlisle That's with plain ...

@DavidCarlisle what has % to do with \includegraphics{example-image}?

4:28 PM

@JosephWright I'm confused

@UlrikeFischer sorry three parallel conversations, I meant " not %

@DavidCarlisle I got a test failure so cut down to the minimum: just the \openin test that is the core of \IfFileExists. It clearly works on Unix as Travis-CI passes the tests, Frank presumably is ok on the Mac, you are OK on Cygwin. But for me on std Windows ... I guess I should check with MiKTeX ...

@DavidCarlisle ;-). I was coming from tex.stackexchange.com/questions/499661/… where quotes don't work correctly either.

@DavidCarlisle, @UlrikeFischer It's a binary bug: we are OK on MiKTeX

with pdftex plain on cygwin I get \F (if the file is there)

@DavidCarlisle Yes, exactly, so it's an issue with the Windows binaries. I'll report on the TL list

4:32 PM

@JosephWright I think this has come up before, isn't there a compile time option to convert the filesystem reported string to utf or not.... you get failures whatever you do in some windows configurations....

@DavidCarlisle I'll raise it and see what I get

@JosephWright which output do you get \T or \F?

@UlrikeFischer \T with TeX Live on Windows, \F with MiKTeX (and implicitly on Linux, as the tests work)

@DavidCarlisle I have put the % at the end of the lines, and took off the \gdef\markleft{\relax} but the gaps still remain.

@JosephWright do you have command_line_encoding = utf8 in your texmf.cnf?

4:37 PM

@UlrikeFischer No idea; I have whatever the vanilla install does

@UlrikeFischer No, seems not, but then lots of users won't either

@JosephWright there has been some discussion on the texlive list to add it by default.

@UlrikeFischer I knew this had come up before:-)

@Levy I believe that the main issue is that each footnote is confined to a single column

@texdr.aft What could be done then?

@Levy what script is this, I can sort of read it but?

4:44 PM

@DavidCarlisle It's just very archaic English without modernizing the character set, if I understand what you're asking

@DavidCarlisle It's mainly from the JOHN WYCLIFFE translation of the bible to english from the Latin Vulgate

Many modern editions of texts from this time would replace ȝ with gh but I see no reason to

@texdr.aft Yes. I'm trying to rescue many etymological backgrounds from the old translations

@DavidCarlisle ;-). @JosephWright I get \F, and I now found also the texmf.cnf where I added the option. I think we should suggest to add it generally but it could have side effects tug.org/pipermail/tex-live/2018-May/041697.html

@Levy to actually answer your question: if you're okay with significantly changing the overall layout you could do what the ESV does and have the text in a single column and the footnotes in two columns. But that's again a huge change and it's totally find if you don't want that

4:49 PM

@UlrikeFischer Well if we don't, things are different between the different operating systems

@Levy no time now but it looks better with \raggedbottom, at least it moves the gaps to a better place

@texdr.aft I thought you were saying to change the footnotes to single column not the text. I think it wouldn't look nice the verses on single column

@Levy the only modernization that I think editions of old texts should do is (1) removing the "long s" and (2) changing the typeface from blackletter to the usual roman

@Levy I do agree

@DavidCarlisle Thank you, I will try it.

@texdr.aft Yes, the long s sometimes most difficult the reading than makes it beautiful. Not to metion that it's almost like the f.

@Levy what's your source for the text?

I know it's the Wycliffe bible, but what edition/website/etc.

4:57 PM

user image

alright cool

It's all the information I have about it

archive.org/details/holybiblecontain03wycluoft

@Levy one idea for fixing your problem: split the page into two vertically. Both halves have the same two-column layout, but the bottom is smaller and contains the footnotes.

this way you have the fixed size for them and they split across correctly

for everything having a fixed height this would work but I'm not sure how to generalize it to variable heights. I think that that would involve modifying the output routine in some way, but it seems possible.

1 hour later…

6:09 PM

Are there any good choices for an argument which can either be nothing or a piece of code?

@FaheemMitha What do you mean by "a piece of code"?

@AlanMunn Well, this macro -> \newcommand{\addblankpage}{\clearpage \phantom{} \clearpage}

I either want to call it, or do nothing.

@FaheemMitha Under what conditions?

@AlanMunn It's part of a macro. I'm wondering how best to pass that as an argument.

@FaheemMitha do you already have a working code?

6:19 PM

@PauloCereda Well, I have working code, but it doesn't resemble what I want.

Basically, the macro includes a PDF using \includepdf, and then adds a blank file after it, or not.

Nothing very exciting.

@FaheemMitha I'm still not sure I understand. When do you call it and when do you do nothing?

@AlanMunn I'd like to pass an argument which says when to call it, and when not to.

Ideally I'd like the argument to be either (a) nothing/empty or (b) the macro above, to create a blank page.

This is an attempt to clean up some old code which is becoming increasingly inconvenient to use.

so \macro{} does nothing and \macro{\addblankpage} does \addblankpage?

@texdr.aft Something like that.

But of course, I'm not sure of the correct syntax. I'm trying to use pgfkeys.

or maybe \macro (no arguments) does nothing and \macro{\whatever} does \whatever

6:31 PM

@marmot -- Hmmm. This keeps changing.

6:57 PM

@marmot Overleaf provides free accounts (with a limited feature set regarding collaboration and 3rd party services like github integration, and a shorter timeout). You don't have to pay anything to them for small tests (but you need a valid email address).

@marmot sorry, I was on the autobahn half of the day, so the answer came a bit late :)

« first day (3176 days earlier) ← previous day next day → last day (1749 days later) »

Transcript for

Jul '1912

$TeX - LaTeX$ TeX, LaTeX and Friends

\begin{chat}

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

$TeX - LaTeX$

site design / logo © 2024 Stack Exchange Inc; legal

mobile