Hi

I have a question.

I just recently reinstalled texlive and tlshell seems to be missing.

@manooooh -- use the "double hyphen" (en-dash) when two authors are involved. That way it's possible to distinguish between different authors when one of them has an already hyphenated name (e.g. Swinnerton-Dyer).

@manooooh -- By the way, how was the eclipse?

@barbarabeeton thanks for the help! So one hyphen is needed

@barbarabeeton meh, it was cloudy and at work I could not see it :/. I talked with coworkers about it

@manooooh -- No. You really want an en-dash, since Newton and Raphson were two different people.

@barbarabeeton oh, I thought you were referring in titles, bibliography etc., not in normal text

Sorry for the misunderstanding

@PhelypeOleinik, thanks.

@azetina, I am no expert at this but if you're on Windows I would check for \2019\bin\win32\tlshell.exe in texlive directory.

@PhelypeOleinik, by the way, I was wondering whether it's feasible to modify TeX output file to enclose two consecutive lines of a paragraph into a box but in such a way so that the box starts in the middle of the first line and ends in the middle of the second (instead of enclosing two whole lines). If can be done, this would be an alternative to soul package to highlight text.

@PhelypeOleinik, I meant to edit the ps or div file generated by TeX.

@bp2017 Its no longer there. I checked

It got deleted in the most recent update.

@bp2017 I also did a fresh install on another computer installing off the net and still ended with the same result

@UlrikeFischer (or someone else interested), just looking at you interesting solution at tex.stackexchange.com/a/5265/87678. It doesn't seem to work for me when I call it with lualatex. Am I missing something? I can produce the final PDFs with lualatex if I call with pdflatex or xelatex and change the \write18 commands to lualatex, just not run the whole thing with lualatex.

When calling with lualatex I get:

$ lualatex --shell-escape temp472
This is LuaTeX, Version 1.07.0 (TeX Live 2019/dev/Debian)
 system commands enabled.
(./temp472.tex
LaTeX2e <2018-12-01>

luaotfload | main : initialization completed in 0.135 seconds
(/usr/share/texlive/texmf-dist/tex/latex/base/article.cls
Document Class: article 2018/09/03 v1.4i Standard LaTeX document class
(/usr/share/texlive/texmf-dist/tex/latex/base/size10.clo))
lualatex --jobname="temp472" "\gdef \conditionmacro{1}\input temp472"
lualatex --jobname="temp4722" "\gdef \conditionmacro{2}\input temp472"

Oh. By accident, I discover that adding the ifplatform package makes it work…

@AlanMunn I get:

With 1, it brings up TeXWorks(?) with the sty file.

2 brings up the Foxit editor, with an error.

(I don't use it, but I have it installed. I was trying to use it to do PDF annotations.)

@AlanMunn Yes, I see his name is mentioned there, once. He's responsible for CTAN's name?

@DavidCarlisle texdoc latex2e.

@DavidCarlisle texdoc -l latex lists a bunch of stuff.

Or perhaps even better package pdftexcmds and \pdf@system.

Looks like someone forgot about this one - github.com/rf-latex/setspace/blob/master/setspace-test.tex

@manooooh Clearly a double hyphen.

@DavidPurton I can't test now but guess that you need the shellesc package - in luatex write18 doesn't work without it.

I just came across a bug report against the setspace package, posted as an answer. I'm not sure why the author thought this was a good place to put it -> tex.stackexchange.com/a/434724/3406

No idea whether it is valid, though.

@FaheemMitha no an answer is not a good place for a bug report. But the chat is not a good place either, so if you think it is important ask the OP to put it on the github or do it yourself.

@FaheemMitha which has "unofficial in its title"

@FaheemMitha which is just a file listing of anything in the doc tree that has the string latex in the file or directory name

@DavidCarlisle So I see.

@DavidCarlisle Yes, I know. But it's a starting point.

@UlrikeFischer It's not my bug report. Ok, I'll add a comment, but where should I tell the poster to post the bug report, then?

@FaheemMitha not really, and nothing that you would call a manual of latex, even the core latex release documentation isn't in there

@FaheemMitha forgot?

@DavidCarlisle If one knows nothing, how do you suggest finding the documentation? BTW, the first hit includes a bunch of links -> file:///usr/share/texlive/texmf-dist/doc/latex/latex-doc-ptr/latex-doc-ptr.pdf

Are those correct?

@DavidCarlisle Yes, it looks like someone starting writing some tests.

@FaheemMitha that is the entire test suite

@FaheemMitha documentation for what? a particular package or the whole tex system?

@DavidCarlisle Oh.

@FaheemMitha it's more than most packages have

@DavidCarlisle General LaTeX documentation.

@FaheemMitha you already know the github, you mentioned it a few line above. And that is not your bug report doesn't matter. You found it, so do something about it.

@UlrikeFischer I know a repository exists. I don't know whether it is a suitable place to put bug reports. And anyway, it's not my bug report.

@bp2017 \par then look at \prevgraf to tell you what line you got to then interrogate the parshape lengths (that you needed to have stored separately before the paragraph) see the shapepar package or several answers on this site

@FaheemMitha it would come to me and I'd close it:-)

If it's a suitable place, I'll add a comment to the poster telling him to post a bug there.

@FaheemMitha where would you get general documentation about everything written in C, or perl?

@DavidCarlisle So it that (a) yes, file a bug or (b) don't bother?

@FaheemMitha b

@DavidCarlisle Harbison and Steele, I suppose. For C. I don't use Perl, so I have no idea.

I had a copy of H&S. Probably still do. Probably horribly out of date, though.

@DavidCarlisle Ok. Thank you.

No, hang on, H&S is a reference manual. I guess a combo of that and the K&R book, at least to start with.

Though really, one would be better off not writing C at all, and using something else.

@FaheemMitha the answer in all the cases is that there is no such thing, the documentaion of the language can not document every program that happens top be written in that language. Actually the latex companion does a good job of a survey of the main packages (as they were 12 years ago) including setspace

@DavidCarlisle Oh, you wrote "everything written in C". Sorry, wasn't reading carefully.

@FaheemMitha pick whichever language you like, C was just an example. The same answer applies to a question about every program written in that language.

Clearly, the answer is, in no one place.

@FaheemMitha so setspace is written in tex that doesn't make it "part of latex2e" and doesn't mean that it can be expected to be mentioned in any core documentation (although as it happens it is mentioned with a large 2 page section)

That top link references mirror.iopb.res.in/tex-archive/macros/latex/doc/usrguide.pdf as the official LaTeX documentation.

Is that still correct?

@FaheemMitha it's a mini user guide on how to get started written from the perspective of someone familiar with latex2.09 and switching to 2e, it is part of the core release, but...

@FaheemMitha make any latex error and the system tells you what the official documentation is, eg:

! LaTeX Error: Missing \begin{document}.

See the LaTeX manual or LaTeX Companion for explanation.

@FaheemMitha the LaTeX manual and the LaTeX Companion constitute the official latex reference, note that errors from tex or errors from latex package code do not have that line

@DavidCarlisle I'm not sure what those documents are. Are they listed in latex-project.org/help/documentation/#general-documentation?

@FaheemMitha they are books you can buy

@DavidCarlisle Oh.

@FaheemMitha just like tex where the reference is the TeXBook, from the same publisher

@DavidCarlisle Ok.

@UlrikeFischer Thanks! I guess that package is a better option again than the pdftexcmds package. I'll take a look.

@DavidPurton I'm sure @DavidCarlisle agrees with me that this is an excellent package and the right choice here. I will edit the answer to add the code.

@UlrikeFischer @DavidPurton yes clearly a brilliant package from a very reputable source.

howdy

@DavidCarlisle What's the right engine agnostic way to test if LaTeX has been called with --shell-escape?

Oh! Maybe \c_sys_shell_escape_int?

@Moriambar Hi! How are you?

hey @CarLaTeX I'm good. And you?

Or \sys_if_shell_unrestricted:TF…

@Moriambar Very well, I'm on holidays at the moment!

@Moriambar Garda Lake :) ^^^

@DavidPurton I thought it was one dash - i.e. -shell-escape.

@DavidPurton oh I was going to add an integer for that to shellesc package, seems like I didn't do it: the l3 integer would work as well

@FaheemMitha one or two makes no difference

@FaheemMitha it was one, but the linux/posix guidelines say two are better for multi-letter arguments and --help should work in all commandline programs taking arguments, so all the web2c related programs just accept one or two dashes for all arguments

@CarLaTeX wow!

unfortunately I'm at the office

@CarLaTeX overcast and cloudy, naturally we have blue skies and sun here!

@DavidCarlisle Ok.

Happy 4th of July!

@PauloCereda not widely celebrated here

@DavidCarlisle nor here. :)

@Moriambar CIAO!

@PauloCereda why should one celebrate the 4th of July?

@Skillmon sorry, I meant for our American friends...

@JosephWright duplicate user login: user192223 and 192222 seems to be the same person: tex.stackexchange.com/questions/498577/…

@DavidCarlisle Yesterday evening there was a storm

@DavidCarlisle Today it's sunny :) ^^^

Happy 4th of July to all the US friends!

@CarLaTeX You're on the Lombardy side, I guess from the picture.

Hi, I think I remember there was an environment that is almost like verbatim but not as harsh with catcodes that can be used in arguments? I cannot remember the name of it

@egreg Sirmione, as usual!

@Bordaigorl alltt from the package of the same name (but unusual to use environments of any sort in the argument of another command) \texttt{....} would be more common

@PauloCereda -- You really should be more specific. After all you're (South) American. And Canada, also in North America, has no special interest in July 4th. Never mind quite a few other countries in the isthmus linking the two main continents. But thanks for the well meant thoughts (although I surely wish it hadn't been hijacked. As expressed by the District of Columbia's representative to Congress, "[The Mall] is the place where we fly kites.")

@barbarabeeton I didn't get that last reference. I take it that's a reference to the National Mall?

But I don't recognize the quote.

@UlrikeFischer tagpdf update has landed

@JosephWright did you get it only today? I already had it yesterday.

@FaheemMitha -- Yes, it is the National Mall. The place where such celebrations as the national folklife festival are held, the front lawn of the Smithsonian and several art museums, and yes, a place where i've joined in kite flying. The quote is from an interview that was aired on the radio this morning. I agree with her, very strongly.

@barbarabeeton Ah, a radio interview.

@UlrikeFischer Just updated my work system; might have been there yesterday

@UlrikeFischer, @DavidCarlisle Karl has sorted l3 dependencies: l3kernel requires l3backend, l3packages and l3experimental require l3kernel

@JosephWright yes I saw the commits (waiting for latex-dev ;-)).

@UlrikeFischer I just updated and checked texlive.tlpdb :)

@bp2017 It should be possible with soul itself. The underlining is done with a macro similar to \fbox but which draws only the bottom rule. Another approach would be to use LuaTeX to insert the rules after TeX has broken the lines into a paragraph.

@bp2017 This whole thing is tricky because at the time TeX builds a paragraph all macros have already been expanded, so you have no control at that point. Workarounds like soul are possible, but they have their limitations. (You have already asked about this before, haven't you?)

@bp2017 And yes, it's possible to add raw contents to the output file. That's what the so-called "drivers" do. They are format (file format, not TeX format) specific instructions which modify the output file directly.

@marmot Do you really think I don't know that \let\frac\undefined will lead to the stated error message? Please, leave silly comments for someone else.

@egreg Sorry, it was only a response to "the only reason I can think of". Nothing else. And the evidence seems to suggest that plain TeX is not the reason. There was nothing hostile nor silly in the comment, only the point that there can be several other reasons.

@marmot It looked hostile.

@marmot The best is removing all those comments. Agree?

@egreg OK, sorry, it was not meant to be. It was really just the statement that there can be other reasons. And yes, I will be happy to remove the comments.

@marmot Anyway, I voted for unclear. The question is bad and I also downvoted it.

@egreg I agree again. (Which is why I told the OP that they should add a complete code...)

@barbarabeeton oh

@DavidCarlisle @egreg anyone remember which kernel patch level the counterwithin was fixed?

The initial release of the kernel with counerwithin had a missing . so instead of 1.1 we got 11.

@daleif 2018-03-08 ltcounts.dtx v1.1k

I'm trying to see if a user has a patch level on their TL18 that is too old

@egreg can we see it from the patch level on the latex format?

THe OP has posted their log file

@PauloCereda Shame on you @barbarabeeton :) The 'American' canard is just that. Words are defined by their usage, and there is no evidence in English that 'American' means anything other than USA-ian. The fact that we have terms like North/Central/South/Latin American isn't evidence that 'American' by itself is the superordinate term. No Canadian would say they are American even though they're happy to say they're North American. The same goes for South/Central/Latin Americans, if speaking English

@barbarabeeton </rant> :)

#########################
# 2018-04-01 PL 3 Release
#########################

2018-04-22  Frank Mittelbach  <Frank.Mittelbach@latex-project.org>

	* ltcounts.dtx: added missing '.' in default for \the<counter>
	when using \counterwithin (issue/38)

2018-04-10  Frank Mittelbach  <Frank.Mittelbach@latex-project.org>

	* docstrip.dtx (subsection{Batchfile commands}): Read 8bit input
	raw to leave high bits in the code to unchanged without utf8
	handling (issue 34)
	Without this change a dtx file with 8bit chars in the code need to

@daleif ^^

The OP is using LaTeX2e <2018-04-01> patch level 2 thus too old

@barbarabeeton Of course, I guess @PauloCereda has a soft spot for canards. :)

@AlanMunn ooh

@AlanMunn OOH

@AlanMunn Well, no marmot would say that they are cute but each of them will agree on super cute. Where is the difference? ;-)

@AlanMunn -- Rant accepted. But, to be fair, the continents should probably be called something else rather than be named after an Italian. I'm glad that I grew up in an area where many of the physical/geographical features kept the names that the original residents used, and were not renamed by Europeans. Similarly where I live now.

@barbarabeeton "land of pizza hawaii"?

@UlrikeFischer -- Not the most desirable pizzas. Although there are some "chain" pizza parlors in this area, the best and most respected still adhere to Italian tradition.

@barbarabeeton True.

@DavidCarlisle iPad Mittelbach? :)

@PauloCereda ?

@DavidCarlisle Frank's email. :)

@PauloCereda it's the latex project one

@barbarabeeton proper Italian pineapples, good.

@DavidCarlisle -- I think they're called "artichokes".

@barbarabeeton wikipedia tells me that nutella is Italian too.....

@DavidCarlisle They also say "In April 2012, Ferrero agreed to pay a $3 million settlement (up to $4 per jar for up to five jars per customer)." for not warning them that this is not to be put on a pizza.

@marmot Following in the glorious tradition of Mars and Guiness

@JosephWright People put Guinness on a pizza? Or Mars? (I only know that the Scottish fry Mars, but putting it on a pizza is a new level. Pffffffffffff.)

@marmot I meant spurious health claims ('A Mars a day helps you work, rest and play', 'Guinness is good for you')

@JosephWright Yes, these spurious health claims are overabundant. (I happened to know a Burger king manager, who also was a runner, and of course would never eat BK himself because it is really bad for the health and performance...)

@JosephWright Is there a way to make l3build ignore Windows/Linux line ending differences?

@PhelypeOleinik It should do anyway ....

@PhelypeOleinik For example, I'm on Windows but Travis-CI is Linux ...

@PhelypeOleinik where do you have a problem with line endings?

@JosephWright do we still have access to svn:-)

@DavidCarlisle Yes, at present

@JosephWright I must be doing something stupid then... That l3debug code I sent you a couple of days ago is failing now that I'm using Windows (I was on Linux back then).

@UlrikeFischer The m3tl-build001 test of the kernel

Comparando os arquivos m3tl-build001.pdftex.log e M3TL-BUILD001.TLG
***** m3tl-build001.pdftex.log
   19:  39:931 140:041 141:141 142:241 143:341 144:441 145:541 146:641 147:741 148:841 149:941 150:051 151:151 152:251 153:351 154:451
   20:
   21:  ============================================================
***** M3TL-BUILD001.TLG
   19:  39:931 140:041 141:141 142:241 143:341 144:441 145:541 146:641 147:741 148:841 149:941 150:051 151:151 152:251 153:351 154:451
   20:  ============================================================

@JosephWright hmm the one I have seems to have problems connecting, you couldn't do an svn blame array..dtx could you...

@UlrikeFischer @JosephWright ^^

@PhelypeOleinik oh, line breaks. That's normally not due to line endings but different path lengths.

@PhelypeOleinik welcome to our world:-) (@UlrikeFischer)

@DavidCarlisle I just wanted to ping you ;-)

@PhelypeOleinik which branch is it?

@DavidCarlisle Need the old non-public SVN?

@DavidCarlisle Checking out to my user area on comedy ..

@JosephWright trying to see when \noexpand got added to {\@tfor\next:=<>clrmbp@!|\do{\if\noexpand\next\NC@char ..

@UlrikeFischer The thing is that I compared a hexdump of both and one has 0d 0a (the generated log) and the other only 0a...

@DavidCarlisle Sure, I'm looking

@DavidCarlisle Yay :D

@UlrikeFischer The debug branch

@PhelypeOleinik Did you perhaps copy code directly, rather than via Git?

@JosephWright I'm getting svn: E170013: Unable to connect to a repository at URL 'svn+ssh://carlisle@comedy.dante.de but I have the same ssh config and public keys as before

@JosephWright I clicked the "download zip" button and unpacked it... Known problem?

@PhelypeOleinik the test m3tl-build001 passes for me and I'm on windows.

@DavidCarlisle You can't get it remotely: have to log into comedy and checkout in your user area on the system

@JosephWright ooh just like the good old days with rcs :-)

@DavidCarlisle It's revision 634, Rainer, 1994

------------------------------------------------------------------------
r634 | schoepf | 1994-03-14 23:27:00 +0100 (Mon, 14 Mar 1994) | 5 lines

        Removed idenfication typeout's.
        Removed the check for the old version of the @tfor macro.
        Moved the driver code in front so that the documentation can be
        processed by simply running the file through LaTeX2e.

@@ -2094,7 +2093,7 @@
 % the list "\NC@list".
 %    \begin{macrocode}
   \@ifundefined{NC@find@\NC@char}%
-    {\@tfor\next:=<>clrmbp@!|\do{\if\next\NC@char
+    {\@tfor\next:=<>clrmbp@!|\do{\if\noexpand\next\NC@char
                     \@warning{Redefining primitive column \NC@char}\fi}%
      \NC@list\expandafter{\the\NC@list\NC@do#1}}%
     {\@warning{Column \NC@char\space is already defined}}%
@@ -2436,39 +2435,6 @@
 % \end{macro}
 % \end{macro}

@DavidCarlisle Perhaps we should give Rainer a heads-up that the SVN is still useful ...

@DavidCarlisle Or I could do a conversion to Git, push to GitHub as a frozen repo

@JosephWright thanks

@JosephWright not sure that it all should be public (although I wouldn't mind)

@DavidCarlisle Well yes, could do the same but keep private, but path of least resistance is just to make sure it doesn't get lost at the moment

@UlrikeFischer I must be doing something silly then... I'll check. Thanks for looking into it :-)

@DavidCarlisle Key thing is the change is not my fault and not your fault

@DavidCarlisle I wonder if Rainer has a GitHub account ...

@JosephWright yes but at that point it was running on rcs via a mail interface concocted by Rainer so all checkins were sent by mail and his mail forwarder picked them up and checked them into an rcs instance on comedy so all checkins would have his name I think.

@DavidCarlisle Ah, I see

@DavidCarlisle Notably that the message doesn't really say anything about this change

@JosephWright I have a horrible feeling who may have written that message....

@PhelypeOleinik not necessarly. David and me have some experiences with test failures due to system differences ;-) But line endings are possible too: I had to change my git settings to avoid that some line endings are changed at checkout.

@DavidCarlisle Do tell :)

@UlrikeFischer The funny thing is that this one single test failed. It could be related to the fact that I didn't use git to clone the repo, as Joseph said. I should probably learn how to use it :-)

@PhelypeOleinik The Git set up is easier than the older SVN one (at least, for the team)

@PauloCereda would have loved the original version control: all front ended by emacs, emacs would do an rcs diff, check it in send a mail to Mainz with the diff where the diff would be patched to the Mainz checkout and checked in there, then send an email to everyone else with the diff, which was then picked up by the emacs mail reader and applied to the local rcs instance we each had, keeping the emailed diffs in sync with local changes was interesting....

@DavidCarlisle ah thanks! I ended up wrapping the thing into a box and then using the box. It's not for the final typesetting, it needed to be passed to one of my macros for internal comments