@barbarabeeton sure and of course taking apart and rebuilding line by line is more or less what the line numbering macros must be doing, but...

@barbarabeeton being American can you explain what's happening? There is a field with lines drawn on it and some big blokes running around trying to play rugby, and then every couple of minutes they stop and cut to the studio where some commentators try to explain what just happened?

@DavidCarlisle -- oh. that's probably the "superbowl". big thing today -- the final playoff of the u.s. football season. (a "collision sport"; rugby is a bit more subtle.) avidly watched by the fans. often notable for the ads. and, of course, the commentary. stretch it out as long as possible, to get in more ads. (i went to see a video of the bolshoi "swan lake" today. seemed much more edifying than football, even if the swan did die.)

@barbarabeeton Did it? Oh, not again!

@egreg -- what do you mean "again"? the superbowl happens only once a year. oh -- you mean that the swan died. yes. true. very sad. but the dancing was superb.

@barbarabeeton That's like poor Mimì: they make her die every time. ;-)

@barbarabeeton I'm watching on the bbc so no ads but that just leaves time for extra explanations for the audience most of whom haven't the faintest idea what's happening.

@egreg -- true indeed.

@egreg at least she stays human

@DavidCarlisle -- well, in the u.s., one gets indoctrinated to football starting with the high school games -- the big game in that league is always on thanksgiving. but soccer (sorry for having to use that name; it's too confusing for us 'mercans to think of two very different field sports called the same thing) is catching on, and may actually win out in the end. but it'll be a while.

@barbarabeeton Meanwhile I received the list of students for the next exams: 93 tomorrow and a whopping 250 for Tuesday. I have to grade them in a week.

@egreg -- oh, dear! guess you won't be getting much sleep until it's over. maybe you should adopt the system allegedly used by some graders -- go to the top of a staircase, drop all the papers from the top, then assign grades according to how far down the stairs they landed. (doesn't really provide any useful feedback to the students though.)

@barbarabeeton I'll find a way. At least there are no lectures, this month.

@egreg Maybe some of them won't turn up?

@barbarabeeton I tried that once. I don't mean I actually was going to assign grades on that basis, but I was curious if it would actually produce any kind of 'spread' at all. Maybe you have to practise, but my students would either all have done very well or all done very poorly. (I'm not sure whether the top of the staircase is meant to be high or low.)

What is the best thing to do if you find out that your package conflicts with another package people want to use as well? The conflict is just a macro name i.e. there is no fundamental conflict. But if I change the name now, then I'll break code which currently works. Whereas if I don't change it, people can't use it along with the other package. (Why does bussproofs have a prooftree environment rather than a bussproof one, as one might expect?)

@cfr -- i'm tackling that problem right now. i have to find a way to make bbold work with the \mathbb defined in amsfonts. i've now spent over 10 hours on the problem (much of which was devoted to figuring out what the problem was, since what it first looked like was a hard loop, which threw me off the track), and no closer to a solution than i was at the beginning, although now i know the cause of the problem. i think there needs to be more communication between developers.

@barbarabeeton That might be tricky, though, given the sheer number of people, including many inactive developers. I've even played with bussproofs but didn't remember it used prooftree. Not sure what I'd use instead, either. The suggestion is to rename prooftrees tableaux but that is a non-starter.

@barbarabeeton I don't think there is a technical solution to my case. Two packages each define a prooftree environment. The only way I can see to practically work around it is to rename the environment. Are there any other possibilities?

@cfr -- how often (if you know) is a single author going to try to use both packages at the same time? the situation is certainly unfortunate. i suspect the only sensible thing to do (since it's impractical if not impossible to change names now) is to analyze the conflicts and document them in an annex to the main documentation. (if you want to write it up for tugboat, that's fine; i'm writing a "case history" of my current problem, with, i hope, a workable solution.)

@barbarabeeton Unfortunately, I suspect we're going to end up with a bunch of copies called tableaux.sty, which conflicts with tableaux.styl. logicmatters.net/latex-for-logicians/trees

@barbarabeeton I have pointed this out and am currently suggesting

\usepackage{prooftrees}

\let\treeproof\prooftree
\undef\prooftree
\undef\endprooftree
\usepackage{bussproofs}

I think this is better than ending up with a bunch of versions called tableaux.sty.

But the suggestion is I rename the package thus, which, as I say, is a non-starter. It is a straight package-name conflict.

@barbarabeeton But I may easily be wrong ...

@egreg yay!

@CarLaTeX ooh :)

@PauloCereda Hallo Paulo! Yesterday egreg-CarLaTeX derby: Juve-Inter :)

Experts, I am stumbed again. I get error "some package has redefined the meaning of the math-mode dollar sign" which seems to come at random location. If I remove some random Latex code, the error goes away. So I can't make MWE. But I have one standalone latex file which shows the error. Can I post a question on this and put link to my Latex file? I am using lualatex. Here is screen shot

It is giving an error about line inside verbatim environment !

if I remove Tikz package. Error goes away. If I remove geometry package, error goes away. If I remove some random Latex code from inside the file, the error goes away.

I just removed a blank line, and the error went away! It seems to be size related. At exactly 150 lines, it gives the error! or so it seems.

@Nasser Foggy crystal ball.

@Nasser you have any tikz in your page head? although really you should make a proper MWE and ask on site

@Nasser the microtype warnings are probably because it doesn't know TU encoding can be used without fontspec, add \usepackage{fontspec} should make them go away)

@DavidCarlisle I put MWE. It is 115 lines. This is the smallest I could make it.

@Nasser This is smaller:

\documentclass[12pt,letterpaper]{article}%
\usepackage[margin=1.1in]{geometry}
\usepackage{tikz,everypage}
\AtBeginDocument{%
  \AddEverypageHook{%
    \small
    \begin{tikzpicture}[remember picture,overlay]
      \path (current page.north west) --  (current page.south west)
            \foreach \i in {1,...,\fakelinenos}
               { node [pos={(\i-.5)/\fakelinenos},
                  xshift=\fakelinenoshift, line number style] {\i} }  ;
    \end{tikzpicture}%
  }%
}

\begin{document}
\vspace*{20cm}

@DavidCarlisle wow! You are amazing. Can I use you version then instead of mine in the post? if that is ok.

@Nasser Why do you ping David?

@TorbjørnT. sorry, wrong ping :) .

@TorbjørnT. can I use your version instead of the one I pasted in my post?

@Nasser And the content of the verbatim environment doesn't have to be that long either, you just need four-five lines, they can be empty.

@Nasser Of course.

@TorbjørnT. Ok, thanks. Will use your MWE since it smaller.

@Nasser And I just took the hint @DavidCarlisle gave you in his comment to your question, then removed some more stuff that didn't remove the error.

@Nasser Just make the picture and box it; then add it in the background. There's no point in remaking it every time.

@Nasser I seem to remember you have already had problems with page breaks in the middle of verbatim.

@egreg thanks. But this would be too advanced for me to do. I am using code from answer which adds line numbers to side of each page. as I mentioned, I did not write the tikz code, I would not know how to.

@Nasser I guess that the workarounds in TikZ for babel shorthands (which use \scantokens) revolt against it.

@Nasser Package background

@egreg thanks,. Will look at background package. But most likely it will be hard for me to do it. I know little tikz as well. But will look at it.

@egreg haven't looked at the example yet but it looks like guess was right. Arguably though latex should try harder to normalise things in page headers. Oh rather in this case the everypage hook thing ought to do the normalisation.

@Nasser Well, I didn't really look at it, I just removed it, recompiled, and saw that the same error still showed up. @egreg's comment made me take a second look, to see where \fakelinenos was defined.

@TorbjørnT. I'm sure I deserved the ping anyway:-)

@TorbjørnT. I commented away all the tikz code which makes line numbers on the side for now. So I can continue working. So I am ok for now. Need to look at background package later. Latex is way tooo complex :)

@Nasser why do you need to change to background?

@DavidCarlisle I do not know. But above, I was told to look at tikz background package as work around to making line numbers instead of what the current code does? I have no idea now. THis is all over my head.

Well there is someone who has slept very badly: tex.stackexchange.com/questions/352276/…

@Nasser latex really is not that complicated, your documents are complicated, but that's a different issue. In any language if you copy arbitrary chunks of code from different people and expect them to work together without knowing how each chunk works you are likely to get clashes with incompatible code. I don't see that is a latex issue.

@DavidCarlisle Thanks! you code works very well and the error went away! But I have been trying to see how to shift the column of the line to the left edge of the page, as it was before and I can't figure it. Before, the line number were a little more pushed to very edge of the page. Now they are close to the text edge. Do you know how to push them back little to left? Is it in this line:

\AtBeginDocument{\AddEverypageHook{%
  \llap{\smash{\raisebox{\dimexpr-\textheight-73pt}{\usebox{\mybox}}}}}}

@Nasser use the shift parameter in your tikz or just add \hspace{...} one or other side of the \usebox

@DavidCarlisle if you mean the line "line number shift=5mm," inside the tikzset{}, then it does not seem to have an effect. I change it and nothing happens. Will look at \hspace now. Will let you know once I figure it out :)

^^ I dunno how can I explain this more. I really don't want to downvote.

@yo' Voting is strange in many cases. Stupid questions are sometimes upvoted to hell... and sometimes answers hard to find get one crammy upvote, if at all.

@yo' s/crammy/crummy ;-)

tikz-uml seems to be rather meh quality.

@UlrikeFischer -- let him create his monster, then produce a minimum example that shows it doesn't work reliably. i suspect that wouldn't be overly difficult.

@barbarabeeton Oh I don't have any problems to let him try ;-) But the outburst was interesting.

@UlrikeFischer -- well, he's just actually described himself. i think most other people here can recognize that.

@UlrikeFischer Not the first time this user has rude reactions. Besides, judging from previous questions, his/her knowledge of TeX is rather primitive.

@egreg one step forward two back;-)

@DavidCarlisle Two steps back: both ¿ and ¡ are missing. :)

@egreg more than that '' has gone as well

@DavidCarlisle Nice done!

@egreg github.com/phi-gamma/luaotfload/commit/…

@DavidCarlisle I see Chris is on form

The train is 30 minutes late. But I got my second glass of wine. ;-)

@HeikoOberdiek: We have recently had a discussion on the forum of the Italian TUG, that uncovered a (known?) bug in he hyperref package, actually nothing more than a typo in a regex specification.

@egreg Be careful, knowing the FFSS, you may risk of becoming an alcoholic!

See this post for further details: it has been purposedly written in English.

@GustavoMezzetti that ping won't reach Heiko (unless you have mod powers to ping anyone) but if you post an issue to github.com/ho-tex/hyperref/issues I'll see what I can do.

@DavidCarlisle Oh, many thanks!

I thought H.O. reads this forum from time to time…

@GustavoMezzetti Yes, but it's best to log bugs in issue trackers

Actually, mine was a way to ask where it should have been reported… because I’m to lazy to find it out by myself! ;-)

@GustavoMezzetti he does, but as a general way the site works if a name does not pop up if you type @ and the first letter the ping will not notify the user. you have to have been in chat within a day or so to be notified that way

@DavidCarlisle I see, I’m still learning how this chat works. Many thanks again

@GustavoMezzetti But he's been taking a slight break from package maintenance so he allowed the hyperref and oberdiek bundles to be maintained from github by a group of people, me mostly, until he has time to pick them up again.

@DavidCarlisle OK, now I’ll post a message where you said me to.

@GustavoMezzetti thanks then it's logged even if I don't get time to look at the issue.

@DavidCarlisle: Mmmh: it seems I’m required to create an account. I don’t like to needlessly create accounts by my name across the Internet, and give away my email addresses… :-(

@GustavoMezzetti gmLatexIssue@gmail.com and a fake name will do if you want to submit something anonymously.

@GustavoMezzetti you can also post to ho-tex at tug.org (but that's a public list so you may not want to do that) I'm going to eat in a few minutes but I'll look at your links later If it's clear enough I'll raise the issue myself

@GustavoMezzetti But bug reporters get a mention every once in a while ;-)

@Johannes_B OK, thanks, I’ll try that.

No, I’m giving up for now: I’m having dinner in a few minutes too. I’m going to postpone this to tomorrow. Many thanks again to everybody.

@GustavoMezzetti Dinner? I am hungry? What do you have? Can i come over? ;-)

@Johannes_B Schnorrer! :-P

@ChristianHupfer Exactly. What are you having for dinner tonight? Can i come over?

;-)

@Johannes_B The dinner will be cold until you would arrive here :D

@ChristianHupfer Nichs Kochen is de Kunst, sonderns warmhalten :-p

@Johannes_B Ok, ich lasse die Nudeln dann mal ein paar Stunden im warmen Wasser ;-)

@ChristianHupfer Oh. That will .. that is ... pasta ... No, please don't. Eat it. Please.

@egreg l3sort moving into the kernel this evening :)

@GustavoMezzetti github.com/ho-tex/hyperref/issues/28

@egreg you'll be inpressed by my Italian translation I'm sure ^^^

@egreg ossa pigre

@DavidCarlisle Not an Italian idiom

@egreg perhaps it'll gain popularity:-)

@DavidCarlisle You were right: it wasn't too hard to get an MWE that showed the problem. tex.stackexchange.com/q/352446/29387

@DavidCarlisle I suppose I can also remove polyglossia too. I'll try both and check if the problem still appears.

Actually, no: without polyglossia the problem goes away.

Hrm.

@DavidCarlisle The issue seems to be the blank line.

I feel pretty stupid now. I suppose I should get used to that.

@JosephWright Nice! Any change?

@Telemachus whether or not you are supposed to have a blank line before \pend I think reledmac could be more forgiving here, although I've not looked at the code in detail. The developer is usually around here so may comment.

@DavidCarlisle I appreciate that, but I think the offending code may be polyglossia. Also, it may be my fault.

That is, what I've discovered is that several little paragraph-related things can be fixed if I place the begin{language} and end{language} commands more carefully.

I'm testing it all out now, but I'll post an answer to my own question when I'm more sure.

To put that another way around, I didn't think that begin{latin} would do anything to paragraphing, even if it was on a separate line. I appear to have been wrong.

I sometimes think of commands as "code" and assume that whitespace is irrelevant to them. That's clearly wrong too.

@Telemachus I find it's always best to blame someone else

@DavidCarlisle Well, fast as thunder! (son come ll fulmine, sono il factotum della città… ;-) I apologize for not hvaing done it myself, but I use to weight carefully subscription to Internet services, I want to evaluate the site’s privacy policy and so on, and I simply hadn’t the time to do all that this evening. Thanks once more.

After all, I knew you are pretty fluent in Italian…

@GustavoMezzetti I was going to post a pointer to github in that guit thread but you have to create an account to post..... :-)

@GustavoMezzetti yes see above: chat.stackexchange.com/transcript/message/35262026#35262026

@GustavoMezzetti as for the suggested regex % add "+" after the space? I don't think it should really be needed as it's quite hard to accidentally get two consecutive spaces in tex

@GustavoMezzetti also in a couple of places it suggets (()|a|b|c) which I think is equivalent to (a|b|c)? isn't it?

@DavidCarlisle I wouldn’t say so: (()|a|b|c) also matches an empty string, but (a|b|c) does not.

@DavidCarlisle Yes, indeed (you have to create an account to post)! But this time it’s was me who posted a pointer…

@GustavoMezzetti with a ? as in : (a|b|c)? (a regex 0-or-1 not a question)

@DavidCarlisle Yes, that would work too (and yes, I had understood that it was not a question!). I’m not claiming that the pattern I devised is the unique solution, or even the most efficient one: I just found it natural to code things that way.

@GustavoMezzetti ooh I'm a "proper authority" that sounds important:-)

And about changing a ` ` into ` +`: I agree that it is not necessary, and indeed I didn’t code it, but wanted to leave a reminder at every spot where the modification could apply.

@DavidCarlisle Ah, you are referring to a chi di dovere!

@GustavoMezzetti apparently:-)

@GustavoMezzetti to be honest I'm less than convinced about all this regex checking anyway. most latex arguments are not checked, if you get an array preamble wrong for example the document just falls over, the regex adds some complexity and I'm not sure it's worth it, but while I'm just "standing in" for Heiko I only want to make minimal fixes so I guess I will fix it up more or less as suggested rather than taking it out.

@DavidCarlisle: Yes, I understand perfectly. Nonetheless, I find t a bit inconsistent to check pdfremotestartview but not pdfstartview.

@GustavoMezzetti sadly if I give just the phrase "chi di dovere" it comes back as "those responsible" which has connotations of blame rather than importance:(

@DavidCarlisle “Who's in charge”, but can be in ironic sense (not necessarily).

Well, it blends the two meanings: the proper authorities, who are also responsible (it is exactly for this reason that they are “proper”).

@egreg not "Who's on first" ? :-)

@DavidCarlisle It's “who has duties” upon the matter.

In any case, a chi di dovere implies that those persons are deemed respectful.

@egreg Let's all just agree to blame Heiko, in a language of our choice:-)

@DavidCarlisle Agreed. Although he's much bigger than I.

Oh, my English! I actually meant “respectable”.

@egreg but not as fast if you are on your bike