Hello!! Could you take a look at my question: tex.stackexchange.com/questions/264290/… about how to write the bibliography in the english language while using the package \usepackage[english,greek]{babel} ?
I am close to finishing my PhD. Some of my peers use LaTeX. Obviously, it looks much better than Word and alike. As I wrote a few papers that go into my final thesis, I however have numerous formats depending on the journal in question. I would like to transform them all into one final document w...
@DavidCarlisle hi David. Not yet I'm afraid. I found I could keep the environment in and delete the contents and it compiles fine. As I add back in, it seems to fail when it has to break over a page. But other tables do that no problem. Will continue investigating...
Thank you for the enquiry, though. I appreciate it.
@JRichardSnape oh that's probably a bug in the page header code, the usual problem is using a multiline header and using \\ assuming it means normal line break but if it's inside the scope of longtable it will be a end of longtable row, any code that may be inserted in such a place should take action to normalise its context (parbox, minipage etc do that automatically (\@parboxrestore) but if the heading code uses lower level access things can go wrong....
\documentclass[a4paper]{article} \usepackage{mathtools} \usepackage[inline]{enumitem} \begin{document} Code: \begin{enumerate*}[label=(\alph*)] \item This is the first item, plus a displayed formula \begin{equation*} \bar{i_s} =2 \end{equation*} \end{enumerate*} \end{document}
It does not compile unless one is using enumerate instead of enumerate*. Any ideas?
\documentclass[a4paper]{article}
\usepackage{mathtools}
\usepackage[inline]{enumitem}
\begin{document}
Code:
\begin{enumerate*}[label=(\alph*)]
\item This is the first item, plus a displayed formula
$
\bar{i_s} =2
$
\end{enumerate*}
\end{document}
@daleif ^^ I would guess that the inline list thing is putting its content in an \hbox to measure something, and display math in an hbox dies as you show....
@DavidCarlisle from the code, yes there is a macro to store something in. It just seems a little strange. And Javier ought to mention in the manual that there are certain things one cannot do in inline lists. I'm guessing we still need paralist for certain things.
@DavidCarlisle directly. I am adapting someone else's template, though - so I guess I need to look at headers etc (maybe remove any redundant usepackage as well - I have not been chopping at all). I think the onus is on me to make the example minimal. I'll probably find it while doing so, if precedent is anything to go by.
@DavidCarlisle And, in fact, your header prompt I think has solved it. The table was breaking over a page that had a footnote as laid out currently. I think there was something in the footnote that was interfering - if I comment out the footnote, I can restore the table and it compiles fine. Now to work out what's in the footnote that it doesn't like and determine a way to sort that out...
@daleif it is consistent with this not giving a paragraph:
\documentclass[a4paper]{article}
\usepackage{mathtools}
\usepackage[inline]{enumitem}
\begin{document}
Code:
\begin{enumerate*}%[label=(\alph*)]
\item This is the first item, plus a displayed
formula
\end{enumerate*}
\end{document}
@daleif ^^ the whole item is set in horizontal mode. So paragraph breaks don't give paragraphs and displayed math doesn't give displayed math. It's pretty odd to want displayed material in an explicitly inline list, anyway.
@DavidCarlisle not really, if you are doing a longer explanation. I might have done this differently if it was me, but I see why the user would have wanted to use this (better structure). But it should at least be noted in the manual that this thing might happen.
\documentclass[a4paper]{article}
\usepackage{mathtools}
\usepackage[inline]{enumitem}
\begin{document}
Code:
\begin{enumerate*}[label=(\alph*)]
\item This is the first item, plus a displayed formula\\
\parbox[t]{\linewidth}{%
\begin{equation*}
\bar{i_s} =2
\end{equation*}}\\
and some more text
\end{enumerate*}
\end{document}
@daleif the whole list is constructed in an hbox then at the end \unhbox \enit@inbox so no vertical mode material will work.
Just to follow up, my longtable & footnote interaction issue appeared to be repeatable. It doesn't seem to matter what the content of the footnote is - plain text triggered the problem. So, I looked through my preamble and found the following which seemed pertinent. I commented that out and, hey presto, everything worked again with the longtable breaking over the footnote.
% Footnote rule
\setlength{\skip\footins}{0.0469in}
\renewcommand\footnoterule{\vspace*{-0.0071in}\setlength\leftskip{0pt}\setlength\rightskip{0pt plus 1fil}\noindent\textcolor{black}{\rule{0.25\columnwidth}{0.0071in}}\vspace*{0.0398in}}
I wondered if it was easy to explain why this preamble would cause the latex to hang if a longtable breaks over a page with a footnote?
@JRichardSnape not sure without a test example I could run, it's pretty odd code though it resets left and right skip globally, those settings are not skoped by any group so it depends where \footnoterule gets executed (at end of minipage and end of each page in the output routine, normally, but who knows...)
@JosephWright not sure where best, probably in luaotfload it could, when defining a new font look through the list of fonts it has already done. for the bare format in the test suite might try using the font loader callback to do just that.
@JosephWright for a start, yes. If that works can decide if it ought to be there always.
@JosephWright perhaps or perhaps into an ini file used to build the lualatex format for the test suite? Not sure. I'll get something working first. If it's in regression-test.tex (or test2e.tex for the 2e suite) it may need to special case fonts already loaded in the format.
@DavidCarlisle I have a 17 line minimal example that demonstrates the problem. Is it bad form to post samples here? I can put in a pastebin if that's better.
\documentclass{article}
\usepackage{lipsum}
\usepackage{longtable}
% Footnote rule copied from my big document - if I comment the following two lines - all OK
\setlength{\skip\footins}{0.0469in}
\renewcommand\footnoterule{\vspace*{-0.0071in}\setlength\leftskip{0pt}\setlength\rightskip{0pt plus 1fil}\noindent\textcolor{black}{\rule{0.25\columnwidth}{0.0071in}}\vspace*{0.0398in}}
\begin{document}
My very elegant mother just sat under newton's protrait.\footnote{One of the few mnemonics I can remember}
my understanding is not great enough... Could you explain / point me to a resource that explains where you find "thousands of lines of \par ->{\unvbox}" and what that means?
I'm keen to gain a bit deeper understanding: the rule is not written by me, nor important to me but I would like to be able to debug such problems for myself.
@JRichardSnape There are other things wrong but the immediate cause of the loop can be fixed by adding \endgraf so you do the vertical space in vertical mode:
\renewcommand\footnoterule{\endgraf\vspace*{-0.0071in}\setlength\leftskip{0pt}\setlength\rightskip{0pt plus 1fil}\noindent\textcolor{black}{\rule{0.25\columnwidth}{0.0071in}}\endgraf\vspace*{0.0398in}}
@JRichardSnape so what I did was add \tracingall before the table and kill the job before the log filled my disk then look at the trace of what it was doing, the log clearly shows a tight loop:
@PauloCereda Yes - I just tried it out on the terminal. pdflatex paused in the terminal with a question mark and I hit enter... Many thousand repetitions later I managed to kill it... Still, at least I have a method now, even if I must approach with caution.
@TorbjรธrnT. Thanks @TorbjørnT., indeed I found that answer when I google tracingall (that's become my first action on encountering programming keywords I don't recognise!)
By the way - thanks for the really helpful atmosphere in this room - it's been a pleasure to interact and I'll no doubt hang out here quite a bit as I work towards completing my thesis... (anything for a distraction!)
@yo' -- in general i agree with you. but what a researcher submits to a journal needs to be more than comprehensible -- it needs to be accurate. by that i mean that notation must be consistent, in math, delimiters must be properly matched, tables and displays must not be so wide that they are cut off by the edge of the paper, etc. [cont'd]
[cont'd] a good proofreader should be able to fix those things, if they're noticed, but even the best proofreader, if not expert in the field, can miss some important things, even miss them enough to not post a query to author. it's embarrassing to have to print errata/corrigenda.
@barbarabeeton that's I think what we agree on. But none of these is dependent on LaTeX per se. I can imagine getting a handwritten article that would be fine, for instance.
@egreg Rule 2: Do not discuss the premises of the question if they are fixed. :D
@DavidCarlisle On the \Ucharcat emulation: what do you think name-wise? I'm wondering about \char_token:nn or \char_generate:nn , 'Generates a character token with the given category code'
@yo' -- yes. and it still happens. (just hope that the author's handwriting is legible ...)
@yo' -- submissions to ams publications are vetted by an editorial board outside the office; they send blind submissions to referees. only after such review, and possible revisions by the author, do they get delivered to the production stream, "approved". we hope that all inaccuracies have been exterminated by then, but there are still queries to authors.
@barbarabeeton the article got almost unchanged by the typesetter, and IIRC the only proofs were fresh reference updates :) So it was, IMHO, a good job on both sides!
@JosephWright generate probably. for lua are you defining it directly or loading ucharcat.sty (and either way do you want me to add error messages for bad catcodes to the latter)
@DavidCarlisle Plan is to have \char_generate:nn as a wrapper, with some filtering (no active chars, no catcode-10 ^^@) to keep consistency as far as possible between macro approach, Lua approach and \Ucharcat
@DavidCarlisle expl3 has no deps beyond LaTeX2e itself so I'll make a copy of the Lua code
@JosephWright yes safe enough: if anyone really has thousands of catcode tables something is wrong (I didn't use very top in case someone used an allocator coming down, just somewhere randomly high:-)
So I have now a working define_font callback that warns me that there is already a font with that name/size defined but it seems I have to return a valid font table (it didn't like nil at all) and if I return a font table (whether it's from font.read_tfm or by reference to the existing font) then it does what it always did and defines duplicate fonts with different ids that show as different names in the log... so one step forward, two steps back...
I could probably use something like tex.sprint(\let\oldname\newname) in the duplicate case but that will show in TeX tracing and I was hoping to do it in lua and really share the font data
@StefanKottwitz -- there's a comment in tex.stackexchange.com/q/266665 pointing to an answer of yours in tex.stackexchange.com/q/14071 ; unfortunately, the latter lacks a visual example (it would be much more compelling with one). the new question does have a good example that maybe could be cribbed, so the new question could be declared a duplicate.
@PauloCereda -- please ask! (why in the world should i mind?)
@PauloCereda -- on a totally different subject, i missed wishing you well on your country's independence day a couple of days ago. belated best wishes.
@barbarabeeton Thank you! :) I stole the \acro macro (should I say m\acro?) definition from ltugboat. I am using in my own thesis and I am loving it, the acronyms look so nice! I had to include an acronym in one of my sections, and it worked like a charm. But then I took a peek at my TOC and noticed the space after the acronym is gobbled. I am not sure if I should use \acro inside sectioning commands, but I was curious if you end up with such situation.
@PauloCereda -- i can't offhand think of any way to check for the space and at the same time keep the \@ effective. (probably @DavidCarlisle could do so.) so i'm just suggesting to use a slash-space after the \acro expression when it happens in a heading. the truth is, tocs are so seldom used in tugboat that we either never noticed it, or just use that hack.
@PauloCereda -- i'm glad you like the tugboat \acro. the only "real" small caps i've ever seen that do justice to acronyms are the ones in the economist.
@barbarabeeton I love so so much, it's the best way of typesetting acronyms I've ever seen in my entire life. :) It's so lovely to see how we learn to enjoy the typesetting itself as much as the content. I'm writing my thesis draft and enjoying every macro I am using. :)
@PauloCereda -- if you get different results between 2014 and 2015, can you determine which versions of which package(s) are involved, please. would like to add that to comments kept about such oddments.
@PauloCereda -- i'll have to report this to christina thiele; i'm pretty sure it was at least partly her idea, and the original macro was our joint effort. (and yes, i'm very fond of it too.)
@barbarabeeton: @Christian is right, I don't get the gobbled space in TL2015, only in TL2014. I added \listfiles and the output is exactly the same. So it has to be the kernel itself, I suspect.
@PauloCereda -- and while you're at it, please add this section test: `\section{Let us talk about \acro{DUCKS}. and other birds, magnificent creatures that can swim or fly or burrow}` -- this will force an additional line in the toc at a place where you can really see what happens after the period.
\subsection{\texttt{\textbackslash @} discards spaces when moving
(pr/3039)}
\begin{verbatim}
>>Number: 3039
>>Category: latex
>>Synopsis: \@ discards spaces when moving
>>Arrival-Date: Sat May 22 09:01:06 1999
>>Originator: Donald Arseneau
>>Description:
The \@ command expands to \spacefactor\@m in auxiliary files,
which then ignores following spaces when it is reprocessed.
\end{verbatim}
@PauloCereda ^^^^
@egreg speaking of speed: reported in 1999, fixed in 2015....
Update: a previous version of this question wrongly attributed the problem to use of minitoc to get section-level tables of contents (\secttoc).
I've got a complicated document with a tables of content, and which switches between one and two-column mode.
I'd like to remove the subsection headi...
@JosephWright so here's a define_font callback that detects duplicate fonts (but doesn't do anything about it)
%\starttext
aa
\directlua{
function mydeffont(n,s,i)
print("here: " .. n .. ' ' .. s .. ' ' .. i)
f=font.read_tfm(n,s)
for ii,vv in font.each() do
if (n == vv.name) then
print('n: ' .. ' ' .. (vv.name or '?') .. ' ' ..(s or 'size') .. ' ' ..(vv.size or 'size2') )
end
end
return f
end
callback.register('define_font',mydeffont)
}
\font\a=cmss10
\font\b=cmbx10
\font\c=cmbx10
\immediate\write16{a: \number\fontid\a}
\immediate\write16{b: \number\fontid\b}
\immediate\write16{c: \number\fontid\c}
and here's some code you can run at intervals to detect duplicate fonts and fix things up so the log does the right thing:
\documentclass{article}
\font\a=cmr10
\font\b=cmr10
\begin{document}
\ifx\directlua\undefined\else
\def\zzz#1#2{%
\global\expandafter\let
\csname#1\expandafter\endcsname
\csname#2\endcsname
}
\directlua{
for i,v in font.each() do
for ii,vv in font.each() do
if (i > ii and v.name == vv.name and v.size == vv.size) then
tex.sprint('\string\\zzz{' ..
string.sub(tex.fontidentifier(ii),2) ..
'}{' ..
string.sub(tex.fontidentifier(i),2) ..
@JosephWright if it was after each \font you could skip the outer loop and just check that font, but I was worried that redefining \font would show up in tracing or have other artefacts but I couldn't make the callback work. may try again but out of time
@PauloCereda -- don't be sorry. if we need to change something in the tugboat classes, better to know it now than be surprised when we first see the glitch in print.
@JosephWright found some information in the change log not in the manual
* The 'define_font' callback interface has changed a little. It is now: retval = function (name,size,fontid)
where fontid is the internal font number of the font that is currently being defined, and retval can be either a table (as before) or a different, previously defined fontid. This is useful if a previous definition can be reused instead of creating a whole new font structure.
so it seems I can return the old number to force sharing.... (will try that in a bit)
@JosephWright bingo, that works, note y and z both logged as \c :
%\starttext
aa
\directlua{
function mydeffont(n,s,i)
print("here: " .. n .. ' ' .. s .. ' ' .. i)
for ii,vv in font.each() do
if (n == vv.name) then
print('n: ' .. ' ' .. (vv.name or '?') .. ' ' ..(s or 'size') .. ' ' ..(vv.size or 'size2') )
f=ii
end
end
return f or font.read_tfm(n,s)
end
callback.register('define_font',mydeffont)
}
\font\a=cmss10
\font\b=cmbx10
\font\c=cmbx10
\immediate\write16{a: \number\fontid\a}
\immediate\write16{b: \number\fontid\b}
\immediate\write16{c: \number\fontid\c}
@JosephWright yes also with another addition to the manual request, apart from anything else check that returning a fontid is supported (a change log from 2009 isn't the most authoritative looking place)
When there are many short footnotes on one page, it looks bad to use a whole line for each. How would one go about typesetting short footnotes in columns, more or less like below, in LaTeX?
\documentclass{article}
\usepackage{showframe}
\begin{document}
\begin{tabular}{lll}
$^1$\,First footn...
@ChristianHupfer I was guessing right. But I also knew that he hadn't really thought it through. Asking for a three column layout without mentioning longer footnotes is incomplete.
Hello chat. I think there's a small problem with amsmath. The basic way of telling \dots that it should give centered dots is using \DOTS inside the next macro, so \dots looks ahead and finds it. In any case, if the next macro is \long it doesn't work.
So, for instance, lots of arrows in the amsmath.dtx file are defined with \renewcommand{\iff} which doesn't work because the asterisk is missing, \renewcommand*{\iff} would work.
But it would be better to fix \dots to work with \long commands.
@Manuel -- there are several items in the bugs list for amsmath regarding \dots, but none that address exactly the problem you mention. please submit a "working" example, either as a question on the regular forum here (unless it's already there and i missed it), or to the ams tech-support address. i can't promise an immediate fix, but i can promise that it will get onto the list with a testable example with which to demonstrate that any future fix is effective.
@egreg -- uh, please be specific here. i will be happy to leave guidance for the next update (which, as you know, is unscheduled), but if it's as nonspecific as your comment here, it will simply get swept under the rug. (you can even send a long list of your specific complaints, off line, please. if a problem can be "fixed" by a change in amsldoc, meaning possibly explaining how to do a workaround until the underlying code problem is fixed, then there might be a chance of doing that.)
@PauloCereda -- oh, good news! you're getting in a lot of mileage (kilometerage? have to ask my canadian friends) this year.
The standard way with amsmath is using \DOTSB inside the macro (\dots looks ahead for \DOTSB).
Note: I don't know why it's important that the macro is not \long so it works, hence the asterisk in \newcommand* is mandatory (which is correct anyways).
\documentclass{scrartcl}
\usepackage{mathtoo...
@barbarabeeton In that answer, if you remove the * after \newcommand (which would make \by a \long macro) the \dots are not correctly aligned.
@barbarabeeton Since in the package there are many definitions like \renewcommand{\iff} those don't work (e.g., \iff \dots \iff doesn't give centered dots, if it was defined with \renewcommand*{\iff} it would work). But the thing is that this seemed to work a few years ago (look at tex.stackexchange.com/questions/649/…), so may be it has to do with recent updates to the kernel?
@egreg I am a huge fan of Claude Rains. :) I once heard a story of him taking his son to watch The invisble man in the movies, in which he had the lead role. As the invisible man only appears in the very last moment of the film, his voice became the trademark. During the session, he was talking with his son about the making of of each scene, when people started to scream because they recognized the same voice speaking in the wild!
@barbarabeeton Yay! For us, it's kilometers too. :)
@egreg -- ah, you meant the .dtx file, not the user doc. yes, that could use quite a bit of help. i'll "strengthen" the commentary in our bugs list in that regard. but again, if you have some specific suggestions, they'd really be welcome.
@Manuel -- (personal opinion: it's a mistake to make all \newcommands \long. there's a reason why knuth made a distinction. no single symbol should be permitted to be \long!) i've come up with a test file for future amsmath reference. i think some of this can be handled by better documentation, one area being in "how to define one's own commands for symbols".
question on your answer: why use scrartcl (other than possible force of habit)? if receiving that in a test file at ams, we would immediately change it to article. we claim to be compatible with the "basic" latex classes, but can't automatically keep up with all others, no matter how worthy.
@barbarabeeton as I added to the answer on the question making it treat long and not long macros the same way is only a small addition, 2 extra lines of code. Seems easier than documenting that they behave differently.
@DavidCarlisle -- in my preliminary writeup for the bugs list, i've given the relevant references in tex.sx, and when i do the detailed writeup, i will mention that. even so, there will still be exceptions that need to be addressed individually (\dotsb etc. when they're at the end of an expression with nothing checkable following. did you test that with the \long addition?)
@barbarabeeton As David says, I think it's better to edit the code so it works in all cases, no need to say that for robust or long commands it doesn't work. By the way, \newcommands are long since the beginning, and years later they thought the same as you, but couldn't โredefineโ the macro, so they added the asterisk variant; so I heard.
@barbarabeeton I add it automatically to all my documents. It wasn't planned as a document to send as a bug report, it was just an answer.
I think it's interesting to see that it did work some time ago. By the way, good addition the fact that it doesn't work with robust commands. (May be adding a few more lines, rather than the two that David says to solve this for โevery caseโ?)
@JosephWright need to try to isolate the code, but seems that my loop iterating through font.each() works in the initex run and in a normal document if explicit in the document but not if inserted in everyjob, I suspect something not set up...
@egreg -- oh dear, i just know you don't care for much of the music i like very much. but maybe you'd like one upcoming offering from the boston early music festival that i'm tempted by -- two virtuoso lutenists "dueling" over selections from the 17th and 18th centuries.
This is TeX, Version 3.14159265 (TeX Live 2015) (preloaded format=tex)
**\showhyphens{Treemunch Thruston}
Underfull \hbox (badness 10000) detected at line 0
[] \tenrm Treemu-nch Thrus-ton
@egreg -- no, i didn't. but that surely would have been exciting! (the other day on the radio, we heard a performance of handel's music for the royal fireworks played by (i think) the band of the queen's horse guards. the tympani had a most distinct sound; it clearly wasn't any "ordinary" symphony orchestra! quite enjoyable.)
@DavidCarlisle well, he seems to be so well known as Bill Thurston, that it would be ok probably. Just like me using Tom Hejda instead of Tomáลก Hejda :) Actually, the TUG article is the first one for which I use the familiar form :-)
@DavidCarlisle I am trying to make a start with TEI-XML. If I use xi:include for the equivalent of TeX \input, I know I need to include the same xmlns declaration in the included file as in the master file. But I am getting errors using entities like – in the included file (parser says it's not defined). What do I need to add? (Feel free to send me to SO, obviously this is not a TeX question)
@DavidCarlisle From trial and error it looks like I need the whole <!DOCTYPE ..> declaration in each subfile.
@AndrewCashner if you use include then each file needs to be well formed so needs at least the entity definitions if not a complete dtd, if you include files as external parsed entities then they are included by the parse of the main document so don't need (and can't have) their own dtd (similarly if included as parsed entities they inherit any namespace settings from the outer document)
@AndrewCashner or simpler is not to use the named character entities. I thought TEI was mainly relaxNG based these days rather than DTD. (I don't use it myself)
Update: a previous version of this question wrongly attributed the problem to use of minitoc to get section-level tables of contents (\secttoc).
I've got a complicated document with a tables of content, and which switches between one and two-column mode.
I'd like to remove the subsection headi...
When there are many short footnotes on one page, it looks bad to use a whole line for each. How would one go about typesetting short footnotes in columns, more or less like below, in LaTeX?
\documentclass{article}
\usepackage{showframe}
\begin{document}
\begin{tabular}{lll}
$^1$\,First footn...
The standard way with amsmath is using \DOTSB inside the macro (\dots looks ahead for \DOTSB).
Note: I don't know why it's important that the macro is not \long so it works, hence the asterisk in \newcommand* is mandatory (which is correct anyways).
\documentclass{scrartcl}
\usepackage{mathtoo...
