« first day (1561 days earlier)      last day (3381 days later) » 
00:00 - 17:0017:00 - 00:00

12:04 AM
pgfplot seems a little strange. options added to {axis} do not affect the addplot, and one much be explicit by addition the options to the \addplot itself:

\documentclass{article}
\usepackage{tikz}
\usetikzlibrary{shapes,arrows}
\usepackage{pgfplots}
\begin{document}
\begin{tikzpicture}[scale=0.4]
  \begin{axis}[anchor=center,domain=0:2*pi,mark=none,smooth]
        \addplot (\x,{sin(deg(\x))}); \end{axis}
\end{tikzpicture}
\begin{tikzpicture}[scale=0.4]
	\begin{axis}[anchor=center,domain=0:2*pi]
I thought the options (mark=none,smooth) on the axis itself will do it. But no, I had to put them on the \addplot itself. a little strange.
 
@Nasser Personally, it makes sense to me. :) It's a plot property. :)
 
@PauloCereda I guess this makes sense when I think about it. May be to allow one to add more plots on same axes, each with different property.
 
@Nasser Matter of taste and usage. :)
 
yo'
@Nasser well, you can want to have boxed plot with a smooth line over it and a messy line too :)
 
But it accepted these options on the axes? And there was no warning and no error. I guess it just ignores options that it does not know what to with them.
 
yo'
12:13 AM
@Nasser because it's difficult to make a system of properties that are all properly used.
 
@yo' well, when I write a function that takes options and such, I would first have a pass to check that all options given are valid for this function, if not, will return an error. But many programs do not check all of these things. too much work I guess. Maple and Mathematica also sometimes do this. I can call a command there with option it does not know about, and it just ignores it.
 
@Nasser you are writing as if there is a compiler with functions that return things, this is TeX, it's a macro replacement language, sensible error recovery is virtually impossible
 
yo'
@Nasser but you have zillions of functions that share almost the same set of options and are coded together. And you even have options to a single command which do nothing in one context but do something in another one
 
but the problem, like in this example, is that a user could be using options, which are not valid for one command, and they would not know, because there is no error, and then wonder why it is not working.
 
yo'
@Nasser that happens
but as I say, it's not just complicated to catch it. It's close to impossible
 
12:20 AM
@Nasser ever tried going \begin{figure}[hello nasser] ?
 
yo'
@DavidCarlisle that's a nice one, yeah :D
 
@yo' I've changed that for 2015, if we ever get it out:-)
 
yo'
much better than \begin{tabular}{hello nasser}
@DavidCarlisle oh no! I use it for placing comments in my code! /kidding
 
@yo' I'm expecting to find out how many people have used H without loading float.....
 
yo'
@DavidCarlisle make H default to: Dear user, you're a failure.
 
12:23 AM
@yo' or I'm sorry, I was young at the time
 
yo'
@DavidCarlisle and I wanted to save 2 csnames.
 
@yo' actually for the float [] argument I don't recall trying to save space I think we just never noticed that it never checked, the code basically came from latex2.09, we just added ! handling. Leslie's book has a quote somewhere to the effect that the amount of error checking is correlated to the weather in seattle at the time he coded the feature..
5
 
yo'
hey whose was the other star?! :D
 
@yo' that was me (lurking, just got back from doing the dishes :-) )
 
yo'
@PaulGessler ah ok :D
 
12:28 AM
Is there any better way to do this than what I've proposed?
0
A: How to expand row height while using long text in multirow?

Paul GesslerHere is one approach, using a p column (from the array package) of the specified width (8 cm) to replace your \parbox and \multirow construction. The rest of the cells to the side are placed in a nested tabular aligned to the [t]op, with the spacing around the outer column zeroed for consistent ...

 
yo'
@PaulGessler who knows, maybe yes. Anyways, it smells to me as a misused table for a list
@DavidCarlisle the funny thing is: with TeX you simply don't need to check, \csname handles it nicely :D
 
@PaulGessler Not using \multirow? ;-) Good night.
 
yo'
@egreg nighty night
 
here is a maple example, where it also does not check for options it does not know what to do with them. Same as latex.
 
@egreg I didn't use multirow... ;-) 'night!
 
12:34 AM
I used mikyMouse as option, and no error.
 
yo'
@Nasser but here it's not the same. Try \tikzset{ bla./style= mikyMouse }
or \usepackage[mikyMouse]{geometry}
 
@yo' I get Package pgfkeys Error: I do not know the key '/tikz/bla./style', to which you passed 'mikyMouse', and I am going to ignore it. Perhaps you misspelled it. \tikzset{ bla./style= mikyMouse } so it still ignored it and that is what Maple did. But at least tikz did generate a warning message, which is good
 
yo'
@Nasser ah sorry, ./ should be /. as in: \tikzset{ bla/.style= mikyMouse }
you should get: I don't know mikyMouse
 
I wish there is a course that teaches latex and tikz and all of these things. I would take it. It is easier to take a course in person to learn these things. I never saw any place that give course in these topics.
I've seen lots of places that give courses in indesign and autocad and MS word and such around.
 
12:49 AM
@Nasser seriously: the documentation for TikZ and pgfplots are some of the best I've seen. I picked it up little by little, just searching the doc when I ran into something I couldn't do. This takes time, simply because the feature set of these drawing packages is huge. Don't get disheartened; you'll get there. :-)
 
yo'
@Nasser my university has one, but it's tough to make such a course. Similarly to any programming, you have soo much varying level of knowledge
@PaulGessler the manual is very long, difficult to navigate (simply because it's too long) and has a long learning curve in using it. The index is difficult, too
 
@PaulGessler I know, but latex and tikz is about programming. Just like there are programming courses in c/java/c++ etc... even though lots of manuals around. It helps to learn things faster when in class and can ask questions.
 
@yo' true, but I can't think of a better way to do it. The fact that I can click on most any command and get to its introduction/definition instantly largely solves the index problem.
@Nasser that's what TeX.SX is for! :-)
 
yo'
@PaulGessler but the index is bad: you have dozens of subitems, the PDF ToC doesn't have letters (nor sub-letters), which makes it even worse.
 
@PaulGessler yes, but it is not really the same as taking a course, with a teacher that can explain things in front of you on the black board and you ask questions, etc...
it does not have to a university. May be few latex/tikz experts can set up such a thing, and charge per person some fee. May be a 1 weeks class, or such type of thing.
 
12:59 AM
@Nasser uktug run latex courses (I think normally Joseph and Nicola give them) egreg gives a latex course at his university as far as I know. trouble is actually preparing a course is hard work and most people don't want to pay as they can come here and get help for free (or to c.t.t or just google) very few people have succeed in making a living from that kind of work
 
1:30 AM
@DavidCarlisle maybe we need to make latex/tikz/etc.. more popular in the market place then?
 
yo'
@Nasser how much of "latex" is sex industry?
 
@yo' these are job skils (indeed.com) I do not think the latex cloth with go there, but I am not really not sure. good point.
 
yo'
@Nasser if you click it, you'll see that in the first page, only 2 or 3 are LaTeX, the rest is latex.
 
@yo' one of the position descriptions may be written by a certain PLOS ONE paper author: Administrative Fellow - Able to work in a latex burdened environment. haha!
 
 
4 hours later…
6:05 AM
@DavidCarlisle: A tabularx environment is processed multiple times (perhaps only twice...). How do I know which cycle I'm in?
@DavidCarlisle: See this question:
0
Q: tabularx processes its body multiple times; how do I know which cycle am I in?

WernerAssume that I enjoy setting my tables using an \itemized coding structure: \begin{mytabular}[.5\textwidth] \item head1 & head2% <-- header row \item abc & def \item ghi & jkl \end{mytabular} The idea is that I always have the first row being a header row, with subsequent rows representin...

 
 
2 hours later…
8:27 AM
Can someone tell me what is the problem here?
\documentclass{article}
\makeatletter
\let\oldref\ref
\def\ref#1{%
\newwrite\myfile
\immediate\openout\myfile=\jobname.txt
\immediate\write\myfile{hello}
}%
\oldref{#1}%
}

\makeatother
\begin{document}
%Here are some refs - [\ref{2014.11.14}]. [\ref{2013.10.05.powai}].
\end{letter}
\end{document}
The error is:
! LaTeX Error: Missing \begin{document}.

See the LaTeX manual or LaTeX Companion for explanation.
Type H <return> for immediate help.
...

l.9 \oldref{#1}
%
?
I'm not writing to the aux file here.
And the apparently similar:
\documentclass{article}
\makeatletter
\def\fee{%
\newwrite\myfile
\immediate\openout\myfile=\jobname.txt
\immediate\write\myfile{hello}
}%
\makeatother
\begin{document}
\fee
\end{document}
works.
Never mind, I think the call to \oldref needed to be inside \ref.
 
@JosephWright: I need to take courage now on creating a PayPal account. I really need to learn how that thingy works, specially when dealing with international transactions. Then I might be able to get broken and subscribe to the UK-TUG. :) Do newcomers receive some starter kit autographed by @David? :)
 
8:44 AM
What is the syntax for writes that append? E.g. suppose in the following I want two hellos.
\documentclass{article}
\makeatletter
\let\oldref\ref
\def\ref#1{%
\newwrite\myfile
\immediate\openout\myfile=\jobname.txt
\immediate\write\myfile{hello}
\oldref{#1}%
}%

\makeatother
\begin{document}
Here are some refs - [\ref{2014.11.14}]. [\ref{2013.10.05.powai}].
\end{document}
 
@PauloCereda Still scratching my head over this one cc: @egreg @cfr
 
@Brent.Longborough Rugby woes.
 
9:21 AM
@DavidCarlisle: I've changed my profile photo again, hopefully now it's ok. :)
 
@PauloCereda You should do one carrying a rugby ball in honour if Italy's match against Ireland
@FaheemMitha \write always appends. As I said yesterday (I'm sure) never put an allocation routine like \newwrite inside a macro , and \immediate\openout is what is flushing your file so don't do that every time either
 
@DavidCarlisle yes, you did say don't put \newwrite inside a macro. But I forgot to ask why.
 
@PauloCereda I'm sure the good people over at the vim chat will appreciate it
 
what does "flushing" mean in this context?
 
9:31 AM
@DavidCarlisle :)
 
@FaheemMitha \openout means start a new file, destroying whatever was there
 
@DavidCarlisle so this should be done outside of a macro? Like at beginning of a document, after begin document?
 
@FaheemMitha transfering data from one temporary storage to another storage? :)
 
It seems one cannot append to an existing file in TeX. Per
8
A: How can I open a file in "append" mode?

Martin ScharrerI agree with Joseph that TeX doesn't support appending to an existing file. However it is possible to read the existing content and write it together with the new content back to the file. This is of course much less efficient than simply appending content, but the only possible way in TeX. The ...

I guess in this case, if one does not close the file, then appending is not an issue.
 
@FaheemMitha No, you can't
@FaheemMitha TeX's not meant for general file manipulation
 
9:34 AM
@JosephWright That's a bummer.
 
@FaheemMitha Why?
@FaheemMitha TeX works on a job basis: you know where you are in a job
 
@JosephWright Because sometimes it is useful.
 
@FaheemMitha You'd otherwise have a risk of material accumulating between runs
 
@JosephWright True. But it is possible to keep a file open during a run, right?
 
@FaheemMitha We've no idea what you are actually up to here!
@FaheemMitha Yes, files get closed by \closeout or at the end of the run
 
9:35 AM
See my example above. Shall I repost?
 
@FaheemMitha That doesn't tell me what you are really up to: I don't get what you really want
 
@JosephWright Well, the context is in:
0
Q: Creating a list of labels automatically for use by datatool

Faheem MithaThe following is partly a followup to an earlier question, and Enrico's answer, as given in Automatically creating a table from datatool using references in the text. The method outlined there works, but is rather slow. It occured to me that one way of improving the speed might be to write the r...

Such as it is.
 
@Werner answered:-)
 
@JosephWright So one should do a one time open during the run too?
 
@FaheemMitha Yes
 
9:38 AM
in Vi and Vim, 3 mins ago, by Paulo Cereda
Friends, how offtopic would a question about writing your own plugin be? Background: I have a tool used for TeX automation and I thought of providing some sort of vim plugin, and I'd appreciate some pointers on how to start.
 
@JosephWright ok
 
@FaheemMitha it can be in a macro so long as it's on;y done once.
 
@DavidCarlisle: ^^ You are a vim master, so you can help. :)
 
@DavidCarlisle You mean only call the macro once?
 
@FaheemMitha Yes
 
9:39 AM
@FaheemMitha If you only want to write one file then you only open one file.
 
@DavidCarlisle Mmm, ok.
 
@JosephWright not much mail this morning
 
So \openout destructively opens the file, so call it only once. What does \newwrite do?
 
@FaheemMitha It defines the command like \@auxout that defined which of the16 output streams you are using (there were examples of all these in the answer I referenced yesterday I think)
 
@DavidCarlisle So this should not be called multiple times either? each time it attaches to a different stream?
 
9:42 AM
@DavidCarlisle Fixed. :)
 
@DavidCarlisle No, but TUG renewals at least done
 
I've never heard of these stream things before, but rumor has it there are 16 of them.
 
@FaheemMitha no allocatiion command like \newwrite \newcount etc should be hidden in a macro and called multiople times. I said that several times yesterday as well,
 
@DavidCarlisle Hoping we'll get some from FMi today on the test release
 
Argggggg, scary!!!
 
9:44 AM
@DavidCarlisle I recall you saying \newwrite should not be put in a macro.
and again here:
15 mins ago, by David Carlisle
@FaheemMitha \write always appends. As I said yesterday (I'm sure) never put an allocation routine like \newwrite inside a macro , and \immediate\openout is what is flushing your file so don't do that every time either
 
@DavidCarlisle <3
 
@DavidCarlisle oops, sorry, I misread
Ok, so both \newwrite and the \openout should be called once. But the close is handled automatically, it seems. Unlike many other languages.
I think R gets really upset if you don't close file handles, but it isn't the most robust language.
 
@FaheemMitha You can close it too with ` \closeout\myfoo`.
 
@PauloCereda right, but I gathered that was optional.
 
@FaheemMitha yes closeout is only needed if you want to re-use the same stream for a different file
 
9:49 AM
@DavidCarlisle do you have a reference for this stream thing?
 
@FaheemMitha Some appendix in the TeXbook IIRC.
 
@PauloCereda Ok
after a run, all state is immediately forgotten, right?
 
@FaheemMitha texbook, texbytopic, the answer about writes that I referenced, ... texbook is of course the offical manual for tex.
 
@FaheemMitha TeX is a batch process so it makes sense to close files at the end of the job: after all, the DVI/PDF file gets closed without you doing anything
 
@DavidCarlisle: I'm now scared too! :)
 
9:54 AM
@PauloCereda cricket players are never scared
 
@DavidCarlisle The wicket is down! The wicket is down!
The wicked wicket of East London. :)
Yay @egreg is here!
 
@PauloCereda he always shows up when there is a cricket discussion.
 
@DavidCarlisle Ok
@JosephWright right
I know English people are supposed to like cricket, but Italians?
 
@FaheemMitha It takes his mind off the rugby
 
@DavidCarlisle ah
 
10:00 AM
@FaheemMitha not to mention Brazillians. English speaking, cricket playing, ducks are (as far as I can tell) a typical Brazilian cultural experience.
 
@PauloCereda "Rugby woes" -- good example of a pleonasrn
 
Can `\newwrite\myfile
\immediate\openout\myfile=\jobname.txt` be used in a preamble?
@DavidCarlisle I dunno about Brazilians.
 
@FaheemMitha Yes
 
@JosephWright Ok, thanks. And they don't have to be wrapped in a macro, right?
 
@FaheemMitha well normally you would put it in a macro like \tableofcontents or in a package, you don't really want to be typing that every document
 
10:04 AM
@DavidCarlisle ok
For now, just testing.
 
10:37 AM
Good maen ...
 
11:00 AM
Hi @christian
 
@Johannes_B: Hallo nach Freiberg;-)
 
@Christian Hello to the Black Forest.
@ChristianHupfer Yesterday was Kuchenbasar. It was nice, and busy. Very busy.
 
@Johannes_B: Kuchenbasar? At the University?
 
@christian Kindergarten
@Christian Kindersachenflohmarkt in der Mensa mit Kuchenbasar vom Kindergarten des Studentenwerks.
 
@Johannes_B: Viele Kuchen/Sachen verkauft?
 
11:07 AM
@ChristianHupfer Two of them weren't fancied. One with peaches and the other with carrots. But all the rest got sold.
@ChristianHupfer We could have been much more succesful. Just next to the mensa is the library with around 160/180 students learning for the exams. Fresh coffee for 50 cents, and a piece of cake for 1 Euro. A bit more effort on the parents side, and we could have ....
 
@Johannes_B, sorry, I was away from keyboard
 
@ChristianHupfer We had Wiener (not enough), we could have had some pizza muffins. Die wären uns de bude eingerannt.
 
@Johannes_B: Nevermind. Even my students with special subjects such as BWL/Rechnungswesen do not plan such an event correctly. Marketing? Unimportant... ;-)
 
@ChristianHupfer I wasn't involved in orga, i was just a helping set of hands.
@JosephWright How would we treat this? tex.stackexchange.com/questions/176407/…
 
@Johannes_B A classic TL: I've gone with 'unclear' here as we'd need more data to help
 
11:21 AM
@JosephWright Thanks. TL?
 
@Johannes_B 'Too Localized': we used to close a lot of questions using that reason until it was removed by the Powers
 
@JosephWright Ah, i remember. The event last eve was quite a success?
 
@Johannes_B Yes
 
@JosephWright Nice :-)
 
@Johannes_B :-)
 
@Johannes_B I provided an answer to the \NewDocumentEnvironment question.
 
@egreg Thanks. +1
 
@Johannes_B Damn! Already rep capped. :)
 
@egreg Like the align business, really people shouldn't be doing this
@egreg I wonder how you manage this!
 
12:27 PM
@egreg I got 20 today :-)
 
yo'
@Johannes_B Blame @David (see the last comment)
 
@yo' every scientific paper should have red and purple ruled tables. That's why I made the package. — David Carlisle 10 mins ago
@DavidCarlisle You made my day once again :-)
@DavidCarlisle Voted to close as unclear tex.stackexchange.com/questions/152619/…
 
@JosephWright I wonder the same. :) /looks suspiciously around
 
@egreg, @DavidCarlisle Questions on case changing sent to LaTeX-L
 
yo'
@PauloCereda CRICKET DUCK!!!
 
12:49 PM
What is more healthy? A vegetarian cutlet with pommes frites/french fries/chips or a regular cutlet with plain regular potatoes?
 
@Johannes_B Probably the latter: a lot of vegetarian meat substitutes are a bit heavy on the fat
 
@yo' Quaaaaaaaaaaaaaaaaaaaack! :)
 
yo'
@PauloCereda :)
 
@DavidCarlisle If Paulo is actually a Duck, then he could make Lots Of Money touring and giving shows. Among other things, he's be the only Duck in the World who knows TeX.
 
@Johannes_B @JosephWright was braver than me and wielded his hammer
 
12:57 PM
@JosephWright I knew it is wrong to be a vegetarien. :-)
 
@JosephWright yes just reading now. Must say I never liked bibtex's {} usage (especially as the braces get in the way of other things, such as ligatures) textcase naturally seems more natural to me. On expanding it's tricky, actually more so once you make simple abbreviations \protected\def so you can't easily expand them without expanding other stuff
@FaheemMitha @JosephWright is offended now.
 
@DavidCarlisle why?
 
@DavidCarlisle On the braces, ligatures shouldn't be an issue if you use input protecting entire words rather than single letters (logically you are protecting 'special phrases' entirely, e.g. Fe for iron cannot have any case change at all).
@DavidCarlisle My reasoning for going with {} was that the textcase approach requires specific control sequences, but if we are talking a general approach we can't assume any particular markup. Picking just 'group' as the markup seemed easiest.
@DavidCarlisle I did say all of this is meant for discussion: the alternative approach is also relatively easy to implement
@DavidCarlisle Abbreviations should not be \protected: we still don't have a solution for this at the document level but at the code level it seems clear that \def\myname{Joseph Wright} is a tl not a function, and it is expandable in terms of content so should be in terms of usage.
@DavidCarlisle Please enumerate points on LaTeX-L: I'm keen to collect this stuff up in an archived way for all concerned
@DavidCarlisle I'll perhaps explore the brace business later today
 
@JosephWright yes I know, and it's not unreasonable, still feels wrong though, as if you know the groups were just there for this marking you could remove them but as they are bare groups you have to leave them, so you can not get a result that is just a sequence of letters, unless I missed something.
@JosephWright probably I should comment on latex-l rather than here, encourage discussion
@JosephWright oh that's what you just said..
 
@DavidCarlisle Well here first is fine so long as some summary goes to the list :-)
 
1:13 PM
@JosephWright yes if we formulated that distinction then fully expanding expandable stuff and not expanding protected stuff seems very reasonable
 
@DavidCarlisle One question as I say is what is 'text': as @egreg has said we can imagine that for example \\ might pop up in such a context, and having to map to deal with these things is sub-optimal
@DavidCarlisle Yes: that's what I do in siunitx for a similar issue (covering LaTeX2e means you need to check for robust commands too)
 
@JosephWright can't you skip past non expandable or protected tokens (if we pretent \\ was etex protected)
 
@DavidCarlisle I'd thought about that too
@DavidCarlisle Yes, just needs an appropriate test and so clarity that this is what is wanted
 
@JosephWright and an easy command like \MakeRobust that makes something etex protected, so at this level just worry about etex.
 
Can anyone tell me why the \input file doesn't produce anything?
\documentclass{article}
\makeatletter
\let\oldref\ref
\newwrite\myfile
\immediate\openout\myfile=\jobname.txt
\def\ref#1{%
\immediate\write\myfile{#1}
\oldref{#1}%
}%

\makeatother
\begin{document}
Here are some refs - [\ref{2014.11.14}]. [\ref{2013.10.05.powai}]. \input{dto6.txt}

\begin{enumerate}
\item\label{2013.10.05.powai} something
\item\label{2014.11.14} something else
\end{enumerate}

\end{document}
The obvious answer is that the file is written once the TeX job is complete.
and the file has been closed.
I guess a workaround would be to check for a non-empty file before starting.
 
1:22 PM
@FaheemMitha The file is still open for writing, so you can't \input it.
 
@egreg Yes, I thought that might be the case.
Workarounds?
 
@FaheemMitha Stick with egreg's solution is the most straightforward thing I can think of. :P
 
@FaheemMitha look at \tableofcintents, or in fact yopu could use its internal form \@starttoc that reads the old file before starting to write it again
 
@FaheemMitha That's why the .toc file is written at the end of the job using data from the .aux file.
 
@egreg I see.
Then the TOC file is used at the next go-around?
 
1:25 PM
@FaheemMitha Yes. It's only opened \AtEndDocument, written and closed.
 
@egreg i guess I could copy to another file, then. What is the syntax for that "opened \AtEndDocument, written and closed."?
 
@FaheemMitha Apparently we are not inclined in doing our homework, aren't we? :)
 
@PauloCereda what homework is that?
 
@FaheemMitha how about reading the references you asked for, the ones starred in the right block? :)
 
@PauloCereda I'm not sure what you are suggesting I read, exactly. I don't think I have two of the three (the books), and I don't think the answer in question has that info.
 
1:35 PM
@FaheemMitha Here's a skeleton
\documentclass{article}
\usepackage{letltxmacro}
\makeatletter
\LetLtxMacro\oldref\ref
\newwrite\myfile
\newif\ifopenmyfile
\AtEndDocument{\openmyfiletrue\immediate\openout\myfile=\jobname.txt}

\def\ref#1{%
  \protected@write\@auxout{\let\faheemwrite\relax}{\faheemwrite{#1}}%
  \oldref{#1}%
}
\def\faheemwrite#1{%
  \ifopenmyfile
    \immediate\write\myfile{#1}%
  \fi}
\makeatother
\begin{document}
Here are some refs - [\ref{2014.11.14}]. [\ref{2013.10.05.powai}].

REFS: \InputIfFileExists{\jobname.txt}{}{??}
 
@FaheemMitha That's the homework I was talking about. @egreg did it for you. ^^
 
@PauloCereda I think you seriously overestimate my skills in this area.
@egreg Thanks very much. Should I post this as a question?
 
@DavidCarlisle @barbarabeeton I have a fix for the problem in this question, but is this usage of \fbox correct?
4
Q: "Display math should end with $$" when using MathML with amsmath and \fbox with tex4ht

NasserI just like to get an understanding on why I get this error, only when I use MathML and not in other configurations. pdflatex work fine with this MWE. Only when there is an \fox{}, will tex4ht gives an error. The MWE is below, showing 3 cases. The case with the \fbox fails only with tex4ht, and...

 
@michal.h21 The usage is incorrect, as the math style is not respected (which it is with \boxed).
 
If I remember correctly, the sole purpose of \makeatletter and \makeatother is to @ behaves correctly. If there is no visible @, are those still necessary? E.g. if the code in question calls @, for example?
 
1:39 PM
@egreg this is what I think as well
 
@FaheemMitha If you want to use \@auxout in the body of a definition you must be in a \makeatletter context.
 
@DavidCarlisle Will see what the feeling is about expandability: you only need to worry by hand if the code needs to be expandable
@DavidCarlisle Fair point
 
@egreg yes, I got that. My question was - suppose there are no visible @'s?
Or should that be \@'s?
 
@FaheemMitha I don't understand.
 
@FaheemMitha if the macros you write to the file have @ in their name you need to make @ a letter locally when you read it back in, otherwise you don't need to do that
 
1:44 PM
@DavidCarlisle Ok.
@egreg I'm asking about the case where there is no @ visible in the code. Does it still need to be wrapped with \makeatletter etc?
 
@FaheemMitha if there is no @ then making @ a letter doesn't do anything, but you may still need a makeatletter within the code at the point you input the file, as noted above
 
@FaheemMitha Better version:
\documentclass{article}
\usepackage{letltxmacro}
\makeatletter
\LetLtxMacro\oldref\ref
\newwrite\myfile
\newif\ifopenmyfile
\AtEndDocument{\openmyfiletrue\immediate\openout\myfile=\jobname.txt}

\def\ref#1{%
  \protected@write\@auxout{}{\string\faheem@write{#1}}%
  \oldref{#1}%
}
\def\faheem@write#1{%
  \ifopenmyfile
    \immediate\write\myfile{#1}%
  \fi}
\makeatother
\begin{document}
Here are some refs - [\ref{2014.11.14}]. [\ref{2013.10.05.powai}].

REFS: \InputIfFileExists{\jobname.txt}{}{??}
 
@michal.h21 It is correct really, and no different from using \mbox (for which there are examples in the latex book). amsmath adds \boxed and \text which are better but that isn't the same thing as saying the core latex functions are incorrect. (whatever @egreg tells you:-)
 
@DavidCarlisle :( so I will post a bug report
 
@JosephWright My impression is that few people, have a feeling for expandability, so I'd not expect too many comments on that part.
 
1:52 PM
@DavidCarlisle Hmm, So to be safe always include the \makeatletter?
 
@FaheemMitha No
 
@DavidCarlisle Possibly not, although the denizens of LaTeX-L might be expected to be a little more aware
@DavidCarlisle Really it comes down to the fact that most languages would allow case changing in a functional sense
 
@JosephWright expansion details seem to escape even tex regulars
 
@DavidCarlisle I guess if left to my own devices I'd rather have it
 
@JosephWright yes
@FaheemMitha It is just \def\makeatletter{\catcode\@11\relax}` that is all it does, so use it or not it isn't exactly going to slow tex down.
 
1:54 PM
@DavidCarlisle As I've commented, the need to worry about contextual changes means that a simple \uppercase will fail in any case, so you need a mapping of some form and thus there is less performance hit in doing it expandably than would otherwise be the case.
@DavidCarlisle BTW, I'll make sure I summarise the discussion here for the list :-)
 
@JosephWright es agreed
@JosephWright ah saves me doing it:-)
 
@DavidCarlisle ok
 
@DavidCarlisle Might be easier anyway: I can 'save up' lots of points (and in any case to date I've not exactly been inundated with feedback)
@DavidCarlisle Hoping @egreg will also chime in either there or here
 
@JosephWright I believe you have explained very well the problem. However, using this in a LaTeX2e context exposes to many problems (expandability above all). The idea of “ignoring” braces around the first argument to a macro is nice, but with \label it would be disastrous.
 
@egreg Yes, but I'm not daft enough to put \label on the list (it's currently opt-in)
@egreg @DavidCarlisle would prefer an opt-out list as for textcase, in which case I would make sure \label was skipped
 
2:02 PM
@JosephWright Whatever. ;-)
 
@egreg There's of course a tension between 'ignore LaTeX2e entirely' (easier but will anyone use/test it) and 'tackle LaTeX2e issues' (more complex, danger of retaining issues into the future)
@egreg See the UTF-8 part of my question, for example :-)
 
@JosephWright Yes, I understand. My plan would be making LaTeX3 run only on XeTeX/LuaTeX.
 
@egreg I'm sort-of minded to the same position, but in English it's hard to say pdfTeX is not usable :-)
@egreg Stability also an issue
 
2:22 PM
@JosephWright if a question is solved through a package/distro update is it closed as OT? (I miss TL...)
 
@clemens Yes, 'off topic by convention'
 
yo'
@clemens yep. "OT because it was resolved by a system update." is one possibility.
 
@JosephWright @yo' thanks
 
@egreg it's tempting but then, if you are so tempted you could, like context, say just luatex, and then avoid many expansion issues by hiding all the testing in lua. But I don't think that's clearly a good choice either. hard to predict the future. cf this comment I just got from Martin:-)
While I can understand why you promote T1, it's 2015, so IMHO people should use luatex/xetex and fontspec. — Martin Schröder 33 mins ago
 
@DavidCarlisle I was just referring to Unicode compliance.
 
yo'
2:37 PM
@DavidCarlisle I agree with your point.
 
@DavidCarlisle We know there are issues with both of those engines at the moment :-)
@egreg Yes, I guessed that: as I've said, I'm not keen on anything to add to inputenc in that regard
@DavidCarlisle I wonder what he feels you do when the font is only available in traditional form :-)
@yo' There are lots of things hiding in the 'use a Unicode engine' statement, for example math mode handling is still not 100% 'right'. It's not so easy :-(
@DavidCarlisle Was hoping we might have a mail from Frank by now
 
@egreg ah. It's tempting to go further and make things simply not work, when having three engines triples the amount of code you need and amount of testing to do. For things that "just work" on all three engines which is how people expect it should be then having three engines is a win, but other times...
 
yo'
@JosephWright that's why I agree with DPC :)
 
@yo' you don't need a reason, you should do that anyway.
 
2:52 PM
Ask Heiko to do some coloured blinkind epilleptic attack forcing disclaimer DON'T DO THIS?
What would we do with this question? tex.stackexchange.com/questions/136360/…
 
@Johannes_B: Sorry, quite busy today, so I can not answer immediately ;-)
 
close too broad (there's lots of things you could do, but they will all break something:-)
 
@ChristianHupfer Freaking WLAN. Every once in a while my machine just doesn't do anything net related making chatting buggy. The two posts above were sent in the wrong order.
@DavidCarlisle followed your lead
 
3:10 PM
Love tihs quote:
> Unit Testing Will Make You Drink
 
4:03 PM
Is there a special name for the line $y=mx with $m=1$? In german i wouold say Ursprungsgerade, but apparently this is true for all $m \in \R$.
@ChristianHupfer @DavidCarlisle @egreg ? ^^^ Just got asked by a pretty girl in the library.
 
@Johannes_B First quadrant bisector?
 
@Johannes_B google says that is "through origin"
@egreg yes but not very catchy name
 
@DavidCarlisle Sounds about right
@DavidCarlisle Admittedly not very exciting
 
@Johannes_B erste Winkelhalbierende
 
@egreg @DavidCarlisle @clemens Thanks. She cannot decide right now.
@clemens When did you change to clemens? my fingers wanted to type cgnieder
 
4:10 PM
@Johannes_B :) Yesterday
 
@Johannes_B: Clemens already answered it. You could say one half of the derivative of x^2 too :-P
 
@DavidCarlisle Mathematicians are not famous for choosing catchy names.
 
@Johannes_B: Confuse a pretty girl Ltd. -- We do our best in that ;-)
 
@ChristianHupfer Showed her @egregs answer on screen. She didn't even (seem to) notice the hidden compliment :-D
 
@egreg: Yes, we Physicists have Quarks and... Bosons and ... uncertainty principle ;-)
 
4:18 PM
@ChristianHupfer We have rings, fields and quivers.
 
@egreg: Sorry, I have to yawn ... :-P
 
@clemens: Musicians... once they get famous... see Prince, for instance. I mean, the artist formely known as Prince. Thankfully Unicode was there for him. :)
 
@egreg Should I write a question for this?
 
@PauloCereda lol
@PauloCereda doesn't he call himself Prince again? I thought so
 
@egreg and lots more
Yes, he's back to Prince.
 
4:25 PM
@clemens I have absolutely no idea. :)
 
@FaheemMitha Not sure. Does it solve your efficiency problem with datatool? If so, add a self-answer.
 
@egreg Haven't got that far yet...
 
@PauloCereda I'm not sure but I believe to remember that the reason he called himself <add symbol> was due to some argument with his record company
 
No, it's because he is wacky.
 
@clemens Yep, recording and copyright. :)
 
4:38 PM
@PauloCereda what? It's not included in unicode... :(
 
@clemens If we had someone who could annoy the Unicode committee... @DavidCarlisle :P
@Johannes_B Sorry, I was lost in the beauty of your eyes. < Say this.
 
@PauloCereda Doesn't sound like me.
 
@Johannes_B Oh. :( Did you get her number?
Das Telephonenumber!
 
@PauloCereda No.
 
@Johannes_B Oh. :(
 
4:43 PM
@PauloCereda Don't want the telephonenumber to die? :-p You will have a hard time in Germany ;-)
 
@Johannes_B Articles, bah. :)
 
Die Frau
Die Maus
Die Katze
@JosephWright Just noticed that i messed up the double backslash. Can you change that, please? tex.stackexchange.com/questions/227217/…
 
@Johannes_B Done
 
@JosephWright Thanks
 
@JosephWright you could probably find a few thousand of those if you wanted to fix them all.
 
00:00 - 17:0017:00 - 00:00

« first day (1561 days earlier)      last day (3381 days later) »