« first day (1447 days earlier)      last day (2560 days later) » 

12:05 AM
@Johannes_B seems I'm too late :)
 
 
8 hours later…
7:38 AM
Good maen ...
 
8:03 AM
Not much activity this morning :-(
 
@ChristianHupfer Not yet
 
@JosephWright Let as continue our discussion from "coloured text in plain TeX"
 
@wipet indeed :-)
 
@JosephWright First. You cannot suppose all. Your code test if \pdfoutput is \relax or \undefined. But somebody can write \let\pdfoutput=a. The probability of this is the same as probability of \csname pdfoutput\endcsname which makes \pdfoutput to \relax. In both cases this is mistake of the user and we cannot solve it by complicated macros.
 
@wipet True, but the \ifx\csname <thing>\endcsname problem is common enough to worry about in 'defensive' programming
@wipet For example, LaTeX's \@ifundefined needs to be expandable and so does leave things equal to \relax. That can easily be available in plain with \input miniltx, so it's something to worry about.
 
8:21 AM
@JosephWright Second. How many \color{spec parameter} will be in the document? We really need to save eq-table by complicated expandafters? Note that the string-pool isn't saved by this.
 
@wipet No idea, and not important
@wipet It's standard good practice not to leave anything around as equal to \relax by accident
 
@JosephWright The usage of \@ifundefined{pdfoutput} (from miniltx) is the same nonsens. It should be to written \ifx\pdfoutput\undefined and all else is mistake of the user.
 
@wipet Yes, but people make mistakes
@wipet Also, people could be loading other third-party code
@wipet I don't see the issue with the standard test construct here
@wipet Normally, I'd only ever use a (hopefully) undefined macro for \let\foo\undefined
 
@JosephWright But when third-party code say \let\pdfoutput=a then we cannot to repair this. LaTeX is full of such redundantly complicated code. This is one of the reason why I never support LaTeX, never use LaTeX and dislike LaTeX.
 
@wipet We can't cover every case but we can cover the common ones
@wipet I know you don't like LaTeX (or I guess ConTeXt/Lollipop/...), but for most users having to learn the low-level stuff is simply not what they want to do
@wipet Providing interfaces for different kinds of code is what most programming is, otherwise we'd all still be writing machine code on punched cards
@wipet In the driver code, Heiko checks that \pdfoutput is still the primitive: I didn't feel like adding that test too :-)
 
8:43 AM
@JosephWright Yes. I copied (from your code) the idea of \colorname macros which I feel as nonstandard. More usual is \csname color:name\endcsname and here is common practide to test such macros against \relax and leave them set as relax if this doesn't full up the eq-table exessively. The string-pool is not spared anyway.
 
8:55 AM
@JosephWright "If \pdfoutput isn't defined then \ifnum\relax>0 crashes.
@wipet Of course, but the number test here is only expanded if the earlier check for existence of \pdfoutput passes" OK, but if `\pdfoutput` is defined (checked by previous pass) then you needn't to put it between `\csname...\endcsname`. Uff.
@JosephWright Why there are three expandafters before \pdfmodetrue in your code? Aren't them redundant? LaTeX packages are full of such peculiarities...
Maybe you mean something like this:
\edef\tmpa{\meaning\pdfoutput} \edef\tmpb{\string\pdfoutput}
\ifx\tmpa\tmpb Is \else Isn't \fi primitive.
 
 
1 hour later…
10:26 AM
@wipet That but also using \pdfprimitive
@wipet In general I don't like anything inside conditionals, hence the \expandafter chain. The only exception I make is for literals (letters, numbers)
I.e I strongly favour using \expandafter to deal with the conditional before inserting any further tokens
@wipet The chain here isn't strictly needed, but while such a construct is always safe the reverse is not the case
 
is there anything to do this: tex.stackexchange.com/questions/145892/… the 'right' way?
 
@Serthy Meaning?
 
10:45 AM
well egreg said in his last statement that he not recommends using this style, so maybe there is an alternative
 
 
2 hours later…
12:29 PM
@DavidCarlisle I think the reason we don't have many questions about \widowpenalties is that it's a hard primitive to use correctly :-)
Of course, l3galley does get this all right
(At the cost of quite a headache for me)
 
@JosephWright :-)
 
@DavidCarlisle I remember spending a lot of time working this out: it's really not clear in the e-TeX manual how things work
 
 
1 hour later…
1:53 PM
Just for information: there was an spam edit for answer tex.stackexchange.com/questions/16850/… Please have a look for this. The spammer was an annonymous user ...
 
2:37 PM
@Kurt I see no recent change to the answer
 
@Kurt @tohecz It happens, the review queue exists also for this purpose.
 
2:54 PM
@egreg well, the two rejects with no approval are a good sign :)
 
@tohecz :)
 
and the fact that we just managed to prove something in algebra is a good sign, too :)
 
@tohecz :-)
 
however, we have the term "continuous" there -- is it still algebra? :P
 
@tohecz I've used continuous in algebra all my life.
 
3:26 PM
@egreg I know, it was a joke :p
 
too many squares here
 
@DavidCarlisle and all are red according to some people, with the exception of the Kurt's one :)
 
firefox crashed for about 5th time today:(
 
@DavidCarlisle not here ...
 
@tohecz yes well I run nightly so it's last night's build... (36.0a)
make that 6:((
7, I've switched to chrome for a bit:-)
 
4:12 PM
Remark. I do hate commutative diagrams, but I do love them.
 
@egreg, the great authority of miktex/windows usage:-)
 
@DavidCarlisle I used Windoze perhaps for a couple of hours total in my life; thus I'm a big expert.
 
@egreg we could direct the OP towards arara or something else that might read comments?
 
@DavidCarlisle You're surely a big arara expert.
 
@egreg and so I commented
 
 
1 hour later…
cis
5:34 PM
@DavidCarlisle And? Are you keen on that cake, I promised, if there comes a "2014-Editon" of blkarray? Which kind of cake do you prefer?
 
Has anyone tried MacTeX on Yosemite? Does the "TeX Distribution" system prefs pane work?
 
@Szabolcs I'll tell you next week. Tomorrow I have my talk at the GuITmeeting, I wouldn't try installing the system now.
@Szabolcs There have been problems also with Mavericks, IIRC
 
@egreg You really mustn't then. Something always breaks. I only installed it on a system that I do not need to depend on.
 
6:15 PM
@egreg ooh how nice! Any preview on the subject?
 
If anybody online right now uses / customizes emacs, could you take a gander at this Q? superuser.com/q/827818 Thanks :)
 
 
1 hour later…
7:47 PM
@SeanAllred you need @PauloCereda
 
8:13 PM
@PauloCereda I'll talk about l3regex
 
8:53 PM
@DavidCarlisle Ah yes, the source of all emacs knowledge :-)
 
9:30 PM
@egreg Cool
I need to decide on the UK-TUG meeting talk(s): probably l3build and perhaps beamer overlays, with a third talk if really required on siunitx revisions!
 
@JosephWright Yay!
@egreg Yay!
@Joseph: get Nicola to talk about hats and araras. :)
 
 
1 hour later…
10:39 PM
Hello ducks! (and people, too...)
 

« first day (1447 days earlier)      last day (2560 days later) »