TeX, LaTeX and Friends - 2015-02-14 (page 1 of 2)

cfr

12:20 AM

@PauloCereda KDE is driving me insane!

Paulo Cereda

@cfr what happened?

cfr

@PauloCereda Arch Linux is fantastic. Except. Except its KDE transitions... are... interesting. To keep Kile working properly, I had to switch to Kf5 Plasma. So Kile works but almost nothing else does. I have the teeniest icons in my icon-task thingie I've seen. (I don't usually use this, but when I added regular launchers, half of the panel disappeared and I had to make a new one.) Konsole won't remember custom short-cuts.

yo'

btw, I've made an achievement: I convinced a bunch of people on Music: Practice & Theory to start visiting the chat regularly

@cfr ah :( beeing a MATE guy myself, I'm afraid I can't help you :(

Paulo Cereda

@cfr Oh my! KDE 4.0 and KDE 5.0 coexisting in the same place?

"Coexist" might not be the word here.

cfr

You can't add windows as tabs to others (and kwin doesn't think 2 firefox windows should be grouped automatically. And the paste appelet thingy (which I used for code snippets) no longer exists. And the date is no longer readable and everyone is on UTC all of the time. And and and. Pah!

@PauloCereda Exactly. If it is coexistence, it is certainly not of the peaceful kind. It is a horrible, horrible mess. (Not as bad as when they brought KDE 4 out. That is, it has only completely crashed once so far. But not good at all.)

I have, for example, both K Info Center (Kf5) and K Help Center (KDE 4). Depending on how you access help from an application, you get one or the other. The former doesn't actually contain any help. The latter contains outdated help.

Paulo Cereda

12:30 AM

@cfr Oh my. Arch is a rolling release, right?

cfr

@PauloCereda Yes. And staying on KDE4 became difficult because, of course, the stuff there started to break or disappear... So Kile lost all plugins, for example, and it was impossible to configure any settings form new applications without switching.

And I would add that I only use the stable repositories. This is not testing.

@PauloCereda On the bright side, at least the cashew has gone ;).

Paulo Cereda

@cfr Ouch, that's very annoying and, worst of all, counterproductive. I was gonna suggest to remove some sort of cached config directory in your home, but since both major releases are in place, we might break one in favour of the other.

@cfr ooh! :)

Surely there's been some sort of packaging conflict.

cfr

@PauloCereda Nope. At least, I haven't ignored any such conflicts. I am quite careful. I removed all packages which were going to conflict prior to switching and I followed the recommendations to manually uninstall core KDE4 stuff prior to installing the new dispaly manager and plasma.

Paulo Cereda

@cfr How about reinstalling both versions?

cfr

Oh, also, Kf5 doesn't appear to run autostart scripts. It just ignores them.

@PauloCereda Why?

I don't think most of this is specific to me...

Paulo Cereda

12:37 AM

@cfr To fix some possible overwwriting.

cfr

The teeny-weeny icons probably is. Other people seem to get that only with Air.

Paulo Cereda

I usually have crazy advices. ;)

cfr

@PauloCereda No, usually I would think it quite sensible. It is just that I've been spending some time on Arch's forums and it seems unlikely we could all solve our problems that way :(.

Indeed, KDE seem to think it reasonable that they've broken startup scripts. Something to do which thinking another component should handle them...

Paulo Cereda

@cfr I thought KDE wouldn't require backwards compatibilty.

yo'

your discussion reminds me of Gnome3...

cfr

12:40 AM

@yo' Except this is like having half your system with Gnome3 and the other with Gnome2...

The bits don't talk to each other.

yo'

@cfr well, I mean the discussion about the features being removed

cfr

@yo' Oh, yes. There is that. You also can't have different wallpaper for different desktops. And they've done something to make almost all the sys tray icons disappear. (Dropbox is lost. However, I only actually used it in a cron job so I'm not terribly bothered.)

Some people got the Dropbox icon back. I think it depends how you start it.

yo'

@cfr there've been quite problems with dropbox some time ago in MATE, also, dropbox has made some changes. I've got a completely new icon now, and I didn't install any upgrade AFAIK

cfr

@yo' Dropbox is not that well behaved.

I can't say whether I've got a new icon or not ;).

I got my SpiderOak icon back. But it doesn't autostart when I log in. Neither does the session remember Konsole tabs. It is very, very icky indeed.

Though setting the preferences appears to lose the application?

Yep. If you say 'OK' in preferences, the application quits.

Paulo Cereda

@cfr: I will set up an Arch VM tomorrow and see if I can help you. :)

cfr

12:48 AM

@PauloCereda I would certainly appreciate any help. But don't go to any trouble.

Paulo Cereda

@cfr Don't worry, it's the kind of fight I like to have with operating systems. :)

@cfr: In other news, I feel English: I'm now a member of UK-TUG. :)

cfr

@PauloCereda I didn't realise that the English had a monopoly on it.

Paulo Cereda

@cfr Oh.

In other news, I got a couple of vegetarian/vegan recipes, I'll try some of them during this Carnaval.

cfr

@PauloCereda Cool. What are you cooking? Oh, by the way, KDE now speaks Wenglish.

Paulo Cereda

@cfr I plan aubergine + shoyu. :)

cfr

1:06 AM

@PauloCereda Ooohhhh! Aubergines are sooooo good.

Faheem Mitha

1:46 AM

Does anyone know a standard way to wrap long lines in TeX? I found

4

Q: wrapping inline long lines and adding background color

Objective I'm trying to achieve the following output. Here's the code I used to produce it. I had to hack it \begin{document} \begin{description} \item In \texttt{./build/org.eclipse.cdt.managedbuilder.core/plugin.xml} \\ there is an extension {\btHL \texttt{org.eclipse.cdt.managedbuilde...

but it looks home-made.

OK, maybe this:

14

A: Wrapping long lines that contain no spaces

You may try the seqsplit package.

Funny, doesn't seem to do anything, if I do:

\gdef \ReferencedIDs{,2014.11.14,2014.11.14,2014.11.17,2014.11.14,2014.07.18.tm,2014.11.12.tm,2013.10.05.kanjur,2013.10.05.powai,2013.10.06,2013.10.28.zhm,2013.10.17,2013.10.22,2013.10.14.ks,2013.10.06,2013.10.17,2013.10.22,2013.10.24,2013.10.28,2013.10.30,2013.11.11,2014.09.05.tm,2014.07.18.tm,2014.11.12.tm,2014.11.17,2015.01.06,2015.01.06}

\seqsplit{%
\ReferencedIDs}

If I insert the string explicitly in there it works. Am I missing something?

yo'

tell me, if I ever travel to the US, do I have a chance to convince Don Knuth to let me play his fabulous organ? :)

Faheem Mitha

2:05 AM

@yo' just fyi, fabulous used to mean imaginary, and still should.

It's an unfortunate example of language corruption.

/me in pedant mode.

yo'

@FaheemMitha I could use "mystical" as well, or simply "awesome" :) and it's not corruption. Corruption is a "consists of" in bad places, or "the most optimal". This is simply a shift of meaning, which is necessary.

Faheem Mitha

@yo' why necessary?

yo'

@FaheemMitha because otherwise "mail" wouldn't be "mail" because it's missing the wax seal.

Faheem Mitha

@yo' personally I'm in favor of making new words and leaving the old ones alone. That is also in the spirit of TeX.

And I don't follow your point about "mail".

yo'

@FaheemMitha 99% of the words we use have shifted, broadened or narrowed their meaning some time in the past.

Faheem Mitha

2:13 AM

@yo' citation, please.

Also, this word happens to have a Latin etymology.

yo'

@FaheemMitha I'm not a linguist, sorry. "Mail" is an example of such a word, as I explained before. "Word" is another one (originally meaning "verb"). Etc.

The words shift their meaning every day.

Faheem Mitha

@yo' hopefully not every day

yo'

@FaheemMitha you can only hope :)

Faheem Mitha

@yo' As I said already, I don't get your point about "Mail", but I'm not familiar with the word.

yo'

@FaheemMitha mail means "anything send by the post services", but originally it was a "letter with a proper wax seal", whereas now it's "anything delivered to your inbox, both physical and electronic"

Faheem Mitha

2:21 AM

@yo' "letter with a proper wax seal"? Really? Hmm.

I guess one learns something new every day.

yo'

@FaheemMitha well, there was no other letters back then ;)

Faheem Mitha

@yo' could have been a generic term for something delivered.

Would a parcel have counted as mail in the 19th century?

yo'

@FaheemMitha which it probably was not. Don't take me too seriously. I just wanted to show that a shift in meaning is necessary, because the world changes all the time

Faheem Mitha

Even now people use the term email to distinguish from regular mail.

@yo' I don't know about necessary. And supposedly the language has not changed much since the Johnson directory was published.

yo'

@FaheemMitha maybe, but many new things were invented that need their words -- either you get new words, or you change the meaning of those you have.

Faheem Mitha

2:25 AM

@yo' new words are good.

the words people come up with are so boring though.

Like telephone. Really? Couldn't you do better?

yo'

@FaheemMitha another word that has changed meaning. Now, it means "personal computer"

Faheem Mitha

@yo' :-)

There is the word smartphone.

IMO phone still just means phone. But I take your point.

Trying to go back to sleep now. Talk to you later.

yo'

gnight

Sean Allred

3:32 AM

@ChristianHupfer :)

I just seem to have a fascination with how macro languages work

or really any language

The urge to test it was unstoppable even though I was pretty sure of the outcome XD

yo'

@SeanAllred learn OCaml. That'll cure you from fascination :D

Sean Allred

@yo' Oh I've already played with it :)

yo'

@SeanAllred should I stop speaking to you then? :P Do you know Coq?

Sean Allred

One of the things I marked about it when learning was its basic similarity with macro languages. It seems to have the same idea of an input stream, but with different syntax rules.

@yo' Never heard of that language, actually

Actually yes, I have!

Just not by that name.

yo'

Coq

In computer science, Coq is an interactive theorem prover. It allows the expression of mathematical assertions, mechanically checks proofs of these assertions, helps to find formal proofs, and extracts a certified program from the constructive proof of its formal specification. Coq works within the theory of the calculus of inductive constructions, a derivative of the calculus of constructions. Coq is not an automated theorem prover but includes automatic theorem proving tactics and various decision procedures. The Association for Computing Machinery presented Coquand, Huet, Paulin-Mohring, Barras...

Sean Allred

3:35 AM

I wanted to do something similar for my thesis (I doubled up with math), but my advisor informed me that it was already done :) he never did tell me the name though – thanks for the link :)

@yo' but please don't stop speaking to me :( (Actually, I've got to run off to sleep now. I've logged over sixty hours this week in training… I am le tired.)

Goodnight, all

yo'

@SeanAllred it was a joke :)

I should go to bed, too, it's 4:40am

Christian Hupfer

@yo': Good night ;-) Or good morning :-P

@SeanAllred: ???

karlkoeller

6:55 AM

I think @egreg should earn a gold badge for not having reached the 200 daily rep yesterday

Something like

5

So, @Joseph and the other mods, please invent such a badge. :-)

This because I think he did it intentionally to give @David a chance to win the weekly rep league..........

Christian Hupfer

8:17 AM

@karlkoeller: I think most of us would get this badge then ;-) Is it designed to be awarded multiple times? :-P

karlkoeller

@Christian No, other people can't. It would be designed only for people that once in a year don't get 200+ rep :-) And, yes, it can be awarded multiple times (maximum once in a year)....

Christian Hupfer

@karlkoeller: :-(((((((((((((((((((((((((((((((((((((((.... Ill - designed batch requirements ;-)

Paulo Cereda

9:53 AM

7

Q: Was Pope Francis a Black Sabbath fan in his youth?

Recently I stumbled upon the following image: Could this be real? I'm guessing this is fake, but I am at work and can't do a complete research about it. The reason I find this may be fake is mostly due to this image found on google: But this one could be the fake one as well (although this...

catholic-church photography celebrities pope

@yo' Wow, I know this theorem prover (I also have the book nearby). Quite interesting.

Christian Hupfer

@PauloCereda: Those are the really important questions ;-)

karlkoeller

@PauloCereda I didn't sleep tonight thinking about that...

Paulo Cereda

@Christian, @karlkoeller: Quite shocking, I know. After all, Pope Benedict was a Rammstein fan. :P

Being Argentine, Pope Francis must like tango. :)

Christian Hupfer

@PauloCereda: Nope, rather a rap fan ... using explicit lyrics ;-)

Vrouvrou

hello

how to obtain this with (1.2) in this place

please

Paulo Cereda

10:08 AM

@ChristianHupfer Das Pope!

Christian Hupfer

@Vrouvrou: Ah, the Maxwell equations with matter ;-)

Vrouvrou

what is the command in latex to obtain this please?

Christian Hupfer

@Vrouvrou: I am sure there are already $\infty$ questions with $\infty^2$ answers to that here on TeX.SX!

Paulo Cereda

@ChristianHupfer How about 2^\infty? It should be bigger. :)

Oh wait.

@egreg: How are you feeling today? We miss you, buddy.

Christian Hupfer

@PauloCereda: Boo

Paulo Cereda

10:13 AM

@ChristianHupfer: I speak English very well! I learned it from a book!

@ChristianHupfer boo!

Das Boo! :)

Christian Hupfer

@PauloCereda: I made no mistake I made no mistake la la la la

Paulo Cereda

@ChristianHupfer Neither do I, I never make mistaeks.

^^ Das ist ein Boo!

Christian Hupfer

@PauloCereda: No, except with German articles :-P

Paulo Cereda

:)

@ChristianHupfer Articles are hard!

:)

Christian Hupfer

@PauloCereda: But necessary :-P .... Looks like a marshmellow on dope :-P

Paulo Cereda

10:15 AM

@Christian: besides, @Johannes was suggesting I should kill ducks with Die Ente. :)

@ChristianHupfer LOL

Christian Hupfer

@PauloCereda: I am pretty sure he meant der Erpel ;-)

Paulo Cereda

@ChristianHupfer ooh more duck family!

Christian Hupfer

10

Q: How to format Maxwell equations?

Is there a variant of split environment from amsmath package which would allow for two aligned equations in a row? I want to format two pairs of Maxwell equations in two rows and put single equation number between the rows. Unfortunetely, split allows only one alignment tab & in a row, and I need...

equations amsmath align

@Vrouvrou: See that question, please, for example

David Carlisle

10:35 AM

@yo' I've used that (a long while back)

Joseph Wright

Morning @DavidCarlisle

David Carlisle

@JosephWright morning. Yes I wrote something about AW and marketing in an original version of that answer, but I deleted it:-)

Joseph Wright

@DavidCarlisle Ah

Paulo Cereda

10:57 AM

I'm definitely going to hell for this. ♥ — Paulo Cereda 20 secs ago

karlkoeller

@egreg hope that you haven't got offended by my joke

David Carlisle

@karlkoeller what about me?????

karlkoeller

@David I thought you hadn't even read the joke... And anyway you have all reasons to feel offended :-)

David Carlisle

@PauloCereda I think you'd better send Psmith on holiday for a couple of weeks. We don't want any cricket news.

Paulo Cereda

@DavidCarlisle Uh-oh.

egreg

11:05 AM

@karlkoeller Sorry, what joke?

@PauloCereda Better, but still uncomfortable.

Paulo Cereda

@egreg :(

egreg

@PauloCereda Don't worry, I'll recover soon.

Joseph Wright

MiKTeX working on Win10 :-)

karlkoeller

@egreg

4 hours ago, by karlkoeller

I think @egreg should earn a gold badge for not having reached the 200 daily rep yesterday

David Carlisle

@karlkoeller now you've offended him twice

karlkoeller

11:20 AM

@DavidCarlisle It was my intention to do that...

egreg

@karlkoeller Well, I had something else to do. ;-) Better, somebody had something to do on me.

karlkoeller

@egreg Well, I knew I hadn't to joke with you...

egreg

@karlkoeller Don't worry: something I should have done some years ago. I tried with \xpatchcmd, but it didn't work. ;-)

karlkoeller

11:35 AM

@egreg ??????????????????????

egreg

@karlkoeller I had small surgery, yesterday. :)

karlkoeller

@egreg Oh I didn't know that. Hope everything has gone all right with it. And, no, in these cases \xpatchcmd doesn't work. :-)

egreg

@karlkoeller That's what I discovered. :( Yes, everything seems going fine, I have to go on Monday to verify.

Paulo Cereda

@egreg We have England vs. Italy today and Juve tomorrow. :)

egreg

@PauloCereda Expect jokes from DPC.

Paulo Cereda

11:45 AM

@egreg As usual. :)

karlkoeller

@egreg my best wishes

Paulo Cereda

I don't have ESPN+, so I won't be able to watch rugby. :(

David Carlisle

@egreg so long as you don't mention the cricket world cup

egreg

@DavidCarlisle Something like “Australia won by 111 runs”?

Paulo Cereda

@egreg ooh!

David Carlisle

11:52 AM

@egreg But I'm sure you won't mention that

Paulo Cereda

!!/cricket

David Carlisle

51 mins ago, by David Carlisle

@PauloCereda I think you'd better send Psmith on holiday for a couple of weeks. We don't want any cricket news.

Paulo Cereda

@DavidCarlisle He's not working. :(

!!/help

@PauloCereda Psmith, the TeX bot: Good morning! I'm Psmith, the friendly TeX bot - the p in my name is silent, as in pshrimp. I'm here as a companion to our fellow users in the typographic land. As you probably noticed, I always reply under Paulo's account, but do not despair, I say, my replies are always preceded by my own name. Enjoy your stay at TeX.sx! If you need any help, just ask our chat residents. Cheerio!

!!/cricket

Psmith, the TeX bot, in fixed font mode: Let's take a look at the last cricket results:

- KwaZulu-Natal Inland 213/10 &  220/10  v Boland 62/2 &  187/10 *
- Eastern Province 170/6 * v North West 437/9
- Western Province 374/10  v Border 262/10 &  190/4 *
- Badureliya Sports Club 208/10  v Ragama Cricket Club 191/10 &  124 *
- Colts Cricket Club 245/10  v Nondescripts Cricket Club 338/10 &  127/7 *
- Sri Lanka Air Force Sports Club 15 &  167/10 * v Tamil Union Cricket and Athletic Club 650/8
- Chilaw Marians Cricket Club 226/10  v Sri Lanka Ports Authority Cricket Club 356/10 &  106/2 *

ooh look at this - Australia 342/9 v England 231/10. :)

!!/eightball Will England win today's match?

@PauloCereda Psmith, the TeX bot: The great 8-ball says: If i recall correctly, yes.

clemens

12:50 PM

@PauloCereda this made me think of you: facebook.com/photo.php?fbid=861241380606103 :)

@PauloCereda Oh, Psmith is back :)

Paulo Cereda

@clemens awwww <3

@clemens He's gone, I'm playing Minecraft in my Mac. :)

yo'

@PauloCereda that's a good one, yeah :D

And for the theorem proving: I dunno, the theorem provers have they own world, and a 2-week workshop wasn't enough to explain even the slighest idea of what is co-induction. We all agreed that they are simply unable to define it properly.

@DavidCarlisle oh no!

David Carlisle

1:32 PM

@yo' it was a "competitor" I worked with the Isabelle theorem proving system for a while.

Paulo Cereda

@DavidCarlisle cool, I heard about it.

@yo' It's virtually impossible to grasp those things easily.

yo'

@PauloCereda especially if the people are not willing to say or write: Definition 1.

Paulo Cereda

@yo' :)

yo'

@PauloCereda because that was the real problem

David Carlisle

3:26 PM

@PauloCereda checking the rugby scores?

Joseph Wright

3:56 PM

@DavidCarlisle Have realised that the case changing stuff needs to be f-type expandable: a quick rewrite :-)

@DavidCarlisle All sensible on etex: we just need the green light

Paulo Cereda

4:13 PM

@DavidCarlisle No one is broadcasting. :(

David Carlisle

Any mathematicians here?, what is 47 - 17 ?

egreg

@DavidCarlisle At least we scored three tries.

Paulo Cereda

No one is broadcasting!

egreg

@PauloCereda 47–17

Paulo Cereda

@egreg Thanks! :) Still confusing though. :)

David Carlisle

4:28 PM

@JosephWright isn't it? (I assume not given your comment:-)

yo'

@DavidCarlisle 30, with the right value of 0.

egreg

@DavidCarlisle :P

David Carlisle

4:44 PM

@egreg It seems it's a bit tricky for you today (I know you are not 100%) but @yo' stepped in and worked it out.

yo'

@DavidCarlisle but don't forget to verify the hypothesis

egreg

@DavidCarlisle Did you see the match or were you watching the cricket?

Joseph Wright

@DavidCarlisle Not at present: Bruno felt that wasn't needed (there is a performance hit), but that then means nested use is not as-expected (or indeed currently working!)

@DavidCarlisle Once I've altered I'll drop the older \tl_expandable_... entirely: I'll need to talk to a few package authors

David Carlisle

@egreg neither, I was grepping my tex tree for etex package use :(

David Carlisle

5:33 PM

@JosephWright l3build support required:-)

egreg

6:39 PM

Back to normal, at least with respect to rep cap. ;)

Paulo Cereda

@egreg Yay! But I reached vote cap. :)

egreg

@PauloCereda Where's the problem? ;-)

Paulo Cereda

@egreg I can only upvote you tomorrow. :)

yo'

@egreg sooooo late :p what's wrong with you :p

egreg

@yo' I'll recover soon, don't worry.

Faheem Mitha

7:17 PM

So, back here asking basic questions. In tex.stackexchange.com/a/228063/3406 I don't see why two \expandafters are necessary. This expands the token after the next one, right?

So \expandafter\oldseqsplit should suffice.

yo'

@FaheemMitha that would expand \oldseqsplit after expanding {

Faheem Mitha

@yo' expanding {? That's a token?

yo'

@FaheemMitha and what else would it be?

Faheem Mitha

@yo' Dunno. A bracket? I thought only macros got expanded.

yo'

@FaheemMitha well, the expansion of { is {, but in order to skip it, you need to use \expandafter on it of course

Faheem Mitha

7:20 PM

My god, egreg didn't hit rep cap yesterday. The foundations of the universe are in danger of crumbling.

I thought the way it works was TeX goes from left to right, and when it sees an \expandafter it expands the one after the next, which would also be \expandafter.

yo'

@FaheemMitha yes

Faheem Mitha

And so it does nothing. And when it hits the second \expandafter, that one tells it to expand the argument, skipping the {?

yo'

The expansion of \xp<token1><token2> (shortcut of \expandafter) is that <token1> is skipped for a while, then <token2> is fully expanded, and then <token1> is expanded

so when <token2> is another \xp, <token3> gets skipped and <token4> expanded.

Faheem Mitha

I somehow got the impression all those { and } were just syntactic sugar.

yo'

@FaheemMitha nothing that survives the reading stage is a syntactic sugar

Faheem Mitha

7:26 PM

@yo' reading stage of TeX?

yo'

\xp really works on the token basis, and the only multi-character tokens are control sequences

@FaheemMitha the one when catcodes are assigned, % skipped, ends of lines changed into \par etc.

Faheem Mitha

@yo' sorry, I don't know what \xp is.

@yo' ok

yo'

@FaheemMitha I use it to say \expandafter because \expandafter is just way too long

Faheem Mitha

@yo' oh.

@yo' so is my summary above correct? And one is left with \oldseqsplit and the expanded \ReferencedIDs?

yo'

@FaheemMitha yep, exactly

Faheem Mitha

7:30 PM

@yo' ok.

So the curly brackets are functional syntax? Sometime people omit them, sometimes for no obvious reason.

yo'

@FaheemMitha that's more complicated (and beyond what I can explain I'm afraid)

Faheem Mitha

@yo' ok

So, is everything in TeX a token?

yo'

@FaheemMitha that's like saying that everything in a novel is a letter. It's true (to some extent), but what does it help with?

Paulo Cereda

@FaheemMitha Take a look at Chapter 5 of the TeXbook (grouping).

David Carlisle

@FaheemMitha anything you can store in a macro is a token, below that are boxes glue lists

Faheem Mitha

7:36 PM

@PauloCereda ok

yo'

@DavidCarlisle or above :-)

Paulo Cereda

@yo' in the middle? :)

Faheem Mitha

@DavidCarlisle "boxes glue lists". are these separate items? :-)

David Carlisle

@yo' only if you are standing on your head:-)

yo'

@DavidCarlisle :)

David Carlisle

7:40 PM

@FaheemMitha you can't really explain an entire processing model in disjointed chat lines, but tokens are the things you process with \newcommand, #1 \expandafter etc, usually after all tokens etc are expanded things get put on to horizontal or vertical lists, and at that level tex deals with liists of character nodes, boxes, glue. these are much harder to manipulate in tex, in some cases you can remove the last item of a list with `\unskip or \lastbox etc but some items are not removable

Faheem Mitha

@DavidCarlisle to a first approximation, you need to worry about the expansion or nonexpansion of tokens as a means of evaluation. is that a fair statement?

And the other stuff sort of "comes around for the ride".

David Carlisle

@FaheemMitha not sure if it's fair or not, you need to worry about expansion but most evaluation, eg incrementing counters or lengths is not done via expansion

Paulo Cereda

@DavidCarlisle: There's a new Bond movie in the making. Will you apply for playing Q? :)

David Carlisle

@PauloCereda No, real secret agents may not apply for fear of revealing their identity

oops

Paulo Cereda

@DavidCarlisle oh no!

But you can get a Jaguar! :)

David Carlisle

7:54 PM

@PauloCereda with a chauffeur:

Paulo Cereda

@DavidCarlisle ooooooh! :)

David Carlisle

@PauloCereda internet is weird enough that I had faith that google for "duck driving jaguar" would turn up something

Paulo Cereda

@DavidCarlisle LOL

Faheem Mitha

Ok, then why is the first \expandafter needed? What's wrong with \oldseqsplit\expandafter? That \expandafter will expand the argument, which is all that is needed anyway.

David Carlisle

@FaheemMitha first expandafter where?

Faheem Mitha

7:59 PM

3

A: seqsplit does not work unless the string is given as an argument directly

The argument to \seqsplit is processed token-by-token, assuming that it's fully expanded. This is not the case when you pass a macro (a single token) as the argument. The following setup redefines \seqsplit to expand its argument before processing it. \documentclass{article} \usepackage{seqspl...

@DavidCarlisle that usage has \expandafter\oldseqsplit\expandafter.

David Carlisle

@FaheemMitha oh well \oldseqsplit\expandafter the argument to oldseqsplit would be \expandafter which wouldn't make sense at all

@FaheemMitha why do you say it would expand the argument? In that case it is the argument.

Faheem Mitha

@DavidCarlisle oh

David Carlisle

@FaheemMitha yes because you want to expand the token after the { before expanding \oldseqsplit so you have to chain \expandafter to make that happen.

Faheem Mitha

So if one has \expandafter\oldseqsplit\expandafter{arg}, then the order is (1) first \expandafter tries to expand the second \expandafter which does nothing. Does the second \expandafter then expand the argument? It doesn't go back to \oldseqsplit?

David Carlisle

@FaheemMitha the second expandafter doesn't do nothing it expands the token after {

Faheem Mitha

8:06 PM

@DavidCarlisle No, I meant: first \expandafter tries to expand the second \expandafter. This expansion has no effect on the second \expandafter. That is, \expandafter expanded is still just \expandafter.

My question is about what happens in what order.

Paulo Cereda

@FaheemMitha Left to right?

:)

Faheem Mitha

Let's use \ea to for \expandafter. So

1) 1st \ea tries to expand 2nd \ea. with no result.

Is then control handed to the second \ea, which then expands the argument?

And at what point does control go back to \oldseqsplit?

yo'

@FaheemMitha the assumption 1) is completely wrong

Faheem Mitha

@yo' oh?

Paulo Cereda

@yo' tihs.

Thsi.

Tish.

Faheem Mitha

8:10 PM

I thought that is what \eadoes.

Paulo Cereda

Stih.

Faheem Mitha

@PauloCereda ??!!

yo'

what you call "tries to expand 2nd \ea" has a big result, namely: the 2nd \ea is expanded.

Faheem Mitha

expanded to what?

isn't it already what it is supposed to be?

yo'

@FaheemMitha expanded to: skip one token, expand the second one

Faheem Mitha

8:11 PM

@yo' oh. I thought expansions expanded to other tokens.

yo'

@FaheemMitha expansions expand to whatever is the expansion of the thing

Paulo Cereda

@FaheemMitha What is an expansion?

Faheem Mitha

@yo' hmm

@PauloCereda Dunno. What happens to your stomach when you eat too much?

yo'

@FaheemMitha when you drink milk over cherries :)

Paulo Cereda

@FaheemMitha If you use this analogy, I can tell you got the expansion concept totally wrong, sorry.

Faheem Mitha

8:15 PM

I thought macro expansion was separate from execution.

yo'

@FaheemMitha it is, see this example:

\def\gobble#1{}

\def\abc{abc}

\edef\testone{X\expandafter\gobble\abc Y}
\def\testtwo{X\expandafter\gobble\abc Y}

\show\testone
\show\testtwo

Faheem Mitha

I.e. one expands out the macro till you can't expand it any more, and then it is executed.

yo'

it will show:

> \testone=macro:
->XbcY.
l.8 \show\testone

?
> \testtwo=macro:
->X\expandafter \gobble \abc Y.
l.9 \show\testtwo

Paulo Cereda

@Faheem: it's the transformation of an occurrence of a lexical/syntatic pattern into the resulting sequence.

yo'

i.e., \expandafter was expanded in the \edef

Faheem Mitha

8:17 PM

But skip next token and expand the one after is a not a sequence, it is an action

yo'

@FaheemMitha whatever it is, it's done during the expansion phase, as you can see from my example, and that's the important thing

Faheem Mitha

I don't understand the syntax of \def\gobble#1{}. Does it mean - take one argument and produce nothing? So first argument disappears?

yo'

@FaheemMitha yep

so \gobble\abc would be nothing, whereas \expandafter\gobble\abc expands \abc and then puts away the first group or token

sorry guys, gotta go shopping or they close in 40 minutes

Faheem Mitha

@yo' ok, thanks.

\edef expands everything immediately.

does not wait till it is called.

In \edef\testone{X\expandafter\gobble\abc Y} still don't undrestand. So \ea jumps to \abc. That expands to abc. Fine.

Then it goes back to \gobble? Why?

Hm, if en.wikibooks.org/wiki/TeX/expandafter is correct, then

\ea is still in control here. first expands \abc, then \gobble.

Presumably if any of those in turn contained macros, they would be executed before moving on.

egreg

8:35 PM

@FaheemMitha You should note that \expandafter disappears after having performed its duty. Let's say we have \def\x{yz}. Then \ea\foo\ea{\x}` will expand \x and make TeX into seeing \foo{yz}

Faheem Mitha

@egreg yes, I see. I still don't see how \ea expands into skipping a token and expanding the one after that. it is surely an action rather than an expansion. like printing a letter.

egreg

@FaheemMitha No, it's pure expansion.

Faheem Mitha

@egreg is printing a token also expansion? if not, why are the two different?

egreg

@FaheemMitha No it's not expansion. The two are performed at different levels.

Faheem Mitha

@egreg oh

This article looks useful:

tug.org/TUGboat/Articles/tb09-1/tb20bechtolsheim.pdf

egreg

8:43 PM

@FaheemMitha An unexpandable token is sent at a different level, where unexpandable tokens are executed. For instance, \relax is not expandable; its execution produces nothing, but it's another matter.

Faheem Mitha

@egreg ok

David Carlisle

@FaheemMitha maybe someone answered while I was out, but in case not "That is, \expandafter expanded is still just \expandafter." No \expandafter expanded isn't itself

egreg

@FaheemMitha The different levels are explained well in TeX by Topic

Faheem Mitha

@DavidCarlisle I was told \expandafter expanded is skip the next token, expand the one after that.

but i'm having difficulty processing that.

egreg

@FaheemMitha Yes, but as soon as it triggers that expansion it disappears from the main input stream.

Faheem Mitha

8:45 PM

@egreg yes, I see that.

David Carlisle

@FaheemMitha given that I've left had a meal cleared up had coffee and conversation is at same point I think I was right with this:-)

1 hour ago, by David Carlisle

@FaheemMitha you can't really explain an entire processing model in disjointed chat lines, but tokens are the things you process with \newcommand, #1 \expandafter etc, usually after all tokens etc are expanded things get put on to horizontal or vertical lists, and at that level tex deals with liists of character nodes, boxes, glue. these are much harder to manipulate in tex, in some cases you can remove the last item of a list with `\unskip or \lastbox etc but some items are not removable

Of course @egreg, being Italian, is scratching his head trying to work out how it's possible to eat a meal in an hour

yo'

@DavidCarlisle one our? not enough even for the starter

Faheem Mitha

So TeX doesn't expand left to itself? It need to be told to do so by \expandafter?

Why isn't there a \expandnext?

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends