@FaheemMitha no quite the opposite, the normal mode of operation is to expand until a non expandable token is head of the list

@DavidCarlisle oh. so \expandafter just changes the order?

tex by topic seems helpful. More readable than the texbook, at any rate.

@FaheemMitha yes if you go \newcommand\zzzz{hello}....\zzz then when tex gets to \zzz that is expandable so it expands to hello and then tex starts at the new head of the list and sees h that isn't expandable it's a character token so it adds the corresponding character node to the current horizontal list. You don't explicitly call for \zzz to be expanded

@DavidCarlisle yes, I see.

@DavidCarlisle An hour? A comfortable meal requires at least three hours.

@egreg and at least 3 such meals a day

@DavidCarlisle Have been out (mum's birthday): will take a look

@JosephWright thanks (and happy birthday)

So, the argument to a \def macro needs to be expanded before it is passed to the macro?

@FaheemMitha no

"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. "

@FaheemMitha actually I'm not sure what you mean by that (but the answer is no anyway) do you mean the argument to a macro defined by \def or the argument to def itself (\def isn't a macro)

@FaheemMitha that is a description of \seqsplit not of macros in general

@DavidCarlisle argument to a macro defined by def.

As in the quote above.

Is \seqsplit special in this respect?

@FaheemMitha every macro does different things.

The issue is that the argument to \seqsplit, namely \ReferencedIDs does not get expanded automatically. Enter \expandafter and all the ensuing wackiness.

@DavidCarlisle how annoying of them

@FaheemMitha consider \def\zz{ab} then you can use \fbox\zz or \fbox{\zz} or \fbox{ab} and they are all the same as \fbox treats its argument as a unit and doesn't care if it is expandable or not.

If \seqsplit expanded \ReferencedIDs there would be nothing special to do.

but now consider \def\zbox{\@gobble#1}

@DavidCarlisle oh

Does \@gobble eat the first argument?

I mean the first letter, I guess.

now \zbox{\zz} and \zbox{ab} are completely different. In the first case you get \@gobble\zz so nothing, in the second case you get \@gobble ab so a is gobbled and you get b.

@FaheemMitha Now if you have \zz but want to get the second behaviour you need \expandafter\zzbox\expandafter{\zz} which is exactly the idiom this conversaion started with

@FaheemMitha \@gobble is defined by \def\@gobble#1{}

@DavidCarlisle right

@FaheemMitha so if zzbox was defined a "losing the first letter" but someone asked why \zzbox{\zz} did not work, the answer would be "The argument to \zzbox is processed token-by-token, assuming that it's fully expanded." But as you see that doesn't apply to \fbox it is a description of the way #1 is used in this particular macro

Hang on, what is \fbox above?

@FaheemMitha standard latex command (randomly chosen)

@DavidCarlisle oh

@FaheemMitha It was the example I used 4 lines up

"In the first case you get \@gobble\zz so nothing". Not sure what that means. So \@gobble eats the first token it sees, which in this case is \zz?

Which means the whole thing disappears?

@FaheemMitha yes

Ok

So, this behavior in the case of \seqsplit, is it a defect?

Or to put it differently, would it be better if it could take macros as arguments?

@FaheemMitha You have to make a choice

@FaheemMitha It can take macros as arguments but it splits the sequence you give it, it doesn't split the sequence that that expands to. It's more general that way as you can always expand the argument first if need be, but if you have a command that internally forces expansion and you want it to work on a list of unexpanded tokens, it's harder

@DavidCarlisle Breakfast is shorter, two hours suffice.

@DavidCarlisle I see.

Yes, that makes sense.

@egreg A working breakfast?

@FaheemMitha Is your aim to remove the first comma you get in the expansion of \ReferencesID?

@FaheemMitha that distinction is inherent in TeX, not just in macros, consider \def\z{abc} then \uppercase{\z} is \z which expands to abc which isn't what people want, you could do \uppercase\expandafter{\z} to get ABC. \uppercase acts on the tokens supplied, it doesn't expand them first

@egreg Not specially.

@FaheemMitha Italians don't work and eat at the same time, what an insult!

I don't know. Does it get in the way of something?

@DavidCarlisle :)

@DavidCarlisle Working on that :-)

@JosephWright I know, but I didn't want to go there:-)

@DavidCarlisle I assume you mean \expandafter\uppercase\expandafter{\z}

@FaheemMitha No:-)

@DavidCarlisle sigh

@FaheemMitha I nearly added the redundant first \expandafter so as not to confuse, but @egreg and @JosephWright are watching and they would have jumped on me:-)

@DavidCarlisle Indeed

@DavidCarlisle well, that's very thoughtful of them.

@FaheemMitha This is tricky: \expandafter\uppercase\expandafter{\x} has the same effect as \uppercase\expandafter{\x} because of how primitives having as argument a <general text> work. They expand tokens until finding { and discarding space tokens and \relax as they go.

\uppercase isn't a macro with an argument defined by #1 it is a tex primitive and so it has its own parse rules.

@DavidCarlisle oh. my bad.

A tex primitive is an atom of the language, so to speak?

@FaheemMitha yes defined in tex-the-program rather than defined within tex with \def

@FaheemMitha As a result, \uppercase isn't sent down to the execution processor before \expandafter has done its job.

@DavidCarlisle ok

@egreg um

I assume this is that stages of tex people talk about. the stomach and so forth

@FaheemMitha \uppercase and \lowercase are very tricky.

I'm sure reading tex by topic would be helpful. I wish I had a paper copy. I see it is free now.

@FaheemMitha Yes, the stomach is what I was referring to.

@egreg ok.

I can try to read it on the computer, I guess. Is tex by topic as good as it gets? from an explaining tex pov?

@FaheemMitha It's excellent and available in print

@FaheemMitha I do have one! lehmanns.de/shop/mathematik-informatik/…

@JosephWright yes, but probably not so easy in india.

@egreg I got a free one as a DANTE member :-)

There might be some print to order outfits in India, but I don't know if any off-hand.

@FaheemMitha Order from Germany?

@JosephWright hmm

@FaheemMitha If that's not an option, make a hardship case and ask UK-TUG, DANTE or TUG (etc.) to support sending you a copy?

@JosephWright Um, making a hardship case would be hard. :-)

but thanks for the suggestion.

@JosephWright it's definitely an option. but would be relatively pricey. depends on shipping costs.

@FaheemMitha Hardship of availability?

@FaheemMitha How much can it cost to send to India? Surely not more than about £20

@JosephWright I don't know if you could call that hardship.

@FaheemMitha You can if we want people to have access to TeX resources

@JosephWright Dunno. I guess I could check Amazon UK too.

Hold on, the german thing is in english, right?

Searching for this keeps bringing up

dickimaw-books.com/latex/novices/index.html

for some reason

@FaheemMitha Have you looked at lulu.com/shop/victor-eijkhout/tex-by-topic/paperback/…?

@TorbjørnT. no, I hadn't seen that. I'll check their shipping rates

Thanks

Is Victor around here?

@JosephWright I got one as a member of GuIT. :)

@FaheemMitha Not very active, however.

Oh, he has an account. Has he been seen here?

Hmm, bitbucket repos.

bitbucket.org/VictorEijkhout/tex-by-topic

@FaheemMitha He gave one answer and disappeared; he last looked at the site last October.

@egreg yes, so I see. That's a shame.

Ok, I'm impressed. The book built from the repos, without fuss, in like 3 seconds flat.

@FaheemMitha according to the BitBucket page, his affiliation is Texas Advanced Computing Center, The University of Texas at Austin.

@PaulGessler Sure. I meant where did he learn his stuff? Auto-didact, perhaps. There doesn't seem to be a TeX school.

But learning the language at that level of detail cannot be easy. I surprised he isn't even a bit interested in spending time here.

@FaheemMitha you only have to read the texbook, then write it out in a less eccentric order.

@DavidCarlisle I think there is a little bit more than that to it.

@FaheemMitha there is art to writing a good text book sure, but the first part is true, you only need to read the texbook, you don't need any special place to learn tex

@DavidCarlisle does the texbook contain everything about tex?

@FaheemMitha if you squint and re-read it and then check what tex does, and then re-read it interpreting the words carefully, then yes, it is very hard to find any behaviour of tex that is not explained in the texbook. Of course there is more to a a tex distribution eg fonts and things, that is different, but the behaviour of tex is almost all in the texbook (and all in tex-the-program) (which like the texbook is available as a paid for book but the source (tex.web) is freely available

@DavidCarlisle Ok.

So what victor covers is all contained in the texbook?

@FaheemMitha Not read it all, just flicked over it from time to time especially if I want to give someone a pointer to a free documentation resource. (it wasn't free originally it was a published book)

@DavidCarlisle ok. anyway, it looks good. and readable. I think I would have a hard time following Knuths book. Though I have not really tried.

Though it has been a long time since I tried reading a computing book. But with TeX I might have to try.

@FaheemMitha it depends what you mean by "covers" and "is contained in" you could answer every question on this site by "read the texbook and the package source" but it wouldn't necessarily make the site more useful, but there is a sense in which that is always the answer.

@DavidCarlisle yes, I see what you mean.

@egreg Good morning. How are you?

@JosephWright Happy birthday!

@egreg: how are you feeling now?

@PauloCereda Better than yesterday, thanks.

@cfr: wishing I could +1 again after you added tlmgr --help :-)

@PaulGessler but I can maybe, where? :)

@egreg <3

@PauloCereda kmlinux.fjfi.cvut.cz/~hejdato1/temp/OKKO.py try hitting 3 on the numpad when you're asked for the number of players :)

@yo' Cool!

@PauloCereda I'd more say: CRAZY!

@yo' :D

but good for testing purposes

@egreg Not taking rest. Aren't you? :-)

@HarishKumar This is taking rest. :)

@egreg I doubted you would say that :-)

Editing tags can find answers!

@lockstep rocks.

@PaulGessler It seemed to call for teaching somebody to fish.... (Only metaphorically, of course.)

anyone with good datatool and hyperref foo who isn't about to go to bed, want to take over:-) tex.stackexchange.com/a/228173/1090