TeX, LaTeX and Friends

2:28 AM

I am having an error with bibtex. bibtex while reading the top level .aux file (created by compiling main.tex) says that it cannot find any \citation commands. That's because the commands are in the second level .aux files (e.g. chapter1.aux, chapter2.aux). How to make bibtex see those \citations commands? As a result of this I don't have any references in the final .pdf.

Raphael

6:42 AM

Some of you may be interested in a new development on the "LaTeX maths in the browser" front:

2

0

Q: A contender for MathJax?

Khan Academy released KaTeX, a library that does essentially the same job as MathJax does today -- only faster. Part of that is probably because (the hard?) part of the layout can be preprocessed on the server. This may also enable mathematics typesetting in apps more easily. As of now, it does ...

Maybe they have use for one or the other interested LaTeX expert?

tohecz

7:22 AM

leaving home and heading for the exam. Cross your fingers please!

David Carlisle

@tohecz good luck, have fun:-)

tohecz

@DavidCarlisle you have a very sarcastic idea of fun, obviously :D

but thanks, anyways!

David Carlisle

@tohecz If in doubt add an extra \expandafter that usually works

3

tohecz

@DavidCarlisle hey, this is not fair! :D

well, I go, bye!

David Carlisle

bye

wipet

7:26 AM

hello David

Is the 50 lines similar to your guess about thousands lines of code?

Joseph Wright

@wipet I susepct @DavidCarlisle, like me, had in mind a more conceptually simple but token-expensive approach

@Raphael More off-topic questions for us :-)

@Raphael More seriously, thanks for the heads-up

wipet

@JosephWright I don't understand. Can you more specify this mind?

Is here (in 50-line solution) something what is doesn't cover the original question?

Joseph Wright

@wipet I can only speak for myself, but my approach would be to define a series of collector macros for each letter of each word that might appear. Thus \foo@collect@first@f, \foo@collect@first@fi, \foo@collect@first@fir and so on for a branch which might end up finding #1 followed by first (each a delimited macro including a \futurelet to find the next token and branch as appropriate).

This needs one auxiliary per char in the possible delimiter plus associated support to handle spaces and so on, so gets very expensive very quickly but is easy to follow conceptually.

Of course, for lots of possibilities I'd define a generator system, but it would still be expensive

@wipet Not a criticism of your approach: I was just commenting on how I'd tackle the same issue. I usually leave this sort of tricky thing to Bruno Le Floch!

wipet

7:50 AM

@JosephWright Maybe I can write a little article about this code with step-by-step description. But I am not sure if somebody interest in this.

Joseph Wright

@wipet I can imagine more efficient ways than my own: I just wouldn't start with them :-) (For example, as it's non-expandable anyway we don't need multiple auxiliaries, as a single one can be re-purposed.)

@wipet Karl is always after articles for TUGBoat

David Carlisle

8:21 AM

@wipet I haven't looked yet but I thought that question would interest you:-)

David Carlisle

8:32 AM

Phew finally I can answer a question, I didn't answer any yesterday as I didn't want an accept to throw me out of parity: