Conversation started Feb 20, 2014 at 0:51.
tohecz
tohecz
Feb 20, 2014 00:51
@DavidCarlisle lol
btw, what is the current maintanance status of tex4ht/htlatex?
David Carlisle
David Carlisle
@tohecz river valley are looking after it as far as I know
tohecz
tohecz
@DavidCarlisle whom?
David Carlisle
David Carlisle
@tohecz river-valley.com
tohecz
tohecz
I would seriously need to de-bug its MathML generator. Because otherwise I'm quite content with the output, but the MathML is not valid (missing closing tags)
David Carlisle
David Carlisle
@tohecz it might be valid in html, even so (but it would be better with them)
tohecz
tohecz
Feb 20, 2014 00:55
@DavidCarlisle not the case, they're really missing
David Carlisle
David Carlisle
@tohecz sure but it may still be valid if you use it in an html page
tohecz
tohecz
and btw, it's incompatible with \DeclarePairedDelimiter, but that's likely a solvable issue
Speravir
Speravir
@tohecz Scroll back here today Michal and Raphael (Raphink) wrote a little bit. Start here: chat.stackexchange.com/transcript/41?m=13856493
tohecz
tohecz
@DavidCarlisle not the case truly. HTML viewing works to some limited extent, not so quite the MathJaX (which gives a "nice" output, but fails to proceed some parts because ov invalid input, and XML parser confirms there're mistakes)
David Carlisle
David Carlisle
@tohecz we are talking past each other:-) If you use an xml parser <math><mi>x</math> is not well formed and a fatal error, if you use an html parser (ie serve the document as text/html rather than application/xhtml+xml, then not only is it not not well formed, it is valid mathml.
tohecz
tohecz
Feb 20, 2014 00:58
@Speravir pinned, thanks :)
@DavidCarlisle ah ok
but then it should not claim it's XML: <math xmlns="http://www.w3.org/1998/Math/MathML" display="block" class="equation">
David Carlisle
David Carlisle
@tohecz If you are using mathjax anyway I would configure tex4ht not to convert the TeX and just to let it pass straight through and let mathjax do that conversion.
tohecz
tohecz
anyways, I need a valid XML input obviously :)
@DavidCarlisle is that possible? And what about user-defined macros, MathJaX-unaware packages etc. ? (I'm just asking, we have not set up the solution for math yet)
David Carlisle
David Carlisle
@tohecz That doesn't claim anything (and it is conforming HTML5 syntax, which happens to overlap with XML syntax) it is the served mime type that matters not what the tags look like
tohecz
tohecz
damn ok. I'll need to get into this. However, I know that I would prefer to have a valid XML structure since that's what we know how to deal with in general
David Carlisle
David Carlisle
@tohecz on the grounds that not touching it and copying it straight to the output must be easier than generating mathml it must be possible. (configuring tex4ht is a black art though:-)
@tohecz yes so would I, but I was just reporting how things are:-)
@tohecz well for any packages it's essentially the same you may have to set up a tex4ht configuration or a mathjax one if it is not already there.
tohecz
tohecz
Feb 20, 2014 01:04
@DavidCarlisle now I lost you :) You mean that tex4ht should be possible to output a low-level tex code of anything?
@DavidCarlisle On the other hand, e.g. tensor package just converts everything into a sequence of \vphantoms, math tokens and su[bp]scripts. But you can't just output \tensor{x^i^j_i^k_m}, since MathJaX will never understand it
David Carlisle
David Carlisle
@tohecz I mean it must (in principle) to tell tex4ht to treat math as a verbatim environment and just pass it through to the output. (In fact I think I had this conversation with @nasser a while back and he got something working)
tohecz
tohecz
@DavidCarlisle @Nasser well, we're somehow in the project in the phase where this is crucial, so I'm interested in any good ideas
David Carlisle
David Carlisle
@tohecz If a package is defined in terms of simple \def or \newcommand you can give the tex code to mathjax and it will interpret the definitions, but if it involves low level box measuring then you have to write javascript as mathjax tex emulation doesn't go that far. (tex4ht of course really uses tex so things are different there)
@tohecz Fred's math ebook project is interesting (and he's doing most of the support for mathml in firefox and webkit at present) ulule.com/mathematics-ebooks
tohecz
tohecz
@DavidCarlisle well, simple \@ifstar fails in MathJaX I think
Anyways, I gotta go to bed. I get up in 6 hours and have a long day to come
David Carlisle
David Carlisle
Feb 20, 2014 01:21
@tohecz you've spent too long here: commands with @ in their name are not simple by definition:-)
tohecz
tohecz
@DavidCarlisle yep, still used a lot by packages. I was thinking: which TeX primitives is MathJaX accepting? If I just "relaxed" all of them, did the full expansion of math, and collected whatever is left...
David Carlisle
David Carlisle
Feb 20, 2014 01:57
@tohecz there is a list...
@tohecz docs.mathjax.org/en/latest/tex.html#supported-latex-commands
N3buchadnezzar
N3buchadnezzar
Feb 20, 2014 02:12
Opposite of slack (regarding a string) ?
Nasser
Nasser
@DavidCarlisle @tohecz I got red marker with message. if you want to know how to pass verbatim latex via tex4ht I can show you how. Is this what is being asked?
Speravir
Speravir
@N3buchadnezzar Perhaps "tight".
Nasser
Nasser
@tohecz Here is the link, tug.org/pipermail/tex4ht/2013q4/000887.html "Is there a way to pass the Latex code directly to the HTML page for MathJax use?" if I find more docs, will let you know.
David Carlisle
David Carlisle
@Nasser yes thanks @tohecz has sensibly gone to bed, if you look up the transcript above you will see we were discussing tex4ht and mathjax and I had a sudden feeling of deja vue and that I'd had the conversation with you at some point:-)
Speravir
Speravir
@N3buchadnezzar Or "taut". Everything according to some dictionaries.
Nasser
Nasser
Feb 20, 2014 02:28
@DavidCarlisle @tohecz Yes, I was searching for it. I thought I documented it. Yes, you did help me on this. But this is the final scoop: One can PASS direct latex to mathjax, but only inline math ! not display math. I am sure of this.
David Carlisle
David Carlisle
@N3buchadnezzar taut or tight in common usage, perhaps "under tension" if you are discussing the science of it
Nasser
Nasser
Feb 20, 2014 02:42
@DavidCarlisle @tohecz I found it. I did document it here : 12000.org/my_notes/faq/LATEX/html_and_latex.htm (just search for the string "second method above is better" on the page, and you'll find how to do it. I wrote: "The current issue with this method though, is that there is no way to tell tex4ht to pass the display math latex code to the HTML page. Only in-line latex code can be passed as is to the HTML page" then there is an example how to do it.
it is important to use the shown tex4ht .cfg file to configure things. But again, this only works for inline latex. so not useful really.
Gonzalo Medina
Gonzalo Medina
Feb 20, 2014 03:40
@JosephWright This might be on interest for you: tex.stackexchange.com/q/161385/3954
 
1 hour later…
Sean Allred
Sean Allred
Feb 20, 2014 04:49
user image
From the arara manual. This made me laugh much more than it should have (esp. ten minutes before midnight with sleeping flatmates!)
(Also, +1 to the arara fan-club :))
Ironically, there are no instructions for Emacs.
 
3 hours later…
michal.h21
michal.h21
Feb 20, 2014 08:04
@tohecz tex4ht conversion to mathml is sensitive on some constructs, see tug.org/applications/tex4ht/mml.html for common problems, maybe it could help
@Nasser there is possibility also for display math: tex.stackexchange.com/a/85388/2891
but there is also problem with math environments, they must be redefined to collect its contents and output it with \unexpanded or token lists
Joseph Wright
Joseph Wright
Feb 20, 2014 08:50
@GonzaloMedina Thanks: I think I see the issue
 
Conversation ended Feb 20, 2014 at 8:50.