Re: [O] ANN/RFC: ox-epub

2017-04-18 Thread Mark Meyer
I will look into Image Export.

Cheers, Mark

On Tue, Apr 18, 2017, at 12:47, Rasmus wrote:
> Hi Mark, 
> 
> Mark Meyer <m...@ofosos.org> writes:
> 
> > I'll keep looking on how to best implement [math support].
> 
> Probably the default solution should be images.  Do you think svg images
> would be an option?  From here it seems it might generally work for at
> least epub-3.
> 
> http://epubtest.org/testsuite/epub3/feature/image-embedding/
> 
> We can produce svg images with dvisvgm already, cf.
> 
> http://orgmode.org/org.html#Math-formatting-in-HTML-export
> 
> Is it possible in epub to specify fallback rendering options like in
> HTML?
> If so it might be possible to put in both an MathML rendering and an svg
> image.
> 
> Thanks,
> Rasmus
> 
> -- 
> And when I’m finished thinking, I have to die a lot
> 
> 


-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] ANN/RFC: ox-epub

2017-04-17 Thread Mark Meyer
>>>>> "Eric" == Eric S Fraga <e.fr...@ucl.ac.uk> writes:

Hi Eric,

Eric> this sounds potentially very useful (to me, at least).  Will
Eric> it export equations and tables, even if as images?  I guess I
Eric> could simply try it... which I'll aim to do this week.

epubtest.org is something aking to an exhaustive test of feature
compliance of reading
systems. http://epubtest.org/testsuite/epub3/feature/mathml/ details the
MathML support for these devices. It looks kind of discouraging.

Now most readers support some kind of JavaScript, but that support is
not clearly specified in the standard. So using MathJax may be possible
on some devices, but not on others.

I'll keep looking on how to best implement this.

Cheers, Mark

-- 
   Mark Meyer
   m...@ofosos.org



Re: [O] ANN/RFC: ox-epub

2017-04-17 Thread Mark Meyer
Hi Eric,

>>>>> "Eric" == Eric S Fraga <e.fr...@ucl.ac.uk> writes:

Eric> this sounds potentially very useful (to me, at least).  Will
Eric> it export equations and tables, even if as images?  I guess I
Eric> could simply try it... which I'll aim to do this week.

Tables should be no problem. Equations I'll say are not supported. I'd
assume you want MathML support, this was delivered with EPUB 3.0.1,
while we're currently targeting EPUB 2.0.1.

Let me look into that.

Cheers, Mark

-- 
   Mark Meyer
   m...@ofosos.org



[O] ANN/RFC: ox-epub

2017-04-16 Thread Mark Meyer
Hi,
this is to announce that ox-epub is now in a somewhat working state and
pretty much aligned with the rest of the ox framework.

https://github.com/ofosos/ox-epub

The current release is 0.2.1 which is right now syncing to MELPA stable.

If you want to generate EPUB directly from org mode, please give it a
spin. The output should always be reasonable and readable in an ebook
reader, but I'm also interested in a more formal way of checking
correctsness. So if you're at it, you may try
https://github.com/IDPF/epubcheck which will give you a detailed listing
of problems with the generated epub.

If you or epubcheck find any reason to complain, please don't hesitate
to open a bug on github, or by directly mailing to me. Please attach any
relevant data, in general the output of epubcheck might be sufficient. I
will try to solve any issues without requesting your data.

Hope you like it.

Cheers, Mark

-- 
   Mark Meyer
   m...@ofosos.org



[O] Using org-export-get-reference for strings

2017-03-30 Thread Mark Meyer
Hi list,
In ox-epub I'm using several references to external stylesheets,
starting with the external name in the user filesystem I transform
these into a kind-of unique id that lives in the EPUB zip file.

Currently I'm doing this using a counter, generating zip entries in the
form of style-1.css. Can I use `org-export-get-reference' to generate
these ids? I would likely pass a simple string as the object to get a
reference to (the path the external stylesheet lives at).

>From what I understand this would generate different references for
`foobar.css' and `../foobar.css' even when these are identical
documents. This should be fixed by passing `(expand-file-name style)' to
`org-export-get-reference'.

I'm not quite sure I understand `org-export-get-reference' correctly,
can you elucidate this?

Cheers, Mark
-- 
   Mark Meyer
   m...@ofosos.org



[O] How to include menu entry in C-x C-e

2017-03-26 Thread Mark Meyer
Hi,
how do I include a menu entry from my module, when people hit C-x C-e?

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



[O] ox-odt backend use of org-export-async-start

2017-03-26 Thread Mark Meyer
Hi,
I'm presently looking through ox-odt/org-odt-export-to-odt. I uses a
quasiquote to include a parameter in the anonymous function that is then
passed to org-export-async-start. Is there any reason to do this? I
would assume with lexical variables this would be superflouus.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] ox-epub

2017-03-13 Thread Mark Meyer
Sorry, the code lives here: https://github.com/ofosos/ox-epub

Also available on MELPA.

Cheers, Mark

-- 
  Mark Meyer
  m...@ofosos.org



[O] ox-epub

2017-03-13 Thread Mark Meyer
Hi list,
I'm the author of ox-epub, which is a very recent extension to org-mode
to author EPUB files directly via the org export layer.

Pank (Rasmus) asked on Github, whether this could be part of org-mode.
I'm all for it. I gather this would be added to the lisp/contrib directory?

Are there any technical barriers to continuing development on ox-epub to
me? The legal stuff is not an issue to me.

Do you find the addition worthwile?

Speak your mind.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Error on html export

2017-02-27 Thread Mark Meyer
On 02/27/17 12:45, Nicolas Goaziou wrote:
> This is a common issue with mixed installations. You need to re-install
> Org without any other version loaded otherwise.

Spot on! Uninstalling org-mode `C-c C-x' and reinstalling worked.

Is there a FAQ for this kind of issues?

Cheers, Mark

-- 
  Mark Meyer
  m...@ofosos.org



[O] Error on html export

2017-02-27 Thread Mark Meyer
Hi,
I'm getting an error on HTML export, as soon as my document has a
`#+TITLE' set.

Here's part of the stack trace:

Debugger entered--Lisp error: (wrong-type-argument listp #("test" 0 4
(:parent (#1
  org-element-set-contents(#("test" 0 4 (:parent (#0
  apply(org-element-set-contents #("test" 0 4 (:parent (#0))) nil)
  #[257 "\211\305\306\307#\310\311\211:\204

I see that `org-element-set-contents' expects a list to operate on, but
is getting passed an array. This is with org-mode 9.0.5.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 19:28, Nicolas Goaziou wrote:
> Since you're writing the back-end, you don't need to
> use :completion-function, do you?
> 
> Just write the TOC the way you like.

I defined `*org-epub-current-file*' as a special variable and set it in
the publish function. Then I extracted the headlines in
`org-epub-template' (should find a better place). And this works like a
charm.

And now I can use the same heading text multiple times and get a proper
export.

Thank you very much.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 20:32, Eric Abrahamsen wrote:
> I put together just the very beginnings of the package, I think it's
> capable of unzipping and re-zipping an Epub file, and generating a new
> (empty) template directory structure. Take a look and see if anything in
> there is interesting to you -- I think its main strength is using esxml
> to generate the XML files, so you don't have to do that with strings.
> The idea was that that would also make it easier to manipulate resources
> later on, while editing. Anyway, it ain't much:
> 
> https://github.com/girzel/epub-mode/

Oh, this is nice :) That's a very good use of esxml. I definitely like it.

I thought about doing the zipping in `org-epub', but had to get some
other stuff working first. But you already wrote the code :)

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 19:28, Nicolas Goaziou wrote:
> Since you're writing the back-end, you don't need to
> use :completion-function, do you?
> 
> Just write the TOC the way you like.

Yay! You're completely right. The use case is, folding multiple
documents into a single EPUB (via a project), but I can easily write the
partial tocs into one/multiple temp buffers and copy them over after that.

Thanks.

Cheers, Mark

-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 19:12, Nicolas Goaziou wrote:
> It is also easy in Org 9 : (org-export-get-reference headline info). See
> its use in the function above.

I am under the impression that I don't have access to the `INFO' plist
that is required by `org-export-get-reference' - at least not from the
`:completion-function'.

I'll definitely looking into `org-export-collect-headlines'.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 18:28, Nicolas Goaziou wrote:
> I don't know the EPUB exporter. You probably could modify the Org source
> itself instead of the output.

I wrote it yesterday, the code's up at

 https://github.com/ofosos/org-epub

it's a little crufty, because I did write it for org 8 initially.

Hmm, I have to think about it. I would like to either have a 1:1 mapping
between the elements `org-element' delvers and the links the export
backend generates or create a function that runs before the export that
generates :custom_id properties for every element that doesn't have one.

This was easier in org 8, since every internal link was just something
like `sec-8-2'.

Thanks for you quick replies.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 17:39, Nicolas Goaziou wrote:
> There isn't. Internal links do not point to paths in the tree. You can,
> however, point to a specific headline using its unique name, if
> possible, or a custom ID otherwise.

What do you mean by unique name? Is this a property that is available
via org-element or do I have to set this in the source document?

I'm trying to link to all the headlines from a table of contents (EPUB
expoerter). Basically I'm using a `:completion-function' to generate all
the meta data and I need to link to most of the headlines down to a
certain level during the toc generation.

Thanks.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



Re: [O] Searching for a headline

2017-02-26 Thread Mark Meyer
On 02/26/17 16:27, Nicolas Goaziou wrote:
> The first argument for `org-publish-external-link' is a so-called
> "fuzzy" search string. In such a string "*whatever" is meant to match
> a headline named "whatever", whatever its level is. IOW, the "*"
> character denotes a headline, not a level.

Can you point me to a way to match a path in the tree?

Thanks.

Cheers, Mark
-- 
  Mark Meyer
  m...@ofosos.org



[O] Searching for a headline

2017-02-26 Thread Mark Meyer
Hi,
I'm using org-publish-external-link to look up a the link target for a
headline. The following code works, although I'd expect it to fail:

(org-publish-resolve-external-link
  (concat "* " (org-element-property :raw-value headline))
  source-file)

``headline'' is a headline as returned by org-element, while
``source-file'' is a ``.org'' file.

I'd expect that this needs to be

(org-publish-resolve-external-link
  (concat (make-string (org-element-property :level headline) ?*)
   " " (org-element-property :raw-value headline))
  source-file)

I.e. for a level 3 headline it should be ``*** foobar'' and not ``*
foobar''. ``org-publish-resolve-external-link'' uses
``org-export-string-to-search-cell'' internally.

Can someone point me to some docuentation on how the searching works?

Cheers, Mark

-- 
  Mark Meyer
  m...@ofosos.org