Re: How to fully control parsing of Org mode links during export

[Orm Finnendahl <orm.finnendahl@selma.hfmdk-frankfurt.de>]

Hi,

 this is how I currently do the debugging of html-export (there may be
more elegant ways, but it works well for me):

- Copy (defun org-export-as ...) from ox.el into a local elisp buffer
  of your liking (any buffer with the name .el should do).

- In that function add the line (setq info-debug info) just after the
  comment lines (Lines 2997/2998):

  ;; Eventually transcode TREE.  Wrap the resulting string into
  ;; a template.
  (setq info-debug info)

- Evaluate the new definition of org-export-as and run the export from
  your org file once.

Now you can access the complete parse tree with

(plist-get info-debug :parse-tree)

(make sure just to use a small document, otherwise the output might be
pretty nasty as the parents of all elements are recursively fully
contained in the parse tree).

You can extract all links of your document with:

(setq all-links
  (org-element-map (plist-get info-debug :parse-tree) 'link 'identity))

And you can check the output string of html with this:

(org-html-link (nth 0 all-links) "" info-debug)

You can then check what you have to do to make org-html-link output
the desired string, or redefine org-html-link to do what you need and
later make a derived backend from the html backend replacing the
org-html-link function or simply load the new definition of
org-html-link using a mechanism like #+SETUPFILE: in your header.

HTH,
Orm

Am Sonntag, den 21. Juli 2024 um 13:19:29 Uhr (-0400) schrieb Robert Weiner:
> Hi Org developers:
> 
> I could really use some help from Org export experts.  I have
> been trying to tweak export link parsing to do what I want
> without any success.  Here is a summary of the issues and related
> questions I have.
> 
> Usinf Org version "9.8-pre":
> 
> 1. When exporting from an Org file to html, I want this
>      [[Emacs#Section-One]] or [[file:Emacs::Section-One]]
>    to be converted into this:
>      <a href="Emacs.html">Emacs</a>#Section-One
>    but instead Org export produces:
>      <a href="Emacs.html">Emacs</a>#Section-One
> 
> 2. When exporting from an Org file to html, I want this
>      "[[One Testing]] Three"
>    to be converted into this:
>      "<a href="One Testing.html">One Testing</a> Three"
>    but instead Org export produces:
>      "<a href="One.html">One</a> <a href="Testing.html">Testing</a> Three"
> 
> 3. When exporting from an Org file and I have defined the html export
>    syntax for mytype: to html, I want this:
>      [[mytype:MyLink]]
>      [[Testing]]
>    to be converted into this:
>      <a href="MyLink.html">MyLink</a>
>      <a href="Testing.html">Testing</a>
>    but instead Org export produces:
>      <a href="<a href="MyLink.html">MyLink</a>.html">mytype:<a
> href="Testing.html">Testing</a></a>
> 
> My questions are:
> 
> 1. How can I change `org-element-link-parser' in tandem with
>    ox-html.el to output simple HTML links without any of the text
>    within double brackets bleeding outside of the HTML link?
> 
> 2. How can I get the exporter to run in my current Emacs process and
>    spawn a new Emacs so that I can easily debug in-process?  I think I have
>    turned off the async export option but it still runs async and I can't
>    step through the exporter functions.
> 
> Thanks very much for any help.
> 
> -- rsw