EWW/SHR Feature Request

EWW/SHR Feature Request

[T. V. Raman]

Now that we have shr and eww, it would be nice to have a function
that takes an html string  and returns a formatted representation
(replete with properties etc.)  something like 
(shr-format-html-string html-string) -> Return: Formatted string
suitable for insertion into a buffer. 
-- 

-- 

Re: EWW/SHR Feature Request

[Nicolas Richard]

"T. V. Raman" <raman@google.com> writes:

> Now that we have shr and eww, it would be nice to have a function
> that takes an html string  and returns a formatted representation
> (replete with properties etc.)  something like 
> (shr-format-html-string html-string) -> Return: Formatted string
> suitable for insertion into a buffer. 

Is it something like this ?

(let ((html-string  "<a href=\"http://www.gnu.org\">GNU website</a>"))
  (with-temp-buffer
    (insert html-string)
    (shr-render-region (point-min) (point-max))
    (buffer-string)))

(the html-string should in fact be a complete html doc, but libxml does
its best to turn bad html like the above into *something*.)

-- 
Nico.

Re: EWW/SHR Feature Request

[Vibhav Pant]

On Fri, Mar 21, 2014 at 9:14 PM, T. V. Raman <raman@google.com> wrote:
> Now that we have shr and eww, it would be nice to have a function
> that takes an html string  and returns a formatted representation
> (replete with properties etc.)  something like
> (shr-format-html-string html-string) -> Return: Formatted string
> suitable for insertion into a buffer.

By formatted string, do you mean the entire content of the page to
which the address redirects too, or the webpage title?

-- 
Vibhav Pant
vibhavp@gmail.com

Re: EWW/SHR Feature Request

[T.V Raman]

I mean the following:

(shr-format-html-string "<bo>Strong</b>")

Returns a string with the font face set to 'bold or whatever else
shr chooses to use.

Here is the code I have in emacspeak:
(defun emacspeak-webutils-html-string (html-string)
  "Return formatted string."
  (or (require 'shr) (error "Need  emacs 24.4"))
  (with-temp-buffer "*html-format*"
                    (setq buffer-undo-list t)
                    (insert html-string)
                    (shr-render-region  (point-min) (point-max))
                    (buffer-string)))


Vibhav Pant writes:
 > On Fri, Mar 21, 2014 at 9:14 PM, T. V. Raman <raman@google.com> wrote:
 > > Now that we have shr and eww, it would be nice to have a function
 > > that takes an html string  and returns a formatted representation
 > > (replete with properties etc.)  something like
 > > (shr-format-html-string html-string) -> Return: Formatted string
 > > suitable for insertion into a buffer.
 > 
 > By formatted string, do you mean the entire content of the page to
 > which the address redirects too, or the webpage title?
 > 
 > -- 
 > Vibhav Pant
 > vibhavp@gmail.com

Re: EWW/SHR Feature Request

[Lars Magne Ingebrigtsen]

raman@google.com (T.V Raman) writes:

>   (with-temp-buffer "*html-format*"
>                     (setq buffer-undo-list t)
>                     (insert html-string)
>                     (shr-render-region  (point-min) (point-max))
>                     (buffer-string)))

Note that `with-temp-buffer' doesn't take a buffer name as a parameter,
so the "*html-format*" string isn't used.  And there is no undo in the
temporary buffers,  so the whole thing just devolves to

(with-temp-buffer
  (insert html-string)
  (shr-render-region (point-min) (point-max))
  (buffer-string))

and I'm not sure that's enough to warrant a function...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: EWW/SHR Feature Request

[Willem Rein Oudshoorn]

"T. V. Raman" <raman@google.com> writes:

> Now that we have shr and eww, it would be nice to have a function
> that takes an html string  and returns a formatted representation
> (replete with properties etc.)  something like 
> (shr-format-html-string html-string) -> Return: Formatted string
> suitable for insertion into a buffer. 

I have been using this function for a while


(defun salesforce-insert-html (html)
  "Insert at current point, the 'rendered' HTML."
  (when html
    (shr-insert-document
     (with-temp-buffer
       (insert html)
       (libxml-parse-html-region (point-min) (point-max))))))


Wim Oudshoorn

Re: EWW/SHR Feature Request

[raman]

Following up to myself:

For now I solved my use-case by advising 
eww-browse-with-external-browser -- that function is presently called by
EWW for content-types matching  
eww-use-external-browser-for-content-type 
the default value of that var is 
"\\`\\(video/\\|audio/\\|application/ogg\\)"
i.e. video/audio types, and those were the types  I wanted to handle
 within Emacs.

Longer term, there are two ways we could go:

1. EWW already has functions for handling PDF, img types etc. -- we
could set EWW up to handle  audio/video via a user configurable
function, with the default handing off  to  a full-blown browser as now.

2. The other alternative would be to process the user's mailcap file,
but that might not be ideal because in general  most users will want the
.mailcap file to govern their primary browser's behavior.
--