Re: [PATCH] org.el: make org-paragraph-fill ignore \[...\] regions starting and ending a line

[Rasmus <rasmus@gmx.us>]

Federico Beffa <beffa@ieee.org> writes:

> On Mon, Aug 11, 2014 at 3:06 PM, Nicolas Goaziou <mail@nicolasgoaziou.fr> wrote:
>> Federico Beffa <beffa@ieee.org> writes:
>>
>
>>> to help me understand what kind of problems one could face with HTML (or
>>> another back-end), could you give a concrete example?
>>
>> line 1
>> line 2
>> \[1+1\]
>>
>> ==>
>>
>> <p>
>> line 1
>> line 2
>> <img... />
>> </p>
>>
>> whereas
>>
>> line 1
>> line 2
>> \begin{equation*}
>> 1+1
>> \end{equation*}
>>
>> ==>
>>
>> <p>
>> line 1
>> line 2
>> </p>
>>
>> <p>
>> <img ... />
>> </p>
>>
>> IOW, there are two different structures in the document:
>>
>>   ((paragraph latex))
>>
>>   vs
>>
>>   ((paragraph) (latex))
>>
>> even though M-q cannot tell that difference (with your proposal, the
>> behaviour would be the same in both cases).
>
> The example highlight the difference that I suggested to remove in the
> very first place (by making \[...\] an environment). This was rejected
> to preserve backward compatibility and that's fine. So I moved on to a
> second proposal: modify the paragraph filling function.
>
> Didn't the following comment in your previous reply refer to the
> second proposal?
>
>> Ignoring \[...\] when filling the paragraph is misleading. You may
>> believe the object doesn't belong to the paragraph at all. I think M-q
>> should, on the contrary, give clues about the structure of the document.
>>
>> Also, it doesn't make a difference when exporting to LaTeX, but it might
>> in back-ends with a different definition for paragraphs (e.g. HTML).
>
> From your sentence in your last reply:
>    "... even though M-q cannot tell that difference (with your proposal,
>    the behaviour would be the same in both cases)."
> I understand that there is no technical deficiency in it. Am I
> therefore correct in saying that if you prefer not to include this
> proposal in org-mode it isn't for technical reasons but it is a matter
> of opinion/taste?
>
> I would like to know, because in any case I would like to use the
> proposed filling function in my copy of org-mode and if you see
> technical problems/errors I very much would like to be aware of them
> and, if possible, avoid/correct them.

How about adding the possibility to add hooks to org-fill-paragraph?
So that people can add "extensions" to fill as they want and the core
function can rely on org-element only?

—Rasmus

-- 
Lasciate ogni speranza, voi che leggete questo.