Re: [O] [bug, patch, ox] INCLUDE and footnotes

[Rasmus]

Hi,

Nicolas Goaziou <address@hidden> writes:

> Rasmus <address@hidden> writes:
>
>>>> * Foo
>>>> [1] foo
>>>>
>>>> * Bar
>>>> Baz[1]
>>>
>>> I'm not sure to understand. Would you mind elaborating?
>>
>> If I have #+INCLUDE: "example-above.org::*Bar" then point-min of the
>> include area will be pushed forward by four since the definition of [1] is
>> changed to fn:1-1 or something like that.  So min-marker should be a
>> marker.  Or I'm misunderstanding something.
>
> No, you're right. However, this raises a question: why are we modifying
> definition at all? We are only interested in its new label, which we can
> get without modifying buffer (i.e. if definition is within range, modify
> it, otherwise, compute new label and store its definition).

We modify buffer because that's what we want to do when including whole
files.

The routines /could/ be split up, I just deemed it "not worth the
trouble".  Operations on the table are of course limited to when it's
needed.  Buffer-editing is not.  It's simple to wrap it in an if-statement
if you think it's worth it, e.g. performance-wise.  I'd only need to move
the catching of new-label back to the footnote-reference.

>> +                (org-with-wide-buffer
>> +                 (let* ((definition (org-footnote-get-definition label))
>> +                        (beginning (line-beginning-position)))

> There's one potential problem here: `org-footnote-get-definition' may
> return a nil value if there is no matching definition for label. Maybe
> throw an error?

Ox already throws an error when a footnote is not found in
org-export-get-footnote-definition which is why I didn't add this.  But I
guess it would be friendly to add another error here stating which *file*
is missing a footnote and I added this now.

> Also, BEGINNING should refer to (nth 1 definition) since you're not
> using `org-footnote-goto-definition' and therefore, not moving point.

Indeed.  Thanks.

> I think you can push once the issues above are fixed. Thank you for the
> work.

Cool.  I think the #+INCLUDE-keyword is quite a bit better in 8.3 now and.
Thanks for all the help on this series of patches (I think four in total)!

—Rasmus

-- 
However beautiful the theory, you should occasionally look at the evidence