Rasmus <rasmus@gmx.us> writes:

> The attached patch fixes this by explicitly saving the footnote section

As per usual my first patch is dodgy.  It occurred to me that Org can
handle several footnote sections (that's how #+INCLUDE supports footnotes,
I guess).  The attached patch how supports export of t2.org in this
example

    $> cat t{1,2}.org

    # this is t1.org
    * foo
    bar[fn:1]
    baz[fn:2]

    * Footnotes

    [fn:1] bar.corp

    * Footnotes

    [fn:2] baz.corp

    # this is t2.org
    #+INCLUDE: "/tmp/t1.org::*foo"  


However, there is a pitch-fall when doing

    #+INCLUDE: "/tmp/t1.org::*foo0"
    #+INCLUDE: "/tmp/t1.org::*foo1"  

Now *foo1 will be inserted under the footnote-heading of *foo0 which means
it won't be exported.  If min-level is used this is not an issue of
course, but that's kind of unexpected (so is the fact that the second
include will become a child of the first include, but that's another
patch) .

Perhaps a better overall approach (if the above limitation is
unacceptable) would to translate all footnotes in an INCLUDEd file to
inline ones, e.g. when including * foo in t1.org above, what would be
inserted is

    * foo
    bar[fn::bar.corp]
    baz[fn::baz.corp]

Yet another solution would be to return a cond of

    (included-text . included-footnotes)

and make sure to insert footnotes at the very end.

WDYT?

—Rasmus

-- 
. . . The proofs are technical in nature and provides no real understanding