Re: [PATCH] Fixed bug in org-entry-get-with-inheritance

[Ilya Shlyakhter <ilya_shl@alum.mit.edu>]

Thanks Bastien.

Property API documentation could be made more precise in some places.
From the documentation of org-entry-get (both the docstring and the
Org manual), it would seem that unless the inherit argument is
non-nil, file-wide and system-wide property settings should not be
checked at all?   The Org manual seems especially clear: "By default,
this only looks at properties defined locally in the entry."
Logically, it makes sense to think of file-wide properties as being on
an implicit "level 0" headline under which all headlines in the file
are grouped, and system-wide properties as being on an implicit "level
-1" headline under which all level-0 headlines are grouped.
But, existing code and Org files might rely on org-entry-get without
inheritance considering file-wide and system-wide properties?

Documentation of the variable org-use-property-inheritance also can be
read to mean that it controls the inheritance of file-wide and
system-wide properties, though currently it does not ("When nil, only
the properties directly given in the current entry count.")

 For org-entry-properties, it says "Get all properties of the entry"
-- but it returns only properties explicitly defined at the entry, not
anything inherited from up the hierarchy or file-wide or system-wide,
right?   It also says, "Keys may occur multiple times if the property
key was used several times." -- but the manual also says that "a
property can only have one entry per Drawer".

Also, #+PROPERTY: lines have an effect no matter where they are in the
file, but what happens when several of them set the same property
isn't fully specified; I assume they're processed top-to-bottom, and
later settings overwrite earlier ones, just as later prop+ settings
append to earlier ones?  (So in this sense it does matter where in the
file a #+PROPERTY line is).  It might also be useful to clarify that
even if a subtree is archived, global property settings in that
subtree continue to have an effect (because they aren't really in a
subtree).

Manual says that #+PROPERTY lines specify "properties that can be
inherited by any entry in a file"; more precise would be that they
specify "property settings inherited by every entry"?  Likewise for
the variable org-global-properties.

On Tue, Mar 18, 2014 at 11:14 AM, Bastien <bzg@gnu.org> wrote:
> Bastien <bzg@gnu.org> writes:
>
>> Ilya Shlyakhter <ilya_shl@alum.mit.edu> writes:
>>
>>> When I open emacs with this file, move to the emacs-lisp block, and
>>> evaluate it, I get "aaa".
>>
>> I can reproduce your problem now, I'm on it, and the problem is real,
>> but I need to make sure all tests pass fine before fixing this.
>
> Okay, so I committed a different fix in maint and master.
>
> Please test it and let me know.
>
> Thanks,
>
> PS: You may want to read the commit message:
> http://orgmode.org/cgit.cgi/org-mode.git/commit/?id=42ee862d
>
> --
>  Bastien