* lisp/org.el (org-auto-fill-function): Make sure `adaptive-fill-mode'
mode is nil when pre-computed `fill-prefix' is the empty string.
Otherwise filling functions from fill.el think it has to be computed
again and overwrite it.
* lisp/org.el (org-end-of-line): Do not call `end-of-visual-line' when
moving to the end of line. Also improve behaviour on elements that
can be hidden.
* testing/lisp/test-org.el: Add tests.
* lisp/org-element.el (org-element--get-next-object-candidates): Fix
parsing of objects of the same type in a single paragraph.
* testing/lisp/test-org-element.el: Add tests.
* lisp/org-element.el (org-element-sub/superscript-successor): Fix
parsing of sub/superscript at beginning of item.
(org-element-latex-or-entity-successor): Fix parsing of latex
fragments at beginning of item.
* testing/lisp/test-org-element.el: Add test.
* lisp/org.el: Remove utf-8 codepoints in docstrings, bytecode doesn't
work when loaded from compressed files.
Probably a bug in Emacs, but since this is unlikely to be fixed in old
Emacsen, simply don't do it.
* lisp/org-compat.el: Make sure that file-name-directory is getting a
stringp. This avoids a possible "(wrong-type-argument stringp nil)"
error when the library in question does not exist.
* lisp/Makefile: Arrange for an uncompiled org-install.el. This also
ensures that if there's an old org-install.el in the installation
directory, it will be replaced and not linger.
* lisp/org-install.el: Provide an empty file that prints a warning
about an outdated configuration.
* lisp/org-element.el (org-element-context,
org-element--get-next-object-candidates): Fix `org-element-context'.
In particular, the restrictions for an object may be different from
those of its container (i.e. table rows and table cells).
* testing/lisp/test-org-element.el: Add tests.
* org-e-odt.el (org-e-odt-styles-dir): Try more directories.
Don't throw an error, just send a message.
(org-e-odt): Fix a bug about deactivating an advice that was
activated by org-odt.el, not org-e-odt.el.
Thanks to Nick Dokos for reporting this.
* org.texi (Installation): Don't mention org-install.el
anymore as the replacement file org-loaddefs.el is now loaded
by org.el.
* orgguide.texi (Installation): Ditto.
* lisp/org-element.el (org-element-example-block-parser,
org-element-src-block-parser): Store value of example-blocks and
src-blocks unescaped.
(org-element-example-block-interpreter,
org-element-src-block-interpreter): Escape value again when storing
it.
* contrib/lisp/org-export.el (org-export-unravel-code): Don't clean
commas from code since org-element already took care of it.
* testing/lisp/test-org-export.el: Update test.
* testing/lisp/test-org-element.el: Add tests.
* lisp/org-src.el (org-escape-code-in-string,
org-unescape-code-in-string, org-escape-code-in-region,
org-unescape-code-in-region): New functions.
(org-edit-src-code, org-edit-src-exit): Use new functions.
* lisp/org.el (org-strip-protective-commas): Removed function.
* lisp/org-exp.el (org-export-select-backend-specific-text): Use new
function.
* lisp/ob.el (org-babel-parse-src-block-match,
org-babel-parse-inline-src-block-match, org-babel-insert-result):
Always escape produced blocks, independently on the language of the
block, if any. Use new functions.
* doc/org.texi: Update documentation.
* testing/lisp/test-ob.el: Update test.
* org-agenda.el (org-agenda-get-timestamps): Check if the item
is an habit when formatting it with `org-agenda-format-item'.
(org-agenda-get-blocks): Fix bug: don't assume the item is an
habit when formatting with `org-agenda-format-item'.
Thanks to George McNinch for reporting this.
* org-fixup.el (org-make-org-version): Use defvar instead of
defconst to define the `org-odt-data-dir' autoload.
Thanks to Glenn Morris for pointing at this.
* org.el (org-calendar-agenda-action-key): Delete an option.
(org-mode-map): Delete its keybinding.
(org-agenda-action-marker, org-mark-entry-for-agenda-action):
Delete.
* org-agenda.el (org-agenda-diary-entry): Don't prevent from
being used outside of Org agendas, as it can be used in
calendar buffers too.
Note that we may have to reimplement the functions related to
"k" in the calendar.
Bastien Guerry
Nicolas Goaziou
Achim Gratz