* lisp/ob-octave.el (org-babel-execute:octave):
-Ensure that the special Octave variable `ans' is bound when
GFX-FILE is non-nil. The glue code in
ORG-BABEL-OCTAVE-WRAPPER-METHOD causes Octave to exit with a
non-zero exit code when `ans' is not bound.
-Change format control string to %S from %s. Ensure the graphics
filename is quoted. If it is not, Octave may create a mis-named
file or fail entirely.
* testing/examples/ob-octave-test.org:
Update the Graphical tests section:
-put in the correct headers;
-add a remark about where to find each test.
* testing/lisp/test-ob-octave.el:
Add the three tests ob-octave/graphics-file,
ob-octave/graphics-file-session and ob-octave/graphics-file-space.
-ob-octave/graphics-file: The first test verifies that the first bug
identified above is fixed; it also verifies that graphics file
creation works correctly for scripting.
-ob-octave/graphics-file-session: The second test verifies graphics
file creation works correctly for sessions. The Octave command
`crash_dumps_octave_core(0)' is included to prevent the creation of
a core file (`octave-workspace').
-ob-octave/graphics-file-space: The third test verifies that a
graphics filename with a space in it is created correctly.
Thanks to Ihor Radchenko for helpful feedback.
Ref: https://list.orgmode.org/8735asbtfe.fsf@localhost/T/#u
* ob-R.el (org-babel-R-assign-elisp): Use the patch from
ccberry@health.ucsd.edu, to print org-list as a one column table as it
was the case in release_9.5. The break in R is due commit b4e437f96 *
ob-core: Resolve named list references to simple lists.
* test-ob-R.el (ob-R-nested-list): New function to test that org list
with multiple level are handled as expected in R.
see https://list.orgmode.org/87bkofh0ir.fsf@localhost/ for context.
* test-ob-R.el (ob-session-R-result-output): This test will check if
output is printed to buffer in a session with :results output. This
test is to prevent the bug mentioned in https://list.orgmode.org/877czca7oj.fsf@u-bordeaux.fr/
does not happen again.
* lisp/org-persist.el (org-persist-write:url): Since the url write
function is called as part of `org-persist-write-all', it is worth
adding a check to avoid re-downloading the file if a file already exists
in the expected location.
* lisp/org.el (org--confirm-resource-safe, org--safe-remote-resource-p):
Replace instances of buffer-file-name
with (buffer-file-name (buffer-base-buffer)) so these functions work in
indirect buffers.
* lisp/ox.el (org-export-link-localise): When no local copy of the link
resource could be fetched, produce a warning message and do nothing
instead of setting the link :path to nil.
* lisp/org.el (org-hide-drawer-startup): Make alias to
`org-cycle-hide-drawer-startup'.
Reported-by: Kévin Le Gouguec <kevin.legouguec@gmail.com>
Link: https://orgmode.org/list/877cyvzrcu.fsf@gmail.com
* lisp/org-element.el (org-element-use-cache): Autoload to silence
byte compiler warning in Emacs repo.
On Emacs's emacs-29 branch, compiling lisp/org/org-loaddefs.el gives
the following warning that stems from an autoloaded defsubst,
org-element-at-point-no-context (new to Org 9.6):
In org-element-at-point-no-context:
org/org-loaddefs.el:846:61: Warning: Unused lexical variable
`org-element-use-cache'
That doesn't show up in the Org repo because mk/org-fixup.el generates
org-loaddefs.el via autoload.el's generate-file-autoloads, which adds
"no-byte-compile: t" to its output file. However, in Emacs 29,
autoload.el is now deprecated, and the Emacs build uses
loaddefs-gen.el's loaddefs-generate--emacs-batch, which does include
no-byte-compile.
* lisp/org.el (org-display-inline-remove-overlay): Force remove
overlay images from Emacs image cache when un-displaying inline image
overlays. This change should: (1) slightly improve Emacs memory usage
when the number of inline images is large; (2) Clear images from cache
every time Emacs removes the overlay for any
reason. (`org-display-inline-remove-overlay' is used in
overlay 'modification-hooks.)
cc5427b7e (org-babel-result-to-file: Fix results file inside
attachment dir, 2022-12-07) carried along an unrelated white space
change upstream of the main change.
* lisp/org-table.el: Explicitly load org-fold-core.
'make single' warns that org-fold-core-ignore-modifications in
org-table.el (introduced a few commits back) is not known to be
defined, and the Emacs repo would show the same warning.
* doc/org-manual.org (Re-aligning and field motion): Update the
command description according to up-to-date docstring.
Reported-by: André A. Gomes <andremegafone@gmail.com>
Link: https://orgmode.org/list/87zgfrvzz1.fsf@localhost
* lisp/org.el (org-display-inline-images): Always refresh cached
images when displaying image overlays. See Emacs bug#59902. This
fixes images not updating when image file is changed on disk.
Reported-by: William Denton <wtd@pobox.com>
Link: https://orgmode.org/list/87wn7287n2.fsf@localhost
* lisp/ox.el (org-export-as): Make sure that narrowing includes top
heading in a subtree when subtree export is requested. This is needed
after 927621910 changed `org-export--get-subtree-options' behavior.
Reported-by: Alejandro Pérez Carballo <apc@umass.edu>
Link: https://orgmode.org/list/m2sfhpffui.fsf@umass.edu
* lisp/org-fold-core.el (org-fold-core-save-visibility): Calculate
folds on runtime, not during macro expansion.
* testing/lisp/test-org.el (test-org/org-ctrl-c-ctrl-c): Add test
ensuring that outline visibility is preserved upon refreshing Org
setup.
Reported-by: Matt Lundin <mdl@imapmail.org>
Link: https://orgmode.org/list/87wn70tlcn.fsf@fastmail.fm
* lisp/org-table.el (orgtbl-to-generic): Disable org-fold checks in
temporary buffer. Folding state is irrelevant there.
Reported-by: Majzoub, Eric <eric.majzoub@umsl.edu>
Link: https://orgmode.org/list/tn0s4e$84u$1@ciao.gmane.io
* lisp/ox-latex.el (org-latex-example-block): Instead of
`org-latex--wrap-label', manually add label after
\begin{<environment>}.
This produces proper label resolution in LaTeX. Labels before the
environment begins can be interpreted as referring to the surrounding
section.
* lisp/org-persist.el (org-persist-write): Do not write data
associated with gpg-encrypted files. This will avoid data leakage.
(org-persist--storage-version): Bump storage version to invalidate
previously stored, potentially leaked, data.
Reported-by: Clément Pit-Claudel <clement.pitclaudel@live.com>
* lisp/org.el (org-make-tags-matcher): Match
<now>/<today>/<tomorrow>/<+3d>/etc time queries. They are allowed by
"11.3.3 Matching tags and properties" section of the manual.
Fixes regression from e022a0cea.
Reported-by: Carlo Tambuatco <oraclmaster@gmail.com>
Link: https://orgmode.org/list/87bkokgtd7.fsf@kyleam.com
* lisp/ob-core.el (org-babel-result-to-file): When file path in the
result contains attachment dir, reduce the path to relative file name
when generating attachment: link.
* testing/lisp/test-ob.el (test-ob-core/dir-attach): Add test.
Reported-by: Johan Tolö <johan@toloe.se>
Link: https://orgmode.org/list/875yesnia6.fsf@toloe.se
* lisp/org-element.el (org-element--cache-active-p): Consider cache
disabled when any of the indirect buffers do not have modification
detection. This can be important when one uses
`combine-after-change-calls' inside an indirect buffers disabling
`after-change-functions' temporarily in the indirect buffer.
Reported-by: Feng Shu <tumashu@163.com>
Link: https://orgmode.org/list/87359z8ssb.fsf@163.com
* lisp/org-fold-core.el (org-fold-core-cycle-over-indirect-buffers):
Fix logic detecting foreign `org-fold-core--indirect-buffers'.
`current-buffer' is not always at `car'. Before this change, capture
buffers could be missed when cycling over indirect buffers causing
subtle bugs.
Reported-by: Alexandros Prekates <aprekates@posteo.net>
Link: https://orgmode.org/list/20221128122841.2246aa08@enous
* lisp/ob-comint.el (org-babel-comint-with-output): Do not rely on
output filter when detecting prompt lines. Search for
`comint-prompt-regexp' in the full output after applying the filter.
This is needed when the filter is supplied with partial prompts.
Reported-by: Frédéric Santos <frederic.santos@u-bordeaux.fr>
Link: https://orgmode.org/list/877czca7oj.fsf@u-bordeaux.fr
* lisp/ox-md.el (org-md-toplevel-hlevel): Change the type from natnum
to integer, which is less accurate but compatible with all Emacs
versions that Org supports.
This is a follow-up to the recent port of 4b2f5425d.
* lisp/org-clock.el (org-x11idle-exists-p): Don't try to
invoke "command x11idle" on MS-Windows and MS-DOS. This fixes a
hang during byte-compilation of Org.
Fix compilation of Org 9.6 on MS-Windows
8b3baadf671b7b3d5058808c4655eca8bd46015e
Eli Zaretskii
Thu Dec 1 14:40:47 2022 +0200
* lisp/ob-clojure.el
(ob-clojure-babashka-command, ob-clojure-nbb-command):
Account for the possibility of nil (commands absent).
* lisp/org.el (org-safe-remote-resources):
* lisp/ox-md.el (org-md-toplevel-hlevel):
Fix incorrect or invalid types.
Fix types in various defcustom declarations
4b2f5425d7af1a6f992ad8ce0a4452a2e778815a
Mattias Engdegård
Thu Dec 1 13:03:19 2022 +0100
[km] The natnum change will be adjusted in an upcoming commit because
natnum not available until Emacs 28.
Three entries describe changes made in the 9.6 release, not 9.5. The
associated commits are 226119124 (ob-core.el/babel: Special handling
for attachment links in src block, 2022-06-14), 7f3a6cf6e
(org-capture: Add template hook properties, 2022-10-07), and b4e437f96
(ob-core: Resolve named list references to simple lists, 2022-11-26).
* lisp/ob-core.el (org-babel-result-cond): Unconditionally return nil
and suppress all the processing for :results discard.
(org-babel-common-header-args-w-values):
(org-babel-sha1-hash): Add the new value to know :results value list.
* doc/org-manual.org (Handling):
* etc/ORG-NEWS (New =:results discard= header argument): Document the
new value.
Reported-by: Daniel Ortmann <daniel.ortmann@oracle.com>
Link: https://orgmode.org/list/87tu2tjary.fsf@localhost
(=python-mode.el (MELPA)= support in =ob-python.el= is deprecated):
Document that we no longer support third-party python-mode.el package
in favor of the built-in python.el.
Link: https://orgmode.org/list/87r0yk7bx8.fsf@localhost
* testing/lisp/test-org-element.el (test-org-element/class):
Uncomment. Whatever caused the failure in the past is not longer the
problem. Tests are passing.