* lisp/org-refile.el (org-refile):
* lisp/org-capture.el (org-capture-store-last-position):
Pass `format` arg to `with-demoted-errors`.
(with-demoted-errors): Warn on missing `format` arg
d52c929e31f60ff0462371bfe27ebd479e3e82bd
Stefan Monnier
Fri Feb 4 19:39:53 2022 -0500
[km] with-demoted-errors has had a format argument since Emacs 24.4.
Passing it avoids an error when compiling with Emacs's current
master (bug#54225). Once that bug is resolved, it will prevent a
warning.
* lisp/org.el (org-copy-visible): Decide whether text is invisible by
calling invisible-p rather than checking whether the invisible
property at point is non-nil.
Text may have a non-nil invisible property but _not_ be hidden from
the user (and thus should be copied by org-copy-visible). For
example, the link itself is shown when org-link-descriptive is nil,
but it still has an invisible property of `org-link'.
* lisp/org.el (org-copy-visible): Don't copy invisible text that
follows invisible text with a different property value.
If org-copy-visible sees that the left bound position has a non-nil
invisible property, it uses next-single-char-property-change to find
the new bound. However, next-single-char-property-change may just
find a bound that still has a _different_ non-nil invisible property.
Reported-by: "Максим Бабушкин" <maxbabushkin@gmail.com>
Link: https://debbugs.gnu.org/49967
* testing/lisp/test-org.el (test-org/copy-visible): Switch defun to
ert-deftest so that the test is actually executed, and resolve
failures in test.
Make two adjustments so that the now executed test passes:
- Correct the end value of the hidden part of the string for one
test case.
- Disable org-unfontify-region so that the invisible properties show
up in the temporary buffer.
I think org-unfontify-region should probably be changed to _not_
remove the invisible property, but it's a longstanding behavior.
Changing it would need more thought and isn't something that
should be done on the bugfix branch.
* lisp/oc-basic.el (org-cite-basic--parse-json): Make date-parsing and
year extraction more resilient. Provide more informative errors when it
fails.
A string-based date is not only indicated by the key 'raw, but also
possibly by the key 'literal.
String-based dates come in various formats, not necessarily yyyy-mm-dd.
So extracting the first sequence of 4 digits is arguably a better
heuristic for getting the publication year than splitting the string on
- and getting the car of that.
On error, include `value' in the message, which contains the original
value with actionable information, whereas the previously included
`date' is always nil in that case.
TINYCHANGE
* lisp/org-list.el (org-list-send-item): Check if dest is a string
before matching it, to allow dest to be a buffer position.
* testing/lisp/test-org-list.el (test-org-list/send-item): Add tests.
* lisp/ol-bibtex.el (org-bibtex-read): Make sure that the return value
is the updated `org-bibtex-entries'. Document the return value
requirement. The assumption about the return value is used by
`org-bibtex-yank'.
Fixes https://list.orgmode.org/86zgn689m5.fsf@krailli2.fritz.box/T/#u
* lisp/ob-tangle.el (org-id-link-to-org-use-id):
* lisp/org-agenda.el (org-agenda-overriding-header): Do not try set
value of a variable defined elsewhere. This may interfere with custom
interface. See
https://list.orgmode.org/orgmode/871r19ryxb.fsf@localhost/
(org-agenda-restrict):
* lisp/ol-bibtex.el (org-bibtex-description): Remove incorrect
comments.
* lisp/ob-gnuplot.el (org-babel-gnuplot-table-to-data): Revert to
using ascii exporter.
* lisp/ox-ascii.el (org-ascii-table-cell): Do not throw error when
CONTENTS is wider than initial cell value.
* testing/examples/agenda-file.org: Add test headings where SCHEDULED
line is not planning.
*
testing/lisp/test-org-agenda.el (test-org-agenda/non-scheduled-re-matces):
Make sure that SCHEDULED lines that do no belong to planning do not
trigger agenda matches.
See https://list.orgmode.org/20220101200103.GB29829@itccanarias.org/T/#t
* lisp/ob-gnuplot.el (org-babel-gnuplot-table-to-data): Switch to 'org
backend when exporting table data to plain text table. ascii backend
incorrectly handles quote transcoders that change the column width and
sometimes throws an error. (ascii backend simply ignores transcoded
table cells and calculates width based on the initial cell text).
* lisp/org-macro.el (org-macro--vc-modified-time): Wrap file in a list
when passing to vc-call to avoid a type error starting with Emacs 28.
Before Emacs's 3572613550f (Fix vc-git-state for filenames with
wildcards, 2021-08-15), a string could be given to vc-git-print-log as
the FILES argument because the argument was passed directly as
vc-git-command's FILE-OR-LIST, which as the name suggests, accepts
either a string or a list. After 3572613550f, passing a string leads
to a type error.
TINYCHANGE
* lisp/ox-beamer.el (org-beamer-export-to-pdf):
* lisp/ox-icalendar.el (org-icalendar-export-to-ics):
* lisp/ox-koma-letter.el (org-koma-letter-export-to-pdf):
* lisp/ox-man.el (org-man-export-to-pdf):
* lisp/ox-texinfo.el (org-texinfo-export-to-info): Quote lambda.
Quote or name lambdas to prevent their compilation into anonymous
functions which cannot be passed to the external async emacs process.
* lisp/org.el (org-priority): When reading numeric priorities, provide
special handling of SPC so that it clears the priority, as advertised.
If the user has configured numeric priorities and tries to clear a
priority by inputting SPC at the prompt, org-priority feeds " " to
string-to-number and ends up with 0 instead of the ?\s (32) that's
used downstream to signal "remove". Explicitly check for " " and
translate it to ?\s.
Reported-by: "Bruce E. Robertson" <brucer42@gmail.com>
Link: https://list.orgmode.org/877dch89s1.fsf@kyleam.com/
* lisp/oc-basic.el (org-cite-basic--key-completion-table): Cache
completions per document. Return nil when no completion is available
so insertion mechanism can properly report the problem.
* doc/org-manual.org (Advanced Export Configuration):
(Extracting Source Code): Rename "Hooks" subheading so that ox-texinfo
does not make it unique by appending " (N)".
The Emacs repo has a check for problematic characters in info nodes
that flagged "Hooks (2)" (bug#52126). As mentioned in that thread,
this case is actually probably not problematic because the name
doesn't _start_ with "(". However, it's easy enough to sidestep the
warning, and adding the extra context doesn't hurt.
* testing/lisp/test-org-attach-git.el: Provide feature.
As of 9044b300e (org-test: Fix ERC errors using latest Emacs master,
2021-11-24), tests must provide a feature.
Ihor Radchenko
Stefan Monnier
Nicolas Goaziou
Kyle Meyer
David Lukes
Sacha Chua
TEC
Bastien
Marco Wahl
Stefan Kangas
TAKAHASHI Yoshio
Sébastien Miquel
Rasmus