* lisp/ob-C.el (org-babel-C-val-to-C-type): Floats should be printed
as string literals to prevent rounding introduced by %f format.
* testing/lisp/test-ob-C.el (ob-C/float-var): Test that floats are not
rounded when passed as an org :var.
* lisp/ob-C.el (org-babel-C-val-to-C-type): Floats should be printed
as string literals to prevent rounding introduced by %f format.
* testing/lisp/test-ob-C.el (ob-C/float-var): Test that floats are not
rounded when passed as an org :var.
* lisp/org-fold-core.el (org-fold-core--keep-overlays): New internal
macro, signaling `org-fold-core-region' to keep overlays in place and
store new overlays for later use in isearch.
(org-fold-core-region): Never remove overlays and store newly created
ones in `org-fold-core--isearch-overlays' when
`org-fold-core--keep-overlays' is non-nil. Remove resolved FIXME.
(org-fold-core--isearch-setup): Advice `isearch-clean-overlays' as
more reliable way to clear all the temporary overlays created for
isearch. `query-replace' and EVIL packages use
`isearch-filter-predicate' separately as thus `isearch-mode-end-hook'
is not sufficient to ensure that isearch overlays are converted back
to text properties.
(org-fold-core--isearch-show-temporary): Do not alter match data, as
isearch expects the match data to stay unchanged. Arrange the
overlays to be kept for isearch consumption (otherwise, isearch will
signal an error; see
https://list.orgmode.org/orgmode/87pmc4smdg.fsf@fastmail.fm/).
(org-fold-core--create-isearch-overlays): Call `org-fold-core-region'
with let-bound `org-fold-core-style' instead of repeating the code
from `org-fold-core-region'.
Reported-by: Michael Dauer <mick.dauer@gmail.com>
Link:
https://list.orgmode.org/orgmode/CAP7OBx+L11ck3Ni6rv94HGU3otdj6C4rG-rMDzkwR1LTj=BWiw@mail.gmail.com/
Link+: https://list.orgmode.org/orgmode/87pmc4smdg.fsf@fastmail.fm/
Link+: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=60399
Link+: https://yhetil.org/emacs-devel/87fs7c10cq.fsf@web.de/
* lisp/org.el (org-mode-syntax-table): New explicitly defined variable
for Org mode syntax table.
(org-mode-tags-syntax-table): Define value in advance, available even
before Org mode is loaded.
(org-mode): Use the new `org-mode-syntax-table' variable.
Reported-by: Daniel Clemente <n142857@gmail.com>
Link: https://orgmode.org/list/CAJKAhPCcMwZS6aYK9MRKsjds9uTivOdfC9nzMok70Df8vR3tPA@mail.gmail.com
* lisp/org.el (org-return): When `org-return-follows-link' is non-nil
and point is over an org-cite citation or citation reference, call `org-open-at-point'.
TINYCHANGE
* lisp/org-element.el (org-element-headline-parser): Store :true-level
property.
(org-element--parse-to): Call `org-get-limited-outline-regexp'
directly rather than via `org-with-limited-levels'. Avoid moving
point unnecessarily. Prefer `org-headline-re' to generate regexps.
* lisp/org-macs.el (org-headline-re):
(org-skip-whitespace): Inline.
* lisp/org-element.el (org-element--current-element): Do not try to
search cache. This is rarely useful and forces an awkward
`org-element-copy' workaround what parsing buffer.
(org-element--parse-elements): Do not use `org-element-copy' for
element at point - it is now safe to assume that
`org-element--current-element' does not return cached.
(org-element--parse-to): Do not force disabled cache when calling
`org-element--current-element'. It is no longer needed.
* lisp/org-element.el (org-element--current-element): Try hard
spending minimal amount of time in testing which parser to call.
Prefer `look-at-p' that does not modify match data and thus save us
significant amount of time. Do not call `org-with-limited-levels' and
instead use the simplest possible tests with all unnecessary branches
removed. Add commentary explaining the importance of writing fast
core when adding new elements.
(org-inlinetask-min-level): Declare variable.
* lisp/org-list.el (org--item-re-cache):
(org-item-re): Cache results rather than calculating regexp every time
the parser is invoked.
* lisp/org-macs.el:
(org--headline-re-cache):
(org--headline-re-cache-no-bol):
(org--headline-re-cache-bol):
(org-headline-re): Prefer plists to store headline regexp cache - the
number of items tends to be within 10-20 and hence hash table is an
overkill and will be slower.
* lisp/org-macs.el (org-headline-re): New function to retrieve cached
or get a new regexp for headline of level TRUE-LEVEL.
(org-get-limited-outline-regexp): Use `org-headline-re'. Add new
optional argument WITH-BOL.
* lisp/org-crypt.el: Call `org-cycle-set-startup-visibility' on the
decrypted entries. Still re-fold if the decrypted entry has been
inside an existing fold.
* lisp/org-cycle.el (org-cycle-set-visibility-according-to-property):
Respect narrowing.
* etc/ORG-NEWS (=org-crypt.el= now applies initial visibility settings
to decrypted entries): Announce the change.
This patch solves a slight annoyance when the freshly decrypted entry
is fully unfolded, including all the property drawers. It will make
more sense to make the folding state follow initial visibility
settings, including VISIBILITY properties.
* lisp/ox.el (org-export-dispatch): Add customizable variables
`org-export-body-only', `org-export-visible-only', and
`org-export-force-publishing', and use them in `org-export-dispatch'.
* doc/org-manual.org (The Export Dispatcher): Document the new export
variables.
* etc/ORG-NEWS (New customization options for ~org-export-dispatch~):
Announce the new customization options.
Currently when calling `org-export-dispatch', two of the export
options can have their defaults specified with customizable variables:
"Export scope" (via `org-export-initial-scope') and "Async export"
(via `org-export-in-background'). This change adds customizable
variables for the "Body only", "Visible only", and "Force publishing"
options as well.
* lisp/org.el (org-add-prop-inherited): Do not modify argument by side
effect. Modifying strings may cause unexpected issues when working
with constant strings.
Link: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=62847
* lisp/org-element.el (org-element-headline-parser):
(org-element-inlinetask-parser):
(org-element-item-parser):
(org-element-example-block-parser):
(org-element-src-block-parser):
(org-element-link-parser):
(org-element-timestamp-parser): Do not preserve text properties in the
element property values that are strings.
* lisp/org.el (org-mode-tags-syntax-table): New variable holding
syntax table for tags.
(org-mode): Initialize tag syntax table.
(org-make-tags-matcher): Match tags using appropriate syntax table.
(org-tags-expand): Do no modify syntax table by side effect.
Reported-by: Mattias Engdegård <mattias.engdegard@gmail.com>
Link: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=63225#68
* lisp/org-element.el (org-element-cache-map): Fix when we use
NEXT-RE/FAIL-RE, the initial cached guess is not right, and the
NEXT-RE hits not-yet-cached element.
* lisp/org-clock.el (org-clock-special-range): Prevent returning
previous week for `thisweek' KEY when WSTART is 0. Treat both 0 and 7
as Sunday.
* testing/lisp/test-org-clock.el (test-org-clock/special-range): New
test for `thisweek' and various WSTART arguments.
It seems only the case of weeks starting on Monday was tested.
Other variants caused shifted intervals.
Bug report: Marcin Borkowski. What is a week?
Mon, 10 Apr 2023 05:35:44 +0200.
<https://list.orgmode.org/87h6to1ka7.fsf@mbork.pl>
`org-persist-gc' is ran before `org-persist-write-all' and we may not
yet have all the persist files written. Absence of a persist files is
thus not a sign that collection is expired.
This reverts commit e11073d17d.
* lisp/org-persist.el (org-persist-gc): Do not store collections that
lack persist file. Such collection may emerge, for example, when
writing fails or gets blocked for some reason (in
`org-persist-before-write-hook' or when associated with encrypted
file).
Reported-by: Gustavo Barros <gusbrs.2016@gmail.com>
Link: https://orgmode.org/list/CAM9ALR_F4gjWxOVh_Eq14NOC9nF6g0Cv1jAKb=p4FjPvjqSkWQ@mail.gmail.com
* lisp/oc.el (org-cite-list-bibliography-files): When the bibliography
path is relative to the exported file location, keep the path
relative. Resolve relative paths for bibliographies from SETUPFILE as
well.
* testing/lisp/test-oc.el (test-org-cite/list-bibliography-files):
* testing/examples/sub-bib/include-relative-bib.org:
* testing/examples/sub-bib/include-global-bib.org: New test.
Link: https://orgmode.org/list/CAO48Bk_upR4h-xd0YL+FxeKtWvDoqH+Eju6F_Vzds_m6oxBKcg@mail.gmail.com
* lisp/ob-lisp (org-babel-execute:lisp): by default, Sly and Slime
talk with different RPC server programs with different names.
Specialize the prefix of eval-and-grab-output to Slime (prefix swank:)
or to Sly (prefix slynk:).
Both Org bugfix's 52dc48050 (Handle compiler warnings from Emacs
master, 2023-04-27) and Emacs master's 4f0849a9e6d (Remove unused
values in effect context, 2023-04-09) comment out an unused condition
in org-babel-insert-result to silence a compiler warning that's new to
the Emacs master branch. However, the changes are slightly different,
so the next sync to emacs-29 will trigger a conflict when merged to
master.
Takes Emacs's variant to 1) avoid the conflict and 2) undo the
unrelated space change from Org's variant. (52dc48050 improved the
comment slightly, but it's not worth the conflict.)
* lisp/ob-core.el (org-babel-insert-result): Silence
ignored-return-value warnings about calls to side-effect-free
functions in the last clause of `cond` statements whose values are
unused.
Remove unused values in effect context
4f0849a9e6d29e25d23e061bd81bacce9468856d
Mattias Engdegård
Sun Apr 9 16:45:40 2023 +0200
* lisp/ox-odt.el (org-odt--translate-latex-fragments): Reference
`org-odt-with-latex' variable in the warning to direct users what to
do. Only display the warning when there are LaTeX fragments to
convert.
* lisp/org.el (org-delete-indentation): Support active region when the
first line in region is a heading. Clarify in the docstring that
region is ignored with prefix argument.
Reported-by: Spencer Baugh <sbaugh@janestreet.com>
Link: https://orgmode.org/list/ier3552edfr.fsf@janestreet.com
* lisp/org.el (org-time-stamp-formats):
(org-timestamp-formats):
(org-time-stamp-rounding-minutes):
(org-timestamp-rounding-minutes):
(org-time-stamp-custom-formats):
(org-timestamp-custom-formats):
(org-time-stamp):
(org-timestamp):
(org-time-stamp-inactive):
(org-timestamp-inactive):
(org-insert-time-stamp):
(org-insert-timestamp):
(org-toggle-time-stamp-overlays):
(org-toggle-timestamp-overlays):
(org-time-stamp-to-now):
(org-timestamp-to-now):
* lisp/ox.el (org-export-time-stamp-file):
(org-export-timestamp-file): Rename using "timestamp" term, keeping
the old name as alias.
* doc/org-manual.org: Update all the uses, adding #+findex and
#+vindex entries. Keep the alias names for searchability.
Adjust all the callers.
The following "time-stamp" uses are unchanged:
1. `org-time-stamp-format' where obsolete and _different_ function
`org-timestamp-format' still exists.
2. :time-stamp-file property in export INFO plist. Changing this
would be breaking.
3. ORG-NEWS remains unchanged.
* lisp/org-attach.el (org-attach-tag): Avoid error thrown when trying
to set tags before first heading. This is not a proper fix, but it
will at least allow attaching files to Org file. The tagging will be
skipped for now, until setting FILETAGS is supported by Org API.
* lisp/org-macs.el (org-id-uuid): Use built-in `time-convert' instead
of function defined in org-compat. `time-convert' can already be used
as it is available in Emacs 26.
* lisp/ob-core.el (org-babel-insert-result): Comment out unused `cond'
clause. Still leave it there for readability of the logic.
* lisp/org.el (org-agenda-restrict-begin):
(org-agenda-restrict):
(org-agenda-restrict-end): Mark variables from org-agenda.
* lisp/org.el (org--collect-keywords-1): Do not inhibit SETUPFILE in
read-only buffers.
It is unclear what kind of edge case the FIXME is trying to solve.
Yet, there is no reason to inhibit SETUPFILE processing in all the
read-only buffers (which may happen, for example, when refreshing Org
setup with C-c C-c). Finally, if there is some bug in Gnus (if any)
we should better report it instead of using unexplained workaround.
* lisp/org-element.el (org-element--cache-find): Do not collect
hashing statistics. According to
https://orgmode.org/list/87sfffawfe.fsf@localhost, it is quite useful
for a fraction of users, yielding up to 30% recent cache queries being
memoized. Add additional commentary. Also, slight refactoring.
(org-element--cache-hash-nocache):
(org-element--cache-hash-statistics):
(org-element-cache-hash-show-statistics): Remove.
* lisp/org-persist.el (org-persist-write:generic): Do not set
:last-access on write, unless writing first time. Otherwise,
:last-access may never trigger expiration.
Also, do not use [^\000] as a poor-man's replacement for
(rx (or any newline)).
* lisp/ob-core.el (org-babel-src-block-regexp):
* lisp/ob-haskell.el (org-babel-haskell-export-to-lhs):
* lisp/org-compat.el (org-hide-block-toggle-all):
* lisp/org-element.el:
* lisp/org-feed.el (org-feed-read-previous-status):
(org-feed-parse-rss-feed):
(org-feed-parse-rss-entry):
* lisp/ox-org.el (org-org-publish-to-org):
* testing/lisp/test-ob-tangle.el:
* lisp/org.el (org-block-regexp):
(org-clock-drawer-re): Use \(.\|\n\) regexp instead of [^\000].
(org-latex-regexps): Do not try to match \000 inside latex fragments -
we now use parser for this purpose.
Reported-by: Tommy Kelly <tommy.kelly@verilab.com>
Link: https://orgmode.org/list/875yfk9vlv.fsf@localhost
* ob-scheme.el (org-babel-expand-body:scheme,
org-babel-expand-header-arg-vars:scheme): Change the way header
argument :var variables are expanded for for Scheme source blocks. Use
`define' instead of wrapping using `let'.
Wrapping binding definitions using `let' can lead to issues with GNU
Guile and potentially other Scheme dialects. GNU Guile will only get
to the body of the let at evaluation time, not at macro expansion
time. If the let form wraps any imports of libraries that define
macros, then those imported macros are seen too late and their
corresponding forms inside the body of the let are not
expanded. Using `define' to define bindings avoids this problem, at
least in GNU Guile.
For more context see the mailing list discussion at: https://lists.gnu.org/archive/html/emacs-orgmode/2023-03/msg00087.html
TINYCHANGE
Fix build when Org's version changes
* lisp/org/org-macs.el (org--inhibit-version-check): Rename from
'org--built-in-p' and make it a defvar. All users changed.
(Bug#62762)
62e4eb8fcf71a852117b372809bd4a7953f9b679
Eli Zaretskii
Mon Apr 24 14:16:05 2023 +0300
* doc/org-manual.org (Exporting): Add cindex entry for both "backend"
and "back-end" for better searchability.
All other changes are trivial.
Note that `org-element-export-snippet-parser' will still use :back-end
property. So will ox.el in INFO channel.
* lisp/ox-org.el (org-org-publish-to-org): Remove setting buffer
modification status.
There is no clear reason why it should be done and none of the tests
are failing without this statement. Can be re-added if a real problem
is encountered, with appropriate explanation in the comment.
* lisp/org-fold.el (org-fold-catch-invisible-edits-commands): New
custom option.
(org-fold-catch-invisible-edits): Mention the new custom option in the
docstring.
(org-fold-check-before-invisible-edit-maybe): New function checking
if edits are safe for `this-command'.
(org-fold--advice-edit-commands): New function advising the functions
with `org-fold-check-before-invisible-edit-maybe'.
* lisp/org.el (org-mode): Advice functions on Org startup.
(org-self-insert-command):
(org-delete-backward-char):
(org-delete-char):
(org-meta-return): Do not call `org-fold-check-before-invisible-edit'
and rely on the new advise mechanism instead.
* etc/ORG-NEWS (Commands affected by ~org-fold-catch-invisible-edits~
can now be customized): Announce the change.
* doc/org-manual.org (Catching invisible edits): Mention new
customization.
* lisp/oc.el (org-cite-activate): Discard match data set by activate
function. We already move the point after citation by force, so can
as well ensure that match-data remain accurate for later font-lock
use.
Reported-by: Lin Jian <me@linj.tech>
Link: https://orgmode.org/list/87sfczztg6.fsf@linj.tech
(org-agenda-highlight-todo): Use `propertize' that copies the argument
string instead of `org-add-props'. Otherwise, if the compiler
optimizes string constants into a single object, the properties may
populate in unexpected places.
Reported-by: Gustavo Barros <gusbrs.2016@gmail.com>
Link: https://orgmode.org/list/CAM9ALR95F_ZHV2_WsqAz0-35-S2rwxbHqsA5VGftvq51Yz3ZAQ@mail.gmail.com
Moves the calls to `org-icalendar-fold-string' out of
`org-icalendar--vevent' and `org-icalendar--vtodo', instead calling it
from `org-icalendar--vcalendar'. This way, the string folding applies
to the entire VCALENDAR rather than just the VEVENT or VTODO, as
required by the iCalendar spec.
See also:
https://list.orgmode.org/m2pm85jyw1.fsf@eglen.org.uk/T/#md08cfd0209f6332a1a0c6a54f08544006ff93aed
* lisp/ox-icalendar.el (org-icalendar--vevent, org-icalendar--vtodo):
Remove call to `org-icalendar-fold-string'.
(org-icalendar--vcalendar): Add call to `org-icalendar-fold-string'.
* lisp/org-macs.el (org--built-in-p): New constant indicating if Org
source is a part of Emacs source tree.
(org-assert-version): Skip check when Org is built-in.
* lisp/ox.el (org-export--prune-tree): If the removed object has
trailing spaces and previous object does not have, keep the trailing
spaces.
* etc/ORG-NEWS (Blank lines after removed objects are not retained
during export): Document the change.
Reported-by: Andrea Lazzarini <andrea.lazzarini1@gmail.com>
Link: https://orgmode.org/list/87o7p7z9k3.fsf@localhost
* lisp/ob-tangle.el (org-babel-tangle-default-file-mode): The docstring
states that it the default value is read-write for the user, read only
for everyone else. This is not consistent with the default value which
is actually read-*execute* for the user, read only for everyone else.
The default value is updated to be read-write for the user, as stated in
the docstring.
Reported-by: Ruijie Yu <ruijie@netyu.xyz>
Link: https://list.orgmode.org/orgmode/sdv3554xtat.fsf@fw.net.yu
* lisp/ob-tangle.el (org-babel-interpret-file-mode): When specifying a
file mode in the "rw-r--r--" style, this is interpreted by splicing the
user, group, and other compenents into a "u=rw,g=r,o=r" style string and
applying `file-modes-symbolic-to-number`. For correct interpretation,
we need to ensure the dashes are removed in this process.
* lisp/org-element.el (org-element--cache-self-verify): Switch to nil
by default, disabling costly cache self-verification. This is
relatively safe as no relevant bug reports have been submitted for a
while. Also, the most critical cache failures are still monitored.
* lisp/org.el (org-image-max-width): New custom variable controlling
max inline image width.
(org--create-inline-image): Use the new variable.
* doc/org-manual.org (Images):
* etc/ORG-NEWS (New customization ~org-image-max-width~ limiting the
displayed inline image width): Document the new variable.
* org-table.el (org-self-insert-command): Use `last-command-event'
instead of `last-input-event'. Using `last-input-event' causes
problems in the presence of `evil-escape'. Consider a buffer that has
`orgtbl-mode' enabled while evil-escape is in use. Assume the
evil-escape sequence is "fd". Typing "f RET" will instead insert "RET
RET" into the buffer, since `last-input-event' is "RET", but
`last-command-event' is "f".
TINYCHANGE
* lisp/org-habit.el (org-habit-get-priority): Rename to
`org-habit-get-urgency'.
* lisp/org-compat.el (org-habit-get-priority): Obsolete the old name.
* lisp/org-agenda.el (org-agenda-sorting-strategy):
* lisp/org-agenda.el (org-search-view):
(org-agenda-get-todos):
(org-agenda-get-timestamps):
(org-agenda-get-progress):
(org-agenda-get-deadlines):
(org-agenda-get-scheduled):
(org-agenda-get-blocks):
(org-entries-lessp): Alter priority-up and priority-down sorting
strategies to only sort by actual priority. Rename the previous
composite sorting rank to urgency, corresponding to the new urgency-up
and urgency-down sorting strategies. Store the new rank in 'urgency
text property. Update the docstrings accordingly. Use `urgency-down'
in place of `priority-down' in the default sorting strategy.
(org-set-sorting-strategy): Use `urgency-down' in place of
`priority-down'.
* etc/ORG-NEWS ("Priority" used to sort items in agenda is renamed to
"urgency"): Document the change.
* doc/org-manual.org (Sorting of agenda items): Update manual using
"urgency" term in place of confusing "priority".
Thanks to Samuel Wales for the idea how to fix the inconsistency.
Reported-by: Jonas Olofsson <jonas.olofsson@apple.com>
Link: https://orgmode.org/list/26396316-1201-4D88-9D81-C87DDDA8885A@apple.com
* lisp/ox-html.el (org-html--id-attr-prefix):
* lisp/ox-odt.el (org-odt--id-attr-prefix): Refactor hardcoded "ID-"
as a new private variable.
* org/ox-html.el (org-html--reference): Add support for the ID
property (previously only supported CUSTOM_ID).
These changes make crossfile ID links work in the exported HTML when
`org-html-prefer-user-labels' is not-nil.
Fixes issue where the ox-icalendar export uses an inconsistent mix of
dos and unix style line endings.
* lisp/ox-icalendar.el (org-icalendar-fold-string): No longer converts
to CRLF, instead delegating that to `org-icalendar--post-process-file'.
(org-icalendar--post-process-file): New function to handle exported
file post-processing. Converts EOL to CRLF, and then runs
`org-icalendar-after-save-hook'.
(org-icalendar-export-to-ics, org-icalendar-export-current-agenda,
org-icalendar--combine-files): Call `org-icalendar--post-process-file'
instead of running `org-icalendar-after-save-hook' directly.
* testing/lisp/test-ox-icalendar.el: New file for unit tests of
ox-icalendar. Add an initial test for CRLF line endings.
See also:
https://list.orgmode.org/87o7oetneo.fsf@localhost/T/#m3e3eb80f9fc51ba75854b33ebfe9ecdefa2ded24https://list.orgmode.org/orgmode/87ilgljv6i.fsf@localhost/
* lisp/ob-exp.el (org-babel-exp-results): Display warning when babel
does not know how to execute exported code block and :results
demands results to be exported.
* lisp/ob-exp.el (org-babel-exp-process-buffer): Do not use
`save-window-excursion'.
* lisp/ox.el (org-export-dispatch):
(org-export--dispatch-ui): Use `save-window-excursion' only when
displaying the dispatch interface.
* lisp/org.el (org-fontify-meta-lines-and-blocks-1): Allow fontifying
src blocks with language specification missing.
* lisp/org-src.el (org-src-block-faces): Clarify in the docstring that
"" language name referes to src blocks with no language. Remove
redundant :version tag.
* etc/ORG-NEWS (~org-src-block-faces~ now accepts empty string ~""~ as
language name): Announce the change.
Link: https://orgmode.org/list/99388ff3711696091f0312a5c3f0bcbd@kragelund.me
* lisp/ob-haskell.el (org-babel-haskell-export-to-lhs): Use
non-interactive `insert-file-contents' + `write-region' to avoid
triggering various interactive hooks. Ensure that temp files are
always deleted.
* lisp/org-agenda.el (org-agenda-write):
* lisp/org-table.el: Simplify code using `write-region'.
* lisp/ox-odt.el (org-odt-template): Use `insert-file-contents' +
`write-region' instead of `find-file-noselect' that may trigger
various hooks. The new approach makes `revert-buffer' not
necessary (and do not trigger `revert-buffer' hooks). Also, the
problem with backups will no longer exists.
Original idea: https://list.orgmode.org/orgmode/20221002035931.12191-1-dafydd.lukes@gmail.com/
* lisp/ob-comint.el (org-babel-comint-with-output): Do not try to
filter out prompts in `comint-output-filter-functions'. The prompts
may arrive there arbitrarily - multiple prompts together, partial
prompts, full prompts, etc. For example "ghci> " prompt may arrive as
"gh" + "ci> " with second part still matching `comint-prompt-regexp'.
As a result, if we keep using `comint-output-filter-functions', the
split prompts may sometimes retain their initial part, littering the
results. Now, we postpone filtering to after receiving the output,
still making sure that agglomerated prompts gets filtered using a
custom regexp derived from `comint-prompt-regexp'.
* lisp/ob-comint.el (org-babel-comint-with-output): Consider that
comint can sometimes agglomerate multiple prompts together even within
a single output increment as passed to
`comint-output-filter-functions'.
Example in GHC comint buffer:
GHCi, version 9.0.2: https://www.haskell.org/ghc/ :? for help
ghci> ghci> :{
main :: IO ()
main = putStrLn "Hello World!"
:}
main
"org-babel-haskell-eoe"
ghci| ghci| ghci| ghci> ghci> Hello World!
ghci> "org-babel-haskell-eoe"
ghci>
* lisp/ob-python.el (python-shell-buffer-name): Remove unneeded
defvar.
(org-babel-python-initiate-session-by-key): Check if session already
existed before run-python. Only wait for initialization if it's a
newly started session. Also simplify the code a bit by combining
multiple setq and let statements into a single let statement. Also
add a comment about why adding to `python-shell-first-prompt-hook'
after `run-python' should be safe from race conditions.
* lisp/org.el (org-remove-timestamp-with-keyword): Use `delete-char'
instead of for-interactive-use-only `backward-delete-char'.
(org-fast-tag-selection): Remove unnecessary `condition-case'.
(org-delete-backward-char): Make it explicit that we fall back to
normal interactive call.
Link: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=62518
* lisp/org-macs.el (org-sxhash-safe): New function to calculate object
hashes. Unlike `sxhash', the new function ensures that
if (= (org-sxhash-safe A) (org-sxhash-safe b)) then (equal A B)
Not just the opposite.
* lisp/ob-core.el (org-babel-temp-stable-file): Use the new function.
Reported-by: Thomas Worthington <thomas.worthington@cosector.com>
Link: https://orgmode.org/list/877cwrcdh2.fsf@localhost
* lisp/org-fold.el (org-fold-show-children): Clarify the docstring
emphasizing that direct children are always displayed. Fix
grandchildren being displayed when the first child has deeper level
than the next children.
* lisp/ob-comint.el (org-babel-comint-with-output): Do not try to
filter out prompts in `comint-output-filter-functions'. The prompts
may arrive there arbitrarily - multiple prompts together, partial
prompts, full prompts, etc. For example "ghci> " prompt may arrive as
"gh" + "ci> " with second part still matching `comint-prompt-regexp'.
As a result, if we keep using `comint-output-filter-functions', the
split prompts may sometimes retain their initial part, littering the
results. Now, we postpone filtering to after receiving the output,
still making sure that agglomerated prompts gets filtered using a
custom regexp derived from `comint-prompt-regexp'.
* lisp/ob-comint.el (org-babel-comint-with-output): Consider that
comint can sometimes agglomerate multiple prompts together even within
a single output increment as passed to
`comint-output-filter-functions'.
Example in GHC comint buffer:
GHCi, version 9.0.2: https://www.haskell.org/ghc/ :? for help
ghci> ghci> :{
main :: IO ()
main = putStrLn "Hello World!"
:}
main
"org-babel-haskell-eoe"
ghci| ghci| ghci| ghci> ghci> Hello World!
ghci> "org-babel-haskell-eoe"
ghci>
* lisp/org-clock.el (org-logind-dbus-session-path): Do not try to use
dbus when `dbus-call-method' errs.
`dbus-call-method' may throw
Debugger entered--Lisp error: (dbus-error "org.freedesktop.login1.NoSessionForPID" "PID 7361 does not belong to any known session")
Do not try to use dbus method in such scenario.
Link: https://builds.sr.ht/~bzg/job/961763
* lisp/org.el (org-at-date-range-p):
(org--math-p):
(org-first-sibling-p): Remove interactive spec.
Predicates like `org-first-sibling-p' are no use when called
interactively, and should not appear in the M-x prompt.
TINYCHANGE
* lisp/org-timer.el (org-logind-dbus-session-path): New variable.
(org-logind-user-idle-seconds): New function.
(org-user-idle-seconds): Use them.
* etc/ORG-NEWS (Add support for ~logind~ idle time in
~org-user-idle-seconds~): Document the new feature.
* ob-shell.el (org-babel-sh-evaluate): Add condition for async within
session. Allow :async header argument to be either t or blank.
* test-ob-shell.el:
(test-ob-shell/session-async-valid-header-arg-values): Check that
:async header works for both t and blank values.
(test-ob-shell/session-async-inserts-uuid-before-results-are-returned):
Check that UUID is used as placeholder until results return.
(test-ob-shell/session-async-evaluation): Check that asynchronously
evaluated results are eventually placed in the buffer.
Link: https://list.orgmode.org/186283d230a.129f5feb61660123.3289004102603503414@excalamus.com/
* ob-R.el (ob-session-async-org-babel-R-evaluate-session): Use
`org-id-uuid' instead of `md5' as results placeholder.
* ob-python.el (org-babel-python-async-evaluate-session): Use
`org-id-uuid' instead of `md5' as results placeholder.
TEC
