* lisp/ob-ocaml.el (org-babel-execute:ocaml): Support for
pretty-printed outputs.
E.g. the following snippet produces multiple lines of output, due to
pretty-printing
#+BEGIN_SRC ocaml :exports code :eval no-export :results verbatim
(* Note, no need to have parentheses around sqrt, as
application binds the tightest *)
let rec gamma n = if n = 0
then (1.0 +. sqrt 5.0) /. 2.0
else 1.0 /. (gamma (n-1) -. 1.0)
let nums = List.map gamma (List.init 5 (fun x -> x))
#+END_SRC
#+NAME: gamma-tbl
#+RESULTS[1fd6a3e846afdef51350eb6d7ba15c6844ccc14e]:
: val gamma : int -> float = <fun>
: val nums : float list =
: [1.6180339887498949; 1.61803398874989468; 1.61803398874989535;
: 1.61803398874989357; 1.61803398874989823]
(In reality this would have 50 or so outputs, point being mathematically
it shouldn't change, but due to floating point errors it does.)
The regexp has the following problems:
- The "." in group 5 (as-was) doesn't match new-lines, so it only
matches the first line, e.g. as a list
: - [1.6180339887498949, 1.61803398874989468, 1.61803398874989535,
- When using "\\(.\\|\n\\)", it includes the starting indentation of the
list, which makes org-babel-script-escape choke (called from
org-babel-ocaml-read-list, from org-babel-ocaml-parse-output, from
org-babel-execute:ocaml, the code being modified.
TINYCHANGE
* lisp/ox-html.el (org-html-format-code, org-html-do-format-code):
Use new export option :html-wrap-src-lines with variable
org-html-wrap-src-lines to control whether source code lines should
be wrapped in code elements or not.
* doc/org-manual.org: Document the new option
* lisp/ox-publish.el (org-publish-resolve-external-link): Signal
`org-link-broken' when `org-link-search' fails to resolve a link.
Without this change, a broken fuzzy link will always abort the export
even when `org-export-with-broken-links' is non-nil.
* lisp/org.el (org-doi-server-url): Prefer https: to http: un URLs.
Update some URLs
946a56a10fed769646a8b7c4ebc53f53c84be896
Paul Eggert
Mon Sep 23 00:12:52 2019 -0700
Notes(km):
- It's not mentioned in 946a56a10's message, but that commit also
drops the dx subdomain, an "earlier syntax which continues to be
supported" according to https://www.doi.org/factsheets/DOI_PURL.html.
- 946a56a10 replaced a link in the manual with https://api.uva.nl/
(which the previous link redirects to). Here we instead use the
replacement link from master's f4083eefd (manual: Fix URL,
2019-06-01).
* org.el (org-activate-links): `match-beginning' and `match-end` should
be called shortly after `re-search-forward'. Otherwise, they may return
values corresponding to a different invocation of `re-search-forward'.
TINYCHANGE
* lisp/org-macs.el (org-do-remove-indentation): Handle full-width characters.
When the least indented line in buffer begins from fullwidth character
like 'あ', a Japanese character, `org-do-remove-indentation' doesn't
remove indentation at all or removes improperly.
e.g. 'あ' with 2 spaces indent
-----[buffer begin]-----
あ
-----[buffer end]-----
`org-do-remove-indentation' does nothing in this buffer.
Expected result is:
-----[buffer begin]-----
あ
-----[buffer end]-----
Reported-by: Yuichiro Hakozaki <sankaku.git@gmail.com>
<http://lists.gnu.org/r/emacs-orgmode/2019-09/msg00075.html>
* lisp/ob-lisp.el (org-babel-execute:lisp): Make sure that when
a singular string value is returned, it has no quotation marks around
it.
Before this change, the code below would produce an incorrect result:
#+BEGIN_SRC lisp :results file
"plot.png"
#+END_SRC
#+RESULTS:
[[file:"plot.png"]]
After the change:
#+BEGIN_SRC lisp :results file
"plot.png"
#+END_SRC
#+RESULTS:
[[file:plot.png]]
TINYCHANGE
Ref. mail "[O] git-annex-related org-attach tests failing on
master"
https://lists.gnu.org/archive/html/emacs-orgmode/2019-09/msg00030.html
* lisp/org-attach-git.el (org-attach-git-commit): Add optional
argument to function contract to make it work with
`org-attach-after-change-hook'. Even though that argument is
not used in the actual code (due to legacy functionality).
* testing/lisp/test-org-attach.el
(test-org-attach/dired-attach-to-next-best-subtree/1)
(test-org-attach/dired-attach-to-next-best-subtree/2):
Modify tests to also work if user has git-annex installed
* testing/lisp/test-org-attach-git.el
(test-org-attach-git/use-annex): Correct errors from previous
commit, where git-annex was refactored out from org-attach
into it's separate module.
In commit ae9cd4370 the filsystem check for org-attach-dir was
removed. Adding it back here, together with an optional flag to
override the check.
To satisfy compatabilty issues with org-brain. Ref. thread here:
- https://github.com/Kungsgeten/org-brain/pull/203
With nanoseconds precision in the timestamp risk for duplicates shall
be ... near 0 ... even in heavy teamwork-environments or heavily
scripted setups.
Jens Lechtenboerger
Nicolas Goaziou
Kovacsics Robert
Anton-Latukha
Robert Irelan
Kyle Meyer
Paul Eggert
Juanma Barranquero
Gustav Wikström
Marco Wahl
Mario Schlegel
Max Mouratov
Mike Ivanov
Carsten Dominik
Michaël Cadilhac
Bastien