1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
|
org-drill NEWS — user-facing highlights since the fork
=======================================================
For the full chronological log, see CHANGELOG.md. This file
collects the changes a user is most likely to feel.
Since the fork (no tagged release yet)
--------------------------------------
* Several long-standing bugs that users reported upstream are
fixed. If you've ever seen any of these, this fork should help:
- "[Y-08-27 Wed 16:%]"-shaped timestamps in DRILL_LAST_REVIEWED
(issue #59).
- Your default-input-method getting silently set to nil after a
drill session (issues #52, #58 — particularly noticeable for
users with non-Latin input methods).
- Drill entries whose answer lived inside a child sub-heading
being silently skipped (issue #13).
- The cloze face leaking onto unrelated org headings (issue #38).
- "Window system frame should be used" error when running
org-drill in a TTY emacsclient (issue #44 — affects tmux
setups in particular).
- org-drill-final-report not firing after an edit-and-resume
cycle (issue #33).
- "Wrong type argument: hash-table-p, nil" stopping a session
after the first new (no-ID) entry, forcing one-at-a-time
invocations (issue #53).
- Crash on package load when the persist file got corrupted
(issue #45).
- Stray "[debug] org-drill: at marker position N" message logged
on every drilled card.
- Cards with DRILL_CARD_TYPE: translate_number hitting a
void-function error.
* Org 9.6 is now the minimum supported version. org-drill calls
org-fold-show-entry / org-fold-show-subtree (introduced in 9.6)
in seven places without guards — the dependency now matches
what the runtime needs, so a too-old Org will fail at install
time with a clear message rather than a cryptic void-function
error during a session.
* org-drill-resume and org-drill-again now produce a clear
user-error when there's no previous session to resume, instead
of a cryptic eieio type error.
* New M-x targets exposed via the existing keymap haven't been
added; everything is still launched via M-x org-drill,
M-x org-drill-cram, M-x org-drill-resume, etc.
For developers
--------------
* Test coverage went from ~12% to ~78% on org-drill.el. The
tests/ directory now holds ~370 ERT tests across 27 files.
* GitHub Actions CI runs the test matrix (Emacs 28.2 / 29.4 /
snapshot), lint, and coverage on every push to main.
* The Makefile gained `make coverage`, `make compile`,
`make lint`, and `make validate-parens` targets. See the
CONTRIBUTING.md file for the full developer setup.
* `make install` was renamed to `make setup` (it installs Cask
dependencies into a project-local .cask/ directory, not the
package onto the user's system).
|
