Ruijie Yu via "General discussions about Org-mode." <emacs-orgmode@gnu.org>
writes:
> Ihor Radchenko <yanta...@posteo.net> writes:
>
>> (let ((org-time-stamp-formats '("%Y-%m-%d" . "%Y-%m-%d %H:%M")))
>> (org-element-timestamp-interpreter timestamp nil))
>
> Thanks for the pointer. I have made this into a macro and redid my
> changes accordingly. Please also let me know if more work should be
> dedicated to avoid code duplication in tests (I see a lot of duplication
> in the portions I touched in testing/lisp/test-org.el).
>
> What currently still fails unexpectedly:
>
> zh_CN.UTF-8:
> test-org-clock/clocktable/lang
> - This test only fails in batch mode, as I mentioned up-thread. I
> might try to convert this into a test that expects failure
> depending on whether Emacs is in batch mode.
> test-org-colview/columns-width
> - This patch did not attempt to fix it. I posted a diff file under
> v2 (not included in this patch), but that didn't work either.
>
> ja_JA.UTF-8:
> test-org-colview/columns-width
> - See above.
>
> fr_FR.UTF-8:
> test-org-clok/org-clock-timestamps-change
> - This is tracked in the other thread I created, and won't be fixed
> by this patch.
>
> en_US.UTF-8 and de_DE.UTF-8:
> No unexpected test failures.
>
> [2. text/x-patch;
> 0001-DRAFT-Fix-dependence-on-locale-in-org-testing-facili.patch]...
>
>
> To summarize, the remaining goal of this patchset is to fix
> `test-org-colview/columns-width', and to optionally expect-failure on
> `test-org-clock/clocktable/lang'.
Iteration v6. Everything, other than the French test and the
batch-non-batch Chinese test, now passes. I'll look up how to determine
whether we are in batch mode or not, and slap that onto the expect tag
for `test-org-clock/clocktable/lang'. And the French test belongs to
its own thread, so I won't try to fix it here.
V6 has two commits: the first commit is the *exact copy* of v5, and the
second commit tries to fix the zh_CN columns-width code & test. Hence I
only attach the patch file for the second commit.
>From 59cbb93b9fc221bdc8ee708b05943a245c41ad25 Mon Sep 17 00:00:00 2001
From: Ruijie Yu <rui...@netyu.xyz>
Date: Tue, 25 Apr 2023 22:56:02 +0800
Subject: [PATCH 2/2] Let org-columns correctly detect string-widths in code
TODO: maybe I should also make a test directly on
`org-columns-add-ellipses'. Will do in next iteration unless
objections.
* lisp/org-colview.el (org-columns--truncate-below-width): add a
helper function that will trim off just enough data from string to
fit into expected width.
(org-columns-add-ellipses): make sure to do truncation correctly
even in CJK locales (where an ellipsis character takes two
spaces).
* testing/lisp/test-org-colview.el
(test-org-colview/substring-below-width): add test to make sure
helper function is correct.
(test-org-colview/columns-width): fix incorrect expectations for
CJK locales about ellipses.
---
lisp/org-colview.el | 26 +++++++++++++++++++++-----
testing/lisp/test-org-colview.el | 19 ++++++++++++++++++-
2 files changed, 39 insertions(+), 6 deletions(-)
diff --git a/lisp/org-colview.el b/lisp/org-colview.el
index 92a3b473d..6311a9274 100644
--- a/lisp/org-colview.el
+++ b/lisp/org-colview.el
@@ -452,14 +452,30 @@ DATELINE is non-nil when the face used should be
"Type \\<org-columns-map>`\\[org-columns-edit-value]' \
to edit property")))))))
+(defun org-columns--truncate-below-width (string width)
+ "Return a substring of STRING no wider than WIDTH.
+This substring must start at 0, and must be the longest possible
+substring whose `string-width' does not exceed WIDTH."
+ (declare (side-effect-free t))
+ (let ((end (min width (length string))) res)
+ (while (and end (>= end 0))
+ (let* ((curr (string-width string 0 end))
+ (excess (- curr width)))
+ (if (> excess 0)
+ (cl-decf end (max 1 (/ excess 2)))
+ (setq res (substring string 0 end) end nil))))
+ res))
+
(defun org-columns-add-ellipses (string width)
"Truncate STRING with WIDTH characters, with ellipses."
(cond
- ((<= (length string) width) string)
- ((<= width (length org-columns-ellipses))
- (substring org-columns-ellipses 0 width))
- (t (concat (substring string 0 (- width (length org-columns-ellipses)))
- org-columns-ellipses))))
+ ((<= (string-width string) width) string)
+ ((<= width (string-width org-columns-ellipses))
+ (org-columns--truncate-below-width org-columns-ellipses width))
+ (t (concat
+ (org-columns--truncate-below-width
+ string (- width (string-width org-columns-ellipses)))
+ org-columns-ellipses))))
(defvar org-columns-full-header-line-format nil
"The full header line format, will be shifted by horizontal scrolling." )
diff --git a/testing/lisp/test-org-colview.el b/testing/lisp/test-org-colview.el
index 9daec18e2..b421fdc41 100644
--- a/testing/lisp/test-org-colview.el
+++ b/testing/lisp/test-org-colview.el
@@ -26,6 +26,23 @@
(require 'org-duration)
(require 'org-inlinetask)
+(ert-deftest test-org-colview/substring-below-width ()
+ "Test `org-columns--truncate-below-width'."
+ (cl-flet ((check (string width expect)
+ (string= expect (org-columns--truncate-below-width
+ string width))))
+ (if (= (char-width ?…) 2)
+ (progn (should (check "12…" 3 "12"))
+ (should (check "1…2" 1 "1"))
+ (should (check "1…2" 2 "1"))
+ (should (check "1…2" 3 "1…"))
+ (should (check "……………………" 7 "………")))
+ (progn (should (check "12…" 4 "12…"))
+ (should (check "1…2" 1 "1"))
+ (should (check "1…2" 2 "1…"))
+ (should (check "1…2" 3 "1…2"))
+ (should (check "……………………" 7 "…………………"))))))
+
(ert-deftest test-org-colview/get-format ()
"Test `org-columns-get-format' specifications."
;; Without any clue, use `org-columns-default-format'.
@@ -160,7 +177,7 @@
(org-columns))
(org-trim (get-char-property (point) 'display)))))
(should
- (equal "1234… |"
+ (equal (if (= 1 (char-width ?…)) "1234… |" "123… |")
(org-test-with-temp-text "* H\n:PROPERTIES:\n:P: 123456\n:END:"
(let ((org-columns-default-format "%5P")
(org-columns-ellipses "…"))
--
2.40.0
Comments welcome, thanks.
--
Best,
RY
[Please note that this mail might go to spam due to some
misconfiguration in my mail server -- still investigating.]
