bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

[Lars Magne Ingebrigtsen]

If you put point after the following expression and hit `C-x C-e':

?\N{HEAVY CHECK MARK}

you'll get an error saying "(void-variable MARK})".  And in an Emacs
Lisp buffer, using forward/backward sexp commands, they do not skip past
the entire expression.




In GNU Emacs 25.1.50.35 (x86_64-unknown-linux-gnu, GTK+ Version 3.4.2)
 of 2016-04-24 built on stories
Repository revision: a28873ce734b3618b0e8f6892e65db24d2f18da8
Windowing system distributor 'The X.Org Foundation', version 11.0.11204000
System Description:	Debian GNU/Linux 7.9 (wheezy)

Configured features:
XPM JPEG TIFF GIF PNG RSVG IMAGEMAGICK SOUND GSETTINGS NOTIFY GNUTLS
LIBXML2 FREETYPE LIBOTF XFT ZLIB TOOLKIT_SCROLL_BARS GTK3 X11


-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

[Philipp Stephani]

[-- Attachment #1.1: Type: text/plain, Size: 503 bytes --]

Lars Magne Ingebrigtsen <larsi@gnus.org> schrieb am So., 24. Apr. 2016 um
14:17 Uhr:

>
> If you put point after the following expression and hit `C-x C-e':
>
> ?\N{HEAVY CHECK MARK}
>
> you'll get an error saying "(void-variable MARK})".  And in an Emacs
> Lisp buffer, using forward/backward sexp commands, they do not skip past
> the entire expression.
>
>
Thanks, I've attached a patch that should fix eval-print-last-expr. This
doesn't fix backward-sexp, because I think that would be much harder.

[-- Attachment #1.2: Type: text/html, Size: 825 bytes --]

[-- Attachment #2: 0001-Detect-named-character-literals.patch --]
[-- Type: application/octet-stream, Size: 2122 bytes --]

From 62eb07c44f5b6c1af27923b5dd971f4429006806 Mon Sep 17 00:00:00 2001
From: Philipp Stephani <phst@google.com>
Date: Sun, 1 May 2016 14:52:49 +0200
Subject: [PATCH] Detect named character literals.

Fixes Bug#23354.

* lisp/progmodes/elisp-mode.el (elisp--preceding-sexp): Skip over
named character literals.
* test/lisp/progmodes/elisp-mode-tests.el
(elisp--preceding-sexp--char-name): Add test for skipping over
named character literals.
---
 lisp/progmodes/elisp-mode.el            | 11 +++++++++++
 test/lisp/progmodes/elisp-mode-tests.el |  6 ++++++
 2 files changed, 17 insertions(+)

diff --git a/lisp/progmodes/elisp-mode.el b/lisp/progmodes/elisp-mode.el
index ca85980..1c72848 100644
--- a/lisp/progmodes/elisp-mode.el
+++ b/lisp/progmodes/elisp-mode.el
@@ -1051,6 +1051,17 @@ elisp--preceding-sexp
 	      ((or (eq (following-char) ?\')
 		   (eq (preceding-char) ?\'))
 	       (setq left-quote ?\`)))
+
+        ;; When after a named character literal, skip over the entire
+        ;; literal, not only its last word.
+        (when (= (preceding-char) ?})
+          (let ((begin (save-excursion
+                         (backward-char)
+                         (skip-syntax-backward "w-")
+                         (backward-char 3)
+                         (when (looking-at-p "\\\\N{") (point)))))
+            (when begin (goto-char begin))))
+
 	(forward-sexp -1)
 	;; If we were after `?\e' (or similar case),
 	;; use the whole thing, not just the `e'.
diff --git a/test/lisp/progmodes/elisp-mode-tests.el b/test/lisp/progmodes/elisp-mode-tests.el
index 1679af3..a7562a0 100644
--- a/test/lisp/progmodes/elisp-mode-tests.el
+++ b/test/lisp/progmodes/elisp-mode-tests.el
@@ -641,5 +641,11 @@ xref-elisp-overloadable-separate-default
   (elisp--xref-find-definitions (eval '(provide 'stephe-leake-feature)))
   nil)
 
+(ert-deftest elisp--preceding-sexp--char-name ()
+  (with-temp-buffer
+    (emacs-lisp-mode)
+    (insert "?\\N{HEAVY CHECK MARK}")
+    (should (equal (elisp--preceding-sexp) ?\N{HEAVY CHECK MARK}))))
+
 (provide 'elisp-mode-tests)
 ;;; elisp-mode-tests.el ends here
-- 
2.8.1

bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

[Lars Ingebrigtsen]

Philipp Stephani <p.stephani2@gmail.com> writes:

> Thanks, I've attached a patch that should fix eval-print-last-expr. This doesn't
> fix backward-sexp, because I think that would be much harder. 

Thanks; applied.  I hope somebody else can fix backward-sexp and
friends.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Philipp Stephani <p.stephani2@gmail.com> writes:
>
>> Thanks, I've attached a patch that should fix
>> eval-print-last-expr. This doesn't
>> fix backward-sexp, because I think that would be much harder. 
>
> Thanks; applied.  I hope somebody else can fix backward-sexp and
> friends.  :-)

I just checked whether things has changed -- and, no, backward-sexp and
friends still do the wrong thing on Unicode literals like

?\N{HEAVY CHECK MARK}

in Emacs Lisp buffers.  In particular, backward-sexp after the final "}"
skips to the start of "MARK", and forward-sexp at the start skips to the
end of "HEAVY".

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#23354: 25.1.50; Unicode literals aren't handled correctly by some commands

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I just checked whether things has changed -- and, no, backward-sexp and
> friends still do the wrong thing on Unicode literals like
>
> ?\N{HEAVY CHECK MARK}
>
> in Emacs Lisp buffers.  In particular, backward-sexp after the final "}"
> skips to the start of "MARK", and forward-sexp at the start skips to the
> end of "HEAVY".

I've now fixed this in Emacs 29.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no