bug#58168: string-lessp glitches and inconsistencies

["Mattias Engdegård" <mattias.engdegard@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1149 bytes --]

1 okt. 2022 kl. 12.02 skrev Lars Ingebrigtsen <larsi@gnus.org>:

> Funnily enough, the latter displays in a different way for me, which may
> or may not be a bug:
> 
> This is with `display-raw-bytes-as-hex' t.

You are right, that is completely broken -- display-raw-bytes-as-hex shouldn't affect the display of C1 controls.
Whether (string 128) displays "\200" or "\x80", however tarted up in a fancy face, it's still a lie. Only something like "\u0080" would actually be correct.

It seems to be a relic from the pre-Unicode days of Emacs: the code responsible muddles the display of raw bytes and unicode controls.
The attached patch untangles the two somewhat and lets display-raw-bytes-as-hex do what its name and documentation suggest, while using a non-confusing display for C1 controls.

The command

  (insert "C1: " (string 128) " raw: " (unibyte-string 128) ".\n")

currently displays

   C1: \200 raw: \200.
or
   C1: \x80 raw: \x80.

depending on display-raw-bytes-as-hex. With the patch, we get

€ €  €C1: \u0080 raw: \200.
or
   C1: \u0080 raw: \x80.

which should satisfy everyone. What about it?


[-- Attachment #2: unicode-escape-display.diff --]
[-- Type: application/octet-stream, Size: 955 bytes --]

diff --git a/src/xdisp.c b/src/xdisp.c
index 55e74a3603..fa4fc2319e 100644
--- a/src/xdisp.c
+++ b/src/xdisp.c
@@ -8179,12 +8179,20 @@ get_next_display_element (struct it *it)
 		char str[10];
 		int len, i;
 
+		const char *format_string;
 		if (CHAR_BYTE8_P (c))
-		  /* Display \200 or \x80 instead of \17777600.  */
-		  c = CHAR_TO_BYTE8 (c);
-		const char *format_string = display_raw_bytes_as_hex
-					    ? "x%02x"
-					    : "%03o";
+		  {
+		    /* A raw byte: display using an octal or hex escape which
+		       would produce this byte in a Lisp string literal.  */
+		    c = CHAR_TO_BYTE8 (c);
+		    format_string = display_raw_bytes_as_hex ? "x%02x" : "%03o";
+		  }
+		else
+		  {
+		    /* A Unicode character not displayed in any other way:
+		       use a Unicode escape.  */
+		    format_string = c <= 0xffff ? "u%04X" : "U%08X";
+		  }
 		len = sprintf (str, format_string, c + 0u);
 
 		XSETINT (it->ctl_chars[0], escape_glyph);

[-- Attachment #3: Type: text/plain, Size: 203 bytes --]



I see that the redisplay-testsuite.el needs amending too; it actually looks buggy in this respect. If the above approach is deemed acceptable, I'll submit a patch that includes that file as well.