bug#73985: fix: charset.max_char may be used before being set

bug#73985: fix: charset.max_char may be used before being set

[altermo31 via Bug reports for GNU Emacs, the Swiss army knife of text editors]

[-- Attachment #1: Type: text/plain, Size: 227 bytes --]

Problem:
If the value to be set to charset.max_char is invalid, an error occurs, but the error uses the yet unset charset.max_char.

Solution:
Use the value that charset.max_char would be set to if the value wasn't invalid.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-fix-charset.max_char-may-be-used-before-being-set.patch --]
[-- Type: text/x-patch; name=0001-fix-charset.max_char-may-be-used-before-being-set.patch, Size: 910 bytes --]

From 64e9557c5f6ae4416a1a12fd6f68f0d19931bb99 Mon Sep 17 00:00:00 2001
From: altermo <107814000+altermo@users.noreply.github.com>
Date: Thu, 24 Oct 2024 13:39:06 +0200
Subject: [PATCH] fix: charset.max_char may be used before being set

---
 src/charset.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/charset.c b/src/charset.c
index e8d0826..14b33d2 100644
--- a/src/charset.c
+++ b/src/charset.c
@@ -1007,7 +1007,7 @@ DEFUN ("define-charset-internal", Fdefine_charset_internal,
 
       i = CODE_POINT_TO_INDEX (&charset, charset.max_code);
       if (MAX_CHAR - charset.code_offset < i)
-	error ("Unsupported max char: %d", charset.max_char);
+	error ("Unsupported max char: %d", i + charset.code_offset);
       charset.max_char = i + charset.code_offset;
       i = CODE_POINT_TO_INDEX (&charset, charset.min_code);
       charset.min_char = i + charset.code_offset;
-- 
2.47.0

bug#73985: fix: charset.max_char may be used before being set

[Eli Zaretskii]

> Date: Thu, 24 Oct 2024 11:46:17 +0000
> From:  altermo31 via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> 
> Problem:
> If the value to be set to charset.max_char is invalid, an error occurs, but the error uses the yet unset charset.max_char.
> 
> Solution:
> Use the value that charset.max_char would be set to if the value wasn't invalid.

Thanks.  You are right that the original code is incorrect, but the
code you propose has a subtle problem: the sum

  i + charset.code_offset

could overflow.  So we need to find a safer way of explaining the
problem.

bug#73985: fix: charset.max_char may be used before being set

[Eli Zaretskii]

> Cc: 73985@debbugs.gnu.org
> Date: Thu, 24 Oct 2024 17:12:13 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> 
> > Date: Thu, 24 Oct 2024 11:46:17 +0000
> > From:  altermo31 via "Bug reports for GNU Emacs,
> >  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> > 
> > Problem:
> > If the value to be set to charset.max_char is invalid, an error occurs, but the error uses the yet unset charset.max_char.
> > 
> > Solution:
> > Use the value that charset.max_char would be set to if the value wasn't invalid.
> 
> Thanks.  You are right that the original code is incorrect, but the
> code you propose has a subtle problem: the sum
> 
>   i + charset.code_offset
> 
> could overflow.  So we need to find a safer way of explaining the
> problem.

I fixed that (on the master branch) like this:

diff --git a/src/charset.c b/src/charset.c
index e8d0826..f7d80cc 100644
--- a/src/charset.c
+++ b/src/charset.c
@@ -1007,7 +1007,8 @@ DEFUN ("define-charset-internal", Fdefine_charset_internal,
 
       i = CODE_POINT_TO_INDEX (&charset, charset.max_code);
       if (MAX_CHAR - charset.code_offset < i)
-	error ("Unsupported max char: %d", charset.max_char);
+	error ("Unsupported max char: %d + %ud > MAX_CHAR (%d)",
+	       i, charset.max_code, MAX_CHAR);
       charset.max_char = i + charset.code_offset;
       i = CODE_POINT_TO_INDEX (&charset, charset.min_code);
       charset.min_char = i + charset.code_offset;