Re: bin/60496: Locale-dependent UTF-8 issue over SSH session

"Valery Ushakov via gnats" <[email protected]> Sat, 25 Jul 2026 22:35:03 +0000 (UTC)
Newsgroups gmane.os.netbsd.bugs
Message-ID <[email protected]>
The following reply was made to PR bin/60496; it has been noted by GNATS.

From: Valery Ushakov <[email protected]>
To: [email protected]
Cc: 
Subject: Re: bin/60496: Locale-dependent UTF-8 issue over SSH session
Date: Sun, 26 Jul 2026 01:33:52 +0300

 On Fri, Jul 24, 2026 at 15:10:00 +0000, [email protected] via gnats wrote:
 
 > The separator should be a space in Norwegian locale (nb_NO.UTF-8).
 > Instead, it sends &#65533;.
 >
 > $ LC_ALL=nb_NO.UTF-8 printf "%'d\n" 12345 | hexdump
 > 0000000 3231 33c2 3534 000a
 > 0000007
 
 It sends 0xc2 byte, which makes an invalid UTF-8 sequence.
 
 The locale source (share/locale/numeric/nb_NO.UTF-8.src in this case)
 has \ua0 non-breaking space (as UTF-8 \xc2\xa0)
 
 The bug is that thousands_sep in vfwprintf.c is char, not a string,
 hence truncated 0xc2.
 
 -uwe