svn commit: r362148 - head/contrib/nvi/common

[Yuri Pankov]

Gleb Smirnoff wrote:
>    Yuri, Zhihao,
> 
> this commit totally broke Russian input for me in nvi. After
> exiting edit mode, nvi immediately converts all text to ???????.
> 
> I don't have any special settings in my environment. All I have
> is "russian" class for my user which yields in these environment
> variables:
> 
> declare -x LANG="ru_RU.UTF-8"
> declare -x MM_CHARSET="UTF-8"
> declare -x XTERM_LOCALE="ru_RU.UTF-8"
> 
> I'm already digging into that problem, but may be you have
> a clue immediately.

My bad, yes, I see the problem, looking into it.

> On Sat, Jun 13, 2020 at 02:11:02PM +0000, Yuri Pankov wrote:
> Y> Author: yuripv
> Y> Date: Sat Jun 13 14:11:02 2020
> Y> New Revision: 362148
> Y> URL: https://svnweb.freebsd.org/changeset/base/362148
> Y>
> Y> Log:
> Y>   nvi: fallback to ISO8859-1 as last resort
> Y>
> Y>   Current logic of using user's locale encoding that is UTF-8 doesn't make
> Y>   much sense if we already failed the looks_utf8() check and skipped
> Y>   encoding set using "fileencoding" as being UTF-8 as well; fallback to
> Y>   ISO8859-1 in that case.
> Y>
> Y>   Reviewed by:	Zhihao Yuan <lichray at gmail.com>
> Y>   Differential Revision:	https://reviews.freebsd.org/D24919
> Y>
> Y> Modified:
> Y>   head/contrib/nvi/common/exf.c
> Y>
> Y> Modified: head/contrib/nvi/common/exf.c
> Y> ==============================================================================
> Y> --- head/contrib/nvi/common/exf.c	Sat Jun 13 09:16:07 2020	(r362147)
> Y> +++ head/contrib/nvi/common/exf.c	Sat Jun 13 14:11:02 2020	(r362148)
> Y> @@ -1237,7 +1237,10 @@ file_encinit(SCR *sp)
> Y>  	}
> Y>
> Y>  	/*
> Y> -	 * Detect UTF-8 and fallback to the locale/preset encoding.
> Y> +	 * 1. Check for valid UTF-8.
> Y> +	 * 2. Check if fallback fileencoding is set and is NOT UTF-8.
> Y> +	 * 3. Check if user locale's encoding is NOT UTF-8.
> Y> +	 * 4. Use ISO8859-1 as last resort.
> Y>  	 *
> Y>  	 * XXX
> Y>  	 * A manually set O_FILEENCODING indicates the "fallback
> Y> @@ -1246,9 +1249,13 @@ file_encinit(SCR *sp)
> Y>  	 */
> Y>  	if (looks_utf8(buf, blen) > 1)
> Y>  		o_set(sp, O_FILEENCODING, OS_STRDUP, "utf-8", 0);
> Y> -	else if (!O_ISSET(sp, O_FILEENCODING) ||
> Y> -	    !strcasecmp(O_STR(sp, O_FILEENCODING), "utf-8"))
> Y> +	else if (O_ISSET(sp, O_FILEENCODING) &&
> Y> +	    strcasecmp(O_STR(sp, O_FILEENCODING), "utf-8") != 0)
> Y> +		/* Use fileencoding as is */ ;
> Y> +	else if (strcasecmp(codeset(), "utf-8") != 0)
> Y>  		o_set(sp, O_FILEENCODING, OS_STRDUP, codeset(), 0);
> Y> +	else
> Y> +		o_set(sp, O_FILEENCODING, OS_STRDUP, "iso8859-1", 0);
> Y>
> Y>  	conv_enc(sp, O_FILEENCODING, 0);
> Y>  #endif