Re: [PATCH] [gdb/cli] Don't emit emojis in MI

[Tom Tromey <tom@tromey.com>]

>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:

Tom> I think we probably should encode unicode characters in MI error strings as
Tom> octal escapes, independent of the sevenbit-strings setting.  This patch does
Tom> not address this part.

I think there are some other bugs concerning this FWIW.

I tend to think MI should not set sevenbit-strings and should just use
the host encoding.  That way, clients could simply use UTF-8 and not
have to deal with the oddities here.

sevenbit-strings also seems pretty questionable to me.  Like maybe it
made sense in the ancient times before the charset code was written, but
nowadays I think it's just a mistake.

Tom> As for implementation, I introduced a function emoji_allowed alongside
Tom> can_emit_style_escape, which defaults to the value of can_emit_style_escape.

Seems reasonable though:

Tom> +  /* True if emojis are allowed on STREAM.  */
Tom> +  virtual bool emoji_allowed ()
Tom> +  {
Tom> +    /* By default, assume that emojis are not allowed on streams that don't
Tom> +       support ANSI escapes.  */
Tom> +    return can_emit_style_escape ();
Tom> +  }

...there's no present need for 'virtual' here, and if we do ever need
it, it is easy to add.  So I would suggest removing that.

Approved-By: Tom Tromey <tom@tromey.com>

Tom