Re: [PATCH] localedata: fix LC_ADDRESS.country_car entries

[Mike Frysinger <vapier at gentoo dot org>]

On 15 Apr 2016 12:13, keld@keldix.com wrote:
> On Fri, Apr 15, 2016 at 12:47:28AM -0400, Carlos O'Donell wrote:
> > On 04/12/2016 11:44 AM, Mike Frysinger wrote:
> > > On 12 Apr 2016 10:33, Chris Leonard wrote:
> > >> It looks like you are routinely removing the human readable comment
> > >> before the entry and only leaving the Unicode point.  I'm of the
> > >> opinion that it is useful for most of the Unicode-converted fields to
> > >> have a human readable comment to lower the barrier to discovery and
> > >> improvement of entries.   is there a strong rationale for deleting the
> > >> comments?
> > > 
> > > the readable comments are out of date when i make changes, and i've
> > > seen some cases (independent of my changes) where they're out of date.
> > > i'd rather push for better tooling and get people used to that than
> > > rely on possibly stale comments.  the other option is where the files
> > > are just always encoded in UTF-8 as discussed in an older thread and
> > > then there's no need for comments at all.
> > 
> > Right, if we get consensus to just use UTF-8 (I can't remember if there
> > some objections last time) then we can do that.
> 
> Well, I think we should still use ASCII for portability. I belive Unicode does
> the same for their data.

CLDR does not -- it is entirely & gratuitously UTF-8 encoded

the Unicode standard (i.e. UCD.zip & Unihan.zip) are mostly in ASCII, but
that's because they're describing the codepoints themselves.  a few of the
files in there do use UTF-8 when it makes sense.
-mike

Attachment: signature.asc
Description: Digital signature