Re: [PATCH v2] localedata: i18napis: update LC_ADDRESS.country_post

[keld at keldix dot com]

On Fri, Aug 05, 2016 at 06:03:07AM +0200, keld@keldix.com wrote:
> On Thu, Aug 04, 2016 at 03:32:56PM -0400, Mike Frysinger wrote:
> > On 01 Jul 2016 15:29, keld@keldix.com wrote:
> > > On Thu, Jun 30, 2016 at 11:28:26AM -0400, Mike Frysinger wrote:
> > > > On 20 Jun 2016 14:51, keld@keldix.com wrote:
> > > > > I am travelling and has limited acces to ressources, but I will return in a few days.
> > > > 
> > > > and now ?
> > > 
> > > I am back.
> > > I will look at it later today.
> > 
> > ping ?
> > -mike
> 
> I did have a look back then, but I had some difficulties seeing what
> was the organisation and purpose of the API, and what could be done with it.
> It looks like it is doing something else than the glibc spec.
> 
> They refer to the code as documentation, which is always good but
> difficult to grasp. I see there is a video, which I will investigate.

I did a little more research on this. I watched the video...

The libaddressinpput is input oriented while the glibc postal address spec is output
oriented. Libaddressinput is for validating input  commonly from websites,
while the glibc spec is for outputting addresses in a culturally acceptable manner,
eg. for writing letters or addresses in some list of persons.

So this is 2 different purposes, which both have their merits.

Furthermore the libaddressinput is in Java and C++ only, so it does not
cover C usage, which is the aim of glibc.

So I think we  should retain our specs and possibly also look into a C interface
for libaddressinput, though we do only have limited functionality for checking input,
as yesexpr and noexpr.

Best regards
Keld