[tz] Irish Standard Time vs Irish Summer Time

[Howard Hinnant]

On Dec 9, 2017, at 4:41 PM, Brian Inglis <Brian.Inglis at systematicsw.ab.ca> wrote:
> 
> Otherwise this is an atypical new case in the data provided by the project.
> It should be flagged for downstream as a possibly breaking change for tests, and
> for code on platforms not using the reference implementation.
> 
> As with removal of historical abbreviations, there is a risk of possible
> downstream impacts post-release, delaying propagation of required data changes.
> 
> It would be useful if major downstream distros, vendors, and platforms could
> test against the experimental repo and provide feedback prior to any release.
> If testing against the experimental repo was incorporated into inter-release CI
> processes, it could provide valuable pre-release feedback, and avoid "surprise"
> regressions on release.

Any major downstream distro, vendor, or platform that is relying on this code/database and is *not* monitoring this mailing list does not have the qualifications to deliver top quality software.

I’m a very minor vendor, and for an unpaid gig at that.  And even I take the time to monitor this mailing list for breaking changes.  The github repo, and the patches that Paul emails here are wonderful.  That imho is more than sufficient notice and availability for pre-release testing.  If I can do it on hobby budget & time, any major vendor/distro that is not doing such monitoring and testing is just incompetent.

All that being said, careful weighing of the benefit/cost ratio by this group for every change is much appreciated.  In the future this database will continue to grow as an international de facto standard for global time keeping.  There is a proposal before the C++ standards committee right now to create an API based not on the code this group supplies, but on the database, and most specifically, the time zone identifiers in the database:

http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0355r4.html

Not to worry, it is the intent of this proposal that vendors of the C++ compiler and library supply the database, and not https://www.iana.org/time-zones.  Indeed, many platforms already do so in the form of the zic-compiled files (though the zic-compiled form is not required).

Howard

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: Message signed with OpenPGP
URL: <http://mm.icann.org/pipermail/tz/attachments/20171209/a787ab87/signature.asc>