Re: [Ntp] Finding leap-seconds.list

[Thomas Peterson <[e-mail]>]

Thank you Paul for the further clarity around the hosting - I was 
unaware that data.iana.org also hosted it simply because search engines 
don't appear to know about it, which is strange to me as the robots.txt 
on www.iana.org and data.iana.org do not prohibit spiders on the URLs in 
question.

Harlan: If yourselves and the IANA webmasters do not object, would it be 
possible to raise a bug with ntp to change the URL used by 
'update-leap'? I would have done so myself however new accounts appear 
to require approval.


Regards


On 09/11/2018 21:12, Paul Eggert wrote:
> On 11/8/18 8:40 PM, Thomas Peterson wrote:
>> I was intending to keep the scope of this discussion solely on what 
>> is hosted under the ietf.org domain.
>
> Although the intent was that leap-seconds.list (as part of tzdb) be 
> distributed via iana.org URLs, it is also distributed under ietf.org 
> because  <https://data.iana.org/time-zones/> and 
> <https://www.ietf.org/timezones/> are clones.
>
> I don't know why the ietf.org URLs exist. I don't even recall knowing 
> about them until this discussion. However, it appers that access via 
> ietf.org is common. If I search Google for "leap-seconds.list" the 
> first hit is the ietf.org copy (the iana.org copy isn't even in the 
> top ten), and I sent ietf.org URLs to you in my previous email because 
> I Googled for files without noticing that the resulting URLs were 
> ietf.org and not iana.org.
>
> In looking for why people seem to prefer ietf.org URLs, I see that NTP 
> started using them for leap-seconds.list in NTP version 4.2.8p9 
> (2016-11-21) presumably because the nist.gov URLs were so flaky. My 
> guess is that the NTP developer in question didn't know that iana.org 
> was intended and ietf.org was (at least from the tzdb point of view) 
> accidental. I assume we can file a bug report with the NTP developers 
> to replace ietf.org URLs with iana.org URLs, though it will take some 
> time for any such change to trickle out to real-world use.
>
> If the intent is to decommission the ietf.org URLs, I think we're all 
> agreed that we should give NTP and other downstream users sufficient 
> advance notice of the switch. I also suggest redirecting 
> leap-seconds.list (and other tzdb-related) HTTPS requests from 
> ietf.org to iana.org after the decommission occurs.
>
> I will ask the iana.org webmasters why there are ietf.org clones of 
> the URLs in question. Quite possibly there's a good reason that I 
> don't know.
>
>
>> There also does not appear to be links from the IANA time zone page 
>> you link nor from any of the downloads that contain specific 
>> reference to the second and third links in your email, so could you 
>> please clarify how those are announced? 
>
> Tzdb releases are announced on [e-mail] which is described 
> at <https://mm.icann.org/mailman/listinfo/tz-announce>. (I don't know 
> why the mailing list is archived is at icann.org and not iana.org; I 
> suppose that's another question I can ask the iana.org webmasters, if 
> there's interest.)
>
> The most recent tzdb release announcement is archived at 
> <https://mm.icann.org/pipermail/tz-announce/2018-October/000052.html>. 
> Releases are distributed via HTTPS (also plain HTTP or FTP, but these 
> are deprecated). Each release announcement contains GPG checksums for 
> the release files, so the leap-seconds.list file in each distribution 
> can be checked by verifying these checksums.
>