Three more tweaks
Hi, I’ve surveyed the entire database for transitions 24h or less apart with one of them being a faux transition. I found only 6. You just eliminated the first. Two more are in Europe/Istanbul and I see no way to eliminate the faux transition. Here are the remaining 3 with explanations and fixes: ——————————— Europe/Simferopol 1996 Mar 31: 1996-03-30 20:59:59u = 1996-03-30 23:59:59 MSK // 03:00:00 E-Eur MSK/MSD 1996-03-30 21:00:00u = 1996-03-31 01:00:00 MSD // 03:00:00 E-Eur MSK/MSD And then 3h later: 1996-03-30 23:59:59u = 1996-03-31 03:59:59 MSD // 03:00:00 E-Eur MSK/MSD 1996-03-31 00:00:00u = 1996-03-31 04:00:00 MSD // 03:00:00 01:00 MSD The second faux transition can be combined with the first actual transition with: diff --git a/europe b/europe index 630c234..1242927 100644 --- a/europe +++ b/europe @@ -2364,7 +2364,7 @@ Zone Europe/Simferopol 2:16:24 - LMT 1880 # changed in May. 2:00 E-Eur EE%sT 1994 May # From IATA SSIM (1994/1997), which also says that Kerch is still like Kiev. - 3:00 E-Eur MSK/MSD 1996 Mar 31 3:00s + 3:00 E-Eur MSK/MSD 1996 Mar 31 0:00s 3:00 1:00 MSD 1996 Oct 27 3:00s # IATA SSIM (1997-09) says Crimea switched to EET/EEST. # Assume it happened in March by not changing the clocks. ——————————— Europe/Sofia 1982 Sep 26: 1982-09-25 22:59:59u = 1982-09-26 01:59:59 EEST // 02:00:00 Bulg EE%sT 1982-09-25 23:00:00u = 1982-09-26 02:00:00 EEST // 02:00:00 C-Eur EE%sT And then 1h later: 1982-09-25 23:59:59u = 1982-09-26 02:59:59 EEST // 02:00:00 C-Eur EE%sT 1982-09-26 00:00:00u = 1982-09-26 02:00:00 EET // 02:00:00 C-Eur EE%sT The first faux transition can be merged with the second actual transition with: diff --git a/europe b/europe index 630c234..f096de5 100644 --- a/europe +++ b/europe @@ -845,7 +845,7 @@ Zone Europe/Sofia 1:33:16 - LMT 1880 1:00 C-Eur CE%sT 1945 1:00 - CET 1945 Apr 2 3:00 2:00 - EET 1979 Mar 31 23:00 - 2:00 Bulg EE%sT 1982 Sep 26 2:00 + 2:00 Bulg EE%sT 1982 Sep 26 3:00 2:00 C-Eur EE%sT 1991 2:00 E-Eur EE%sT 1997 2:00 EU EE%sT ——————————— Europe/Tallinn 1999 Nov 1: 1999-10-31 00:59:59u = 1999-10-31 03:59:59 EEST // 02:00:00 EU EE%sT 1999-10-31 01:00:00u = 1999-10-31 03:00:00 EET // 02:00:00 EU EE%sT And then 21h later: 1999-10-31 21:59:59u = 1999-10-31 23:59:59 EET // 02:00:00 EU EE%sT 1999-10-31 22:00:00u = 1999-11-01 00:00:00 EET // 02:00:00 EET The second faux transition can be combined with the first actual transition with: diff --git a/europe b/europe index 630c234..c788416 100644 --- a/europe +++ b/europe @@ -1084,7 +1084,7 @@ Zone Europe/Tallinn 1:39:00 - LMT 1880 3:00 Russia MSK/MSD 1989 Mar 26 2:00s 2:00 1:00 EEST 1989 Sep 24 2:00s 2:00 C-Eur EE%sT 1998 Sep 22 - 2:00 EU EE%sT 1999 Nov 1 + 2:00 EU EE%sT 1999 Oct 31 4:00 2:00 - EET 2002 Feb 21 2:00 EU EE%sT Thanks, Howard
Thanks again. I installed the attached patch into the experimental version.
On Jun 20, 2015, at 8:41 PM, Paul Eggert <eggert@cs.ucla.edu> wrote:
Thanks again. I installed the attached patch into the experimental version. <0001-Tidy-up-data-entries-for-more-locations.patch>
Thanks Paul. At first I wasn’t sure if this was a misunderstanding I had of the database (again) or not. And truth be told I found 12 closely spaced transitions, but six were explained away by two bugs in the parser I'm working on. And so I’m not confident I have a reliable parser yet. Howard
participants (2)
-
Howard Hinnant -
Paul Eggert