Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Just FTR, this is upstream response: > Yes, tzinfo supports the vanguard format introduced in tzdata 2018d that became the main format in 2018e. > Conversions to and from local time will be accurate with any release supporting system zoneinfo files. > I would recommend using the latest v1.2.5

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Thx for the answer. I opened upstream ticket to doublecheck this: https://github.com/tzinfo/tzinfo/issues/88 V. Dne 24.5.2018 v 19:32 Patsy Franklin napsal(a): > Hi, > > Thank you for your input on this! > > The concern is that, in some cases, applications that parse the tzdata > files were

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

On Wed, May 23, 2018 at 3:23 AM, Florian Weimer wrote: > On 05/22/2018 05:06 PM, Lennart Poettering wrote: > >> The old format was nicely documented in tzfile(5): >> >> http://man7.org/linux/man-pages/man5/tzfile.5.html >> >> Through this it kinda has become API. Which

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Hi, I agree that the documentation has been very good. From what I've seen, the upstream maintainer continues to update the documentation as needed. Thanks for your input! Patsy On Tue, May 22, 2018 at 11:06 AM, Lennart Poettering wrote: > On Di, 22.05.18 16:24, Jan

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Hi, Thanks for the question! No, this change should not change the signature at the beginning of the file. My understanding is that it should only impact application provided parsers that read the time zone data files and expect the SAVE/DST values to only be positive. Thanks, Patsy On Tue,

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Hi, Thank you for your input on this! The concern is that, in some cases, applications that parse the tzdata files were not expecting to find negative DST offsets (negative SAVE values in the data files) despite the fact that POSIX appears to support this. As this is now the default format for

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

Is this going to be backward compatible or not? rubygem-tzinfo is using tzdata as data source, so I wonder if any action is required. V. Dne 22.5.2018 v 16:24 Jan Kurik napsal(a): > = Proposed System Wide Change: The tzdata transition to 'vanguard' format = >

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

On 05/22/2018 05:06 PM, Lennart Poettering wrote: The old format was nicely documented in tzfile(5): http://man7.org/linux/man-pages/man5/tzfile.5.html Through this it kinda has become API. Which doesn't mean it can't be changed, but can we please make sure that the new format is as well

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

On Di, 22.05.18 16:24, Jan Kurik (jku...@redhat.com) wrote: > = Proposed System Wide Change: The tzdata transition to 'vanguard' format = > https://fedoraproject.org/wiki/Changes/TZDATA-VANGUARD > > > Owner(s): > * Patsy Franklin > > > As of tzdata-2018e, the upstream will now default to

Re: F29 System Wide Change: The tzdata transition to 'vanguard' format

On Tue, May 22, 2018 at 04:24:05PM +0200, Jan Kurik wrote: > = Proposed System Wide Change: The tzdata transition to 'vanguard' format = > https://fedoraproject.org/wiki/Changes/TZDATA-VANGUARD Hi, does the new format change the magic at the beginning of the file? In systemd upstream we recently