[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Pulp-list] Date and Time Feedback Requested!



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/21/2011 03:55 AM, Jason L Connor wrote:
> Sorry about the delay for responding to this one.
> 
> On Mon, 2011-04-18 at 15:22 -0500, Jeff Ortel wrote:
>> Why not just standardize on a standard format like ISO 8601?
> 
> After looking into iso8601, it seems to have everything we're looking
> for:
>       * representation of date and time with timezone information
>       * representation of time intervals and repeating time intervals
>       * readily supported by all major languages
>       * support in python's standard library
>       * python-iso8601 package already in fedora and epel for some of
>         the more esoteric functionality
> 
> Given this, I would actually like to adopt it for all of pulp's date and
> time input and output. We already support it on parts of our API. I vote
> we standardize on it and convert the rest of pulp to its use.

Sounds good to me.

>> So, the pulp server will convert all dates to UTC before
>> using/storing, right? 
> 
> I'm actually not sure of the right approach here. I think we want pulp
> to utilize date and times according to the current local time, taking
> into account daylight savings time, timezones, etc. Personally, I think
> it would be easier to store date and time information by the local
> timezone. But make the timezone information explicit.

So you're suggesting we take a time submitted from TZ A (the caller) and
convert it to TZ B (the server) and store that in the DB? What happens
on the server on the days where daylights saving flips, do we need some
sort of task to run around updating things we've stored or is that
applied when it's pulled out of the DB again?

I'm not sure I follow how that's easier than converting to UTC, though I
don't actually have a strong opinion either way.

I'm going to guess that storing the local time will make logging a whole
lot easier since we can just log the value and not have to rely on the
log file reader to convert a UTC stamp. That sort of thing could save a
ton of time debugging if the logs were easily readable like that. Again
though, that's a guess and I may be missing something.

- -- 
Jay Dobies
RHCE# 805008743336126
Freenode: jdob
http://pulpproject.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJNsDNuAAoJEOMmcTqOSQHCqE8H/A+1EGdaMCALXrQp0hhX81QX
61IOIzC25GYcvLTli/bvAybYnNeNbMUok4EU12XfefWEf0k38TMhDMKw0uENDHLg
QWH3d6NLqTbpprCfBvZ7Ct85jU1zUrI9ogMGCau8TqGRDhOU33W9LJq6uczY6+YG
6CipG8T5EDkq5iZsGoDiuCGTpx67aE6k+Ayoy5oJeGxa6a+BSmP8pzyJ3Gk9QOij
4KRJK7BiUXiVm+X5FnTJol5U5VeF5vtndaIGKqiWxDbMmTq2sL/6i3G38BTAY2ka
xdGACz1+5RvWu8TJlfK2RTJufggYWC6DxJt/YtE8ps8KIx76sCDjZsuzqXAAp0A=
=PhEb
-----END PGP SIGNATURE-----


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]