DateTime.iso8601 fails with an error if a second fraction is present#2941
DateTime.iso8601 fails with an error if a second fraction is present#2941azolotko wants to merge 1 commit intojruby:jruby-1_7from azolotko:2883-iso8601-sec-fraction
Conversation
|
Keep in mind that |
|
@BanzaiMan Yes, absolutely. Please correct me if I'm wrong, but MRI seems to do a similar thing: https://github.com/ruby/ruby/blob/trunk/ext/date/date_parse.c#L2363 |
|
@BanzaiMan in this case it is acceptable as date.rb has gone native somewhere along MRI 1.9.3 + this is a follow-up on a previous date/format.rb patch: b2e1dc1?diff=unified @azolotko test/spec will be fine under regressions ... just be aware to filter it out in 1.8 mode (still tested against under jruby-1_7) |
|
There seems to be some difference in meaning of |
|
Having the spec in spec/ruby (RubySpec) would be useful as well. |
There was a problem hiding this comment.
What about Float("0#{sec_fraction}") ?
There was a problem hiding this comment.
cause of MRI compatibility:
1.9.3-p551 :002 > date = DateTime.iso8601('2014-07-08T17:51:36.013Z')
=> #<DateTime: 2014-07-08T17:51:36+00:00 ((2456847j,64296s,13000000n),+0s,2299161j)>
1.9.3-p551 :003 > date.sec_fraction
=> (13/1000)
1.9.3-p551 :004 > date.sec_fraction.class
=> Rational
1.9.3-p551 :005 > date.second_fraction
=> (13/1000)
|
cherry-picked commit on jruby-1_7 as 4e822a1 ... great work Alex, esp. that you matched MRI ... thanks. |
4 participants
Fixes #2883
I would be happy to add a test, if I only knew where to put it.
Thanks!