JRuby

Time#iso8601 output of nanoseconds lost in --1.9 mode

Details

  • Type: Bug Bug
  • Status: Closed Closed
  • Priority: Major Major
  • Resolution: Won't Fix
  • Affects Version/s: JRuby 1.6
  • Fix Version/s: None
  • Component/s: None
  • Labels:
    None
  • Environment:
    OSX 10.6.7, JRuby 1.6.0 (via rvm) in --1.9 mode
  • Number of attachments :
    0

Description

In --1.9 mode the nanoseconds part is dropped when Time#iso8601 is passed an argument:

ruby --1.8 -rtime -e 'p Time.utc(2010, 12, 31, 23, 59, 59, 1).iso8601(9)'
# => "2010-12-31T23:59:59.000001000Z"

ruby --1.9 -rtime -e 'p Time.utc(2010, 12, 31, 23, 59, 59, 1).iso8601(9)'
# => "2010-12-31T23:59:59.000000000Z"

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Hiro Asari added a comment -

With 1.9, time library ends up calling Time#subsec (https://github.com/ruby/ruby/blob/d737ae54496e6ea5a4d9f054245503ca06a5bfcd/lib/time.rb#L524) to construct that part of the formatted string. Right now, we discard anything below milliseconds inside that method, since JVM does not provide that level of temporal granularity. (https://github.com/jruby/jruby/blob/36b741d4e6f9c870cfb7c3467ff3726eac0f5795/src/org/jruby/RubyTime.java#L645)

Is there anything we can do to get around this problem?

Show
Hiro Asari added a comment - With 1.9, time library ends up calling Time#subsec ( https://github.com/ruby/ruby/blob/d737ae54496e6ea5a4d9f054245503ca06a5bfcd/lib/time.rb#L524 ) to construct that part of the formatted string. Right now, we discard anything below milliseconds inside that method, since JVM does not provide that level of temporal granularity. ( https://github.com/jruby/jruby/blob/36b741d4e6f9c870cfb7c3467ff3726eac0f5795/src/org/jruby/RubyTime.java#L645 ) Is there anything we can do to get around this problem?
Hide
Permalink
Dan Kubb added a comment -

Whoops, I probably should've titled this with "milliseconds" instead of "nanoseconds" even though in the Time#iso8601 I pass in 9 for the scale, but I specify the milliseconds in the constructor.

Show
Dan Kubb added a comment - Whoops, I probably should've titled this with "milliseconds" instead of "nanoseconds" even though in the Time#iso8601 I pass in 9 for the scale, but I specify the milliseconds in the constructor.
Hide
Permalink
Hiro Asari added a comment -

The 7th argument in Time#utc is microseconds, which JVM doesn't have. The argument to Time#iso8601 specifies how many digits the resulting string has after the decimal point.

$ ruby1.9 -v -rtime -e 'p Time.utc(2010, 12, 31, 23, 59, 59, 1).iso8601(9)'
ruby 1.9.3dev (2011-03-30 trunk 31218) [x86_64-darwin10.6.0]
"2010-12-31T23:59:59.000001000Z"
Show
Hiro Asari added a comment - The 7th argument in Time#utc is microseconds, which JVM doesn't have. The argument to Time#iso8601 specifies how many digits the resulting string has after the decimal point. $ ruby1.9 -v -rtime -e 'p Time.utc(2010, 12, 31, 23, 59, 59, 1).iso8601(9)' ruby 1.9.3dev (2011-03-30 trunk 31218) [x86_64-darwin10.6.0] "2010-12-31T23:59:59.000001000Z"
Hide
Permalink
Hiro Asari added a comment -

I'm marking this as WONTFIX for now, since JVM can't provide a reasonable temporal granularity required here.

Show
Hiro Asari added a comment - I'm marking this as WONTFIX for now, since JVM can't provide a reasonable temporal granularity required here.

People

  • Assignee:
    Hiro Asari
    Reporter:
    Dan Kubb
Vote (1)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved: