Bug #11595: Time#utc? and Time#gmt? return misleading results based on $TZ

• Status: Rejected
• Assignee: akr (Akira Tanaka)
• Target version: -
• ruby -v: ruby 2.2.3p173 (2015-08-18 revision 51636) [x86_64-darwin14]
• Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: UNKNOWN

[davidcelis (David Celis)]

There is an issue with Time#utc? and its alias, Time#gmt?, that return misleading results based on the value of the TZ environment variable. It seems that the only way for a Time instance to return true for utc? is if you explicitly call #utc on it before:

ENV['TZ'] = 'UTC'
# => "UTC"
time = Time.now
# => 2015-10-14 19:30:00 +0000
time.utc?
# => false
time = time.utc
# => 2015-10-14 19:30:00 UTC
time.utc?
# => true

This seems misleading based on the value of $TZ being "UTC". The expected result for calling Time.now.utc? in this case would be true, as would that be expected for time zones that are considered links to "UTC" based on the tzdata list. These include "UTC", "GMT", "Etc/UTC", "Etc/GMT", "Universal", etc.

[nkmrya (Yasuhiro Nakamura)]

• Assignee set to akr (Akira Tanaka)

[nkmrya (Yasuhiro Nakamura)]

• File time_utc.patch time_utc.patch added

Anyway I write a patch.

Is it a bug? or specification?

[davidcelis (David Celis)]

Yasuhiro Nakamura wrote:

Anyway I write a patch.

Is it a bug? or specification?

To me it seems like a bug, since the expectation is that Time#utc? would return true for any time with an offset of 0

[avit (Andrew Vit)]

Careful: not every time with offset 0 is UTC.

A time zone with offset +0100/-0100 may have DST rules, and should not be utc?.

[akr (Akira Tanaka)]

• Status changed from Open to Rejected

Current behavior is intentional.

Time#utc? returns the mode of Time object, not the time zone offset is zero.

The mode affects several methods.

For example, Time#to_s generates "UTC" for UTC Time objects and
numeric timezone offset string for non-UTC Time objects.
The result of Time#utc? method means the difference of them.

% TZ=GMT ruby -e '
u = Time.utc(2000)
l = Time.local(2000)
p [u, u.utc?]
p [l, l.utc?]'       
[2000-01-01 00:00:00 UTC, true]
[2000-01-01 00:00:00 +0000, false]