lpeterse/haskell-utc

A pragmatic time and date library (Haskell)

haskell-utc's Issues

renderRfc3339 interprets offset as minutes instead of seconds

Consequently, renderRfc3339 is always off by a factor of 60:

> renderRfc3339  (Local (Just 3600) (epoch :: DateTime)) :: Maybe Text
Just "1970-01-01T00:00:00+60:00"

parseRfc3339 uses the offset correctly however, consequently renderRfc3339 and parseRfc3339 don't form an identity relationship anymore:

(parseRfc3339 "1970-01-01T00:00:00+01:00"  :: Maybe (Local DateTime)) >>= renderRfc3339) :: Maybe Text 
Just "1970-01-01T00:00:00+60:00"

renderRfc3339 and parseRfc3339 interpret datetime of RFC3339 string as UTC regardless of UTC offset

RFC3339 strings consist of a local time and an UTC offset. renderRfc3339 and parseRfc3339 however always interpret the local time of the string as utc time: Instead of offseting the utc field by offset before rendering and after parsing to convert the utc time of the Local type to the local time of the RFC3339 string and vice versa, these functions just treat RFC3339 as if it were a UTC time and an additional offset (just like Local works internally).

From RFC3339

Numeric offsets are calculated as "local time minus UTC". So the
equivalent time in UTC can be determined by subtracting the offset
from the local time. For example, 18:50:00-04:00 is the same time as
22:50:00Z. (This example shows negative offsets handled by adding
the absolute value of the offset.)

This bug is caused by not operating on the Local type, but on the internal type wrapped by Local whose IsTime and IsDate instances don't know about the offset. It is not trivial to fix, since in rfc3339Builder IsDate (Local t) IsTime (Local t) can't be deduced. It would require IsDate t and IsTime t to imply IsDate (Local t) and IsTime (Local t) which is currently not the case and might be tricky to achieve.

addSecondFractions fails if fraction -1 is added

To reproduce:

addSecondFractions (-1.0) epoch :: IO Time
*** Exception: UtcException "Time: setSecondFraction 1 % 1 00:00:00"

This happens, because with that value, frcs would be 0 in addSecondFractions, so setSecond (frcs + 1.0) fails of course.

Consider adding the package to Stackage

It would be nice if this package would be available on Stackage. Provided that everything builds with the current snapshot (master appears to do so, but v0.2.0.1 does not), doing so is as simple as submitting a pull request to fpco/stackage to add a few lines of metadata to the build constraints file. See the guide for details.

Release with fixed version constraints

Would be great to be able to build directly from hackage again!

Build fails with GHC 8

Data/UTC/Class/HasUnixTime.hs:39:3: error:
• Constraint ‘Monad m’ in the type of ‘getUnixTime’
constrains only the class type variables
Use ConstrainedClassMethods to allow it
• When checking the class method:
getUnixTime :: forall (m :: * -> *).
HasUnixTime m =>
forall a. (Monad m, IsUnixTime a) => m a
In the class declaration for ‘HasUnixTime’
cabal: Leaving directory '/tmp/cabal-tmp-9305/utc-0.2.0.1'
cabal: Error: some packages failed to install:
utc-0.2.0.1 failed during the building phase. The exception was:
ExitFailure 1

utc-0.1.0.0 test suite failure

Please see http://hydra.cryp.to/build/778539/log/raw for a complete log of the failing build.