Commits

Show all
Author Commit Message Labels Comments Date
Bryan O'Sullivan
Add ToJSON and FromJSON instances for LocalTime
Bryan O'Sullivan
Add ToJSON and FromJSON instances for Day This fixes gh-252.
Bryan O'Sullivan
Fix some warnings
Bryan O'Sullivan
Build fix for benchmarks
Bryan O'Sullivan
Support time zones of the form +HH, with minutes omitted This finishes off the cherry-picking of gh-270. I think.
Bryan O'Sullivan
Merge
Bryan O'Sullivan
Merge pull request #270 from lpsmith/utcoffset-fixes Utc offset fixes
lpsmith
Fix utc offset parsing As I understand ISO 8601:2004, ±HHMM is not a valid syntax for specifying a UTC offset in the context we are using it. Only ±HH and ±HH:MM are valid in the extended ISO 8601 formats. I probably would weakly prefer to not accept ±HHMM, however, because the the ToJSON ZonedTime instance has been emitting this syntax since aeson version 0.6.1.0 released on 2012-12-14, our hands may well be tie…
lpsmith
Add regression tests for utc offset parsing
Bryan O'Sullivan
Fix the negative timezone offset bug revealed in the previous commit This is a simpler fix than gh-270, and doesn't try to do several things at once.
Bryan O'Sullivan
Add a negative timezone test with a non-zero number of minutes This is cherry-picked from gh-270.
Bryan O'Sullivan
Make the Parser.Time module private
Bryan O'Sullivan
More doc improvements
Bryan O'Sullivan
Move typeMismatch, and improve its docs
Bryan O'Sullivan
A few doc improvements
Bryan O'Sullivan
Update (and simplify) benchmark results
Bryan O'Sullivan
Make UTCTime encoding 3x faster It's now over 10x faster than the formatTime-based approach. In this instance, the win comes from fixed-width arithmetic (Int64 vs Integer).
Bryan O'Sullivan
Make ZonedTime encoding 3x faster
Bryan O'Sullivan
Move the encode function to a more convenient place
Bryan O'Sullivan
Speed up UTCTime encoding by 3x
Bryan O'Sullivan
Drop some dead code
Bryan O'Sullivan
Move Pico conversion code
Bryan O'Sullivan
Add date/time encoding benchmarks
Bryan O'Sullivan
A few minor cleanups
Bryan O'Sullivan
Better living through GHC's constructed product return Allocating a polymorphic tuple makes GHC unable to see that there's actually no allocation needed in the parsePicos loop.
Bryan O'Sullivan
Merge
lpsmith
Fix integer overflow and optimize subsecond parsing Previously we had this behavior: > decode "[\"2014-07-18 12:00:00.99999999999999Z\"]" :: Maybe [UTCTime] Just [2014-07-18 12:00:00.999999999999 UTC] > decode "[\"2014-07-18 12:00:00.9999999999999999999999Z\"]" :: Maybe [UTCTime] Just [2014-07-18 12:00:00.000186471204 UTC] Now the digits past picoseconds are truncated in all cases.
lpsmith
Fix imports for GHC < 7.10
Bryan O'Sullivan
Add a few zero fractions for good measure
Bryan O'Sullivan
Rename some variables
  1. Prev
  2. Next