ac7cf58962995825464af08ae6fa5e006c94f3fa |
|
27-Feb-2013 |
Elliott Hughes <enh@google.com> |
Switch to using icu4c 50's TimeZoneNames API. This takes us down from ~330ms/locale on prime to ~70ms/locale. This also fixes a bug in DateFormatSymbols.getZoneStrings and adds a test so we don't regress. Change-Id: I13663a659666c0a7d87a299235b75fa4e9fef69d
|
f20e96718a936499da309766da7f36f123b43d93 |
|
18-Jan-2013 |
Elliott Hughes <enh@google.com> |
Improve our time zone name collection. A variety of bugs here. We weren't using dates in the current year, we were assuming that icu4c has correct time zone transition data (it usually doesn't), and in the fallback code (which we'd previously never hit) we weren't honoring the caller's request for standard/daylight time. This patch fixes: Africa/Tripoli: long='Eastern European Standard Time' short='Eastern European Standard Time'! America/Araguaina: long='Brasilia Standard Time' short='Brasilia Standard Time'! America/Campo_Grande: long='Amazon Standard Time' short='Amazon Standard Time'! America/Cuiaba: long='Amazon Standard Time' short='Amazon Standard Time'! America/Havana: long='Cuba Standard Time' short='Cuba Standard Time'! America/Santiago: long='Chile Standard Time' short='Chile Standard Time'! America/Sao_Paulo: long='Brasilia Standard Time' short='Brasilia Standard Time'! Antarctica/Palmer: long='Chile Standard Time' short='Chile Standard Time'! Asia/Jerusalem: long='Israel Standard Time' short='Israel Standard Time'! Asia/Tehran: long='Iran Standard Time' short='Iran Standard Time'! Asia/Tel_Aviv: long='Israel Standard Time' short='Israel Standard Time'! Brazil/East: long='Brasilia Standard Time' short='Brasilia Standard Time'! Chile/Continental: long='Chile Standard Time' short='Chile Standard Time'! Chile/EasterIsland: long='Easter Island Standard Time' short='Easter Island Standard Time'! Cuba: long='Cuba Standard Time' short='Cuba Standard Time'! Iran: long='Iran Standard Time' short='Iran Standard Time'! Israel: long='Israel Standard Time' short='Israel Standard Time'! Libya: long='Eastern European Standard Time' short='Eastern European Standard Time'! Pacific/Apia: long='Samoa Standard Time' short='Samoa Standard Time'! Pacific/Easter: long='Easter Island Standard Time' short='Easter Island Standard Time'! Pacific/Fiji: long='Fiji Standard Time' short='Fiji Standard Time'! It also fixes Pacific/Fakaofo, which was confused between GMT+14:00 and GMT+13:00. We need a special case for Pacific/Apia because icu4c 4.9 doesn't believe Samoa has ever used daylight time. I've also added a test to keep us out of mischief in the face of future upgrades to the Olson tzdata and/or icu4c. Bug: 7955614 Bug: 8026776 Change-Id: I7ed551a691f29e26dd71456d492493d2a4fce945
|
687409006d21cb01c56b3a3acc180952bb4d7e3e |
|
22-Jun-2011 |
Jesse Wilson <jessewilson@google.com> |
Don't call overrideable methods from a constructor. Conflicts: luni/src/test/java/libcore/java/util/TimeZoneTest.java (cherry-pick of 7dbf334facc7f43c18244150d2052ae4ec8c5e16.) Change-Id: I0aa0f3c8776538076bdfed070da08e65aa959486
|
78c3de051d68b703af480778c100ca335690b250 |
|
30-Jul-2012 |
Elliott Hughes <enh@google.com> |
Fix TimeZone's handling of Australia/Lord_Howe. Australia/Lord_Howe has a half hour difference between standard and daylight time, rather than the usual hour. Our ZoneInfo implementation ignored this. Fix that oversight, make SimpleDateFormat actually use this information, and prevent TimeZone.getTimeZone("GMT") and TimeZone.getTimeZone("UTC") from being quite as expensive as they accidentally were. Longer term I think we should probably remove all uses of getDSTSavings from libcore in favor of TimeZone.getOffset, but this is probably a useful step forwards anyway. It fixes Australia/Lord_Howe in the meantime and it means that anyone else who's using getDSTSavings won't be bitten (even if they too really ought to be using TimeZone.getOffset). Bug: 4723412 Bug: http://code.google.com/p/android/issues/detail?id=24684 Change-Id: I5d50afecbe1453157e9c8f0b88305a258a3ba2e0
|
91348f798cddae22b59ee1a17bd24315c0897f6f |
|
23-Jun-2012 |
Elliott Hughes <enh@google.com> |
Be more careful when parsing custom zone ids. The old code was allowing invalid ids. Bug: 6556561 Change-Id: I691d33fa133527a76bbffa4e3b56a023c389ca8f
|
6fb123c8681a795c65cbd67ffcc0ef641c473288 |
|
10-Jan-2012 |
Elliott Hughes <enh@google.com> |
Make it clearer that TimeZone.getTimeZone(null) throws NPE on purpose. Bug: http://code.google.com/p/android/issues/detail?id=24036 Change-Id: I3464dc5094274931d67e3ca4a5bc553fa1a6dc96
|
d5250b7d0dbc1e01948a5e7551deefa697baf156 |
|
05-May-2011 |
Elliott Hughes <enh@google.com> |
Add an explicit test for a public bug that we'd already fixed. Bug: http://code.google.com/p/android/issues/detail?id=16608 Change-Id: I30cf4dede3839ee5d7cc9e55e04cf99149294aa6
|
6d82ce5d35a5e84aedf08528fd98b849f3f565a6 |
|
15-Feb-2011 |
Elliott Hughes <enh@google.com> |
Move ZoneInfo and ZoneInfoDB into libcore.util. Change-Id: I91385512ec27aa2aebdc163ae204e542efa58f21
|
65c53ceda4279ee7c19cf42d7c265cd601340ca0 |
|
15-Feb-2011 |
Elliott Hughes <enh@google.com> |
Improve TimeZone's behavior for "pre-historic" dates. By pre-historic I mean "before the beginning of our table of transitions". It's not entirely obvious what the right behavior is, but it's pretty obvious that our old behavior was unhelpful and inconsistent. This is an attempt at a reasonable semantics, backed by tests. Also add a test for an unrelated TimeZone bug (already fixed). Bug: http://code.google.com/p/android/issues/detail?id=14395 Bug: http://code.google.com/p/android/issues/detail?id=11918 Change-Id: I230ab21f9a38e37d6d0a44f97d5a8ac8b01df7d9
|
cf988a1dfb7f4a13bd797ce3a424242474b7c152 |
|
26-Jan-2011 |
Elliott Hughes <enh@google.com> |
Fix time zone formatting for custom SimpleTimeZone instances. Rather than return null for unknown time zones, as our code assumed, ICU unhelpfully pretends unknown time zones are GMT. We need to check whether we got what we asked for, which is a bit of a pain. Bug: 3049014 Change-Id: Ie7f0043b8da622767538ed537a9e2d927652215e
|
c4823abbed65fe80790ad215d6b96679c25b3f68 |
|
28-Sep-2010 |
Elliott Hughes <enh@google.com> |
Add submitter's test for a publicly-reported SimpleTimeZone.clone bug I already fixed. Bug: http://code.google.com/p/android/issues/detail?id=11542 Change-Id: Ic13e6083ec26e9d8725e94294f7cd69902e69dd8
|
4557728efb66c455a52b7669a8eefef7a9e54854 |
|
11-Aug-2010 |
Jesse Wilson <jessewilson@google.com> |
Moving tests to be under the libcore.* package. This is indended to make it easier to run on VMs that restrict the packages from which application classes can be loaded. For example, on the RI you need to use the bootclasspath to load these tests. Change-Id: I52193f35c5fcca18b5a3e1d280505b1e29b388af
|