Hi, > Might be ok to change the tests so they don't expect the number format to > change.
I've gone through all of the other failures and I don't think there’s any major concerns. The drop down (now fixed) seems to be the only regression issue. - The negative number / locale tests (40 odd tests) would require a lot of changes and then some of the tests wouldn’t be testing anything really. It may be that these tests work fine on windows but not sure. Either way it’s not a SDK issue but a Flash Player issue. - The chart tests give different results depending on what hemisphere you’re in. Well time zone actually, but it's more noticeable for northern vs southern hemisphere. An educated guess this is that this is related to the dates of daylight saving time and the date axis. The date axis calculations assume there always 24 hours in a day which doesn’t happen when you go in/out of daylight savings time and thus you can get a day off - which is what we are seeing. Changing the tests to not use local time failed to fix the issue (but may of and then all bitmaps needed updating?), but changing time zone and retuning tests did fix it. - The failing 12 mirrors tests I was a little more concerned about. But if I run the tests on the previous three versions of the SDK they also failed so it's unlikely to be a regression issue. - That leaves 8 failing tests that are also not regression issues as I also ran these on the previous three versions of the SDK. If anyone wants to tests these failing tests on windows that would be appreciated. I can supply failure.txt files listing the failing tests and they only take a few minutes to run. Thanks, Justin
