Build is environment dependent - take 2 - failing tests about smooks HOT 6 CLOSED

Hi Ioan. I'm getting this error on my machine using the changes you made for #55.

I'd suspect this is a JDK version issue. Smooks was always build on Java v5 and v6, using the Sun/Oracle JDK.

from smooks.

Btw... I've no issue with getting the build running on Java7 and on other JDK implementations (OpenJDK etc). I do think however that the Sun/Oracle JDKs should be the baseline. I'd also like it to continue working on Java 5 and Java 6 if possible.

from smooks.

I think it's ok to maintain compatibility with older JDKs if it's not expensive. However I don't have access to Java 5 and I have adopted OpenJDK - which is the basis for Oracle JDK. Also there should be very few/minor differences between them.

Personally I like driving change. For me supporting Java 7 as a minimum makes sens - since older JDK6 is EOL for some time now. However I do understand the need for backwards compatibility and I think we should have a clear road-map for this - if it makes sense.

For example: Smoks version 1.x should support Java 5 & 6 . Smooks 2.x - may require minimum Java 1.7 - if it makes sense.

from smooks.

+1 to everything you say there.

from smooks.

My locale is ro_RO and DateDecoding fails with java.text.ParseException: Unparseable date: "Wed Nov 15 13:45:28 EST 2006" .

I've seen this while debugging. The exception is not propagated which is bad - we should know if the text failed to parse because a date could not be parsed.

The test fails to decode using ro_RO. The test changes the default locale to en-IE during runSmooksTransform but date decoding is done during the getBean('order').

So I think we have two issues:

• a locale issue during getBean - should we use the default locale - en-IN if none is specified instead or the system locale. Either way we should document this.
• the date decoding exception is not propagated

@tfennelly: WDYT?

from smooks.

fixed #55

from smooks.