http://gwt-code-reviews.appspot.com/1601805/diff/1/user/src/com/google/web/bindery/autobean/shared/ValueCodex.java File user/src/com/google/web/bindery/autobean/shared/ValueCodex.java (right):
http://gwt-code-reviews.appspot.com/1601805/diff/1/user/src/com/google/web/bindery/autobean/shared/ValueCodex.java#newcode368 user/src/com/google/web/bindery/autobean/shared/ValueCodex.java:368: if (clazz.isEnum()) { On 2011/12/06 16:29:24, tbroyer wrote:
(I'll try to do that testing right now though, so we can have numbers
on which
to base our decision).
I did my homework: https://docs.google.com/spreadsheet/ccc?key=0Agcd-Zsy2T-YdGtvV2tOS1c0VlBYV0t6Vl9HTmlOa2c Note that it's by no mean a scientific experiment (machine has more than 15 days of uptime, I had Chrome running at the same time –though I didn't touch the machine while the benchmarks run–). I'm not sure my tests are OK too (are you supposed to do loops in your benchmarks? or the Benchmark machinery does that for you?) Anyway, as I interpret it, there's an overall gain in removing the test, except for the case where you want to encode an enum value for which getClass().isEnum()==false (MyEnum.BAR in the test). Based on that, I kept the change in. Let me know if you want me to revert it. http://gwt-code-reviews.appspot.com/1601805/ -- http://groups.google.com/group/Google-Web-Toolkit-Contributors