[gwt-contrib] Re: dev tests failing locally?

2016-04-19 Thread Colin Alworth
Confirmed, isPublic is true in all three cases, as expected, without the old code in place. On Tuesday, April 19, 2016 at 9:20:10 AM UTC-5, Colin Alworth wrote: > > There is no MOD_DEFAULT, at least not yet. I initially attempted to add > one and tried to copy the same date from

[gwt-contrib] Re: dev tests failing locally?

2016-04-19 Thread Colin Alworth
There is no MOD_DEFAULT, at least not yet. I initially attempted to add one and tried to copy the same date from org.objectweb.asm, but their Opcodes type doesn't really have a field that seems to match (no ACC_DEFAULT, etc). If memory serves, bytecode doesn't hold this information, but just

[gwt-contrib] Re: dev tests failing locally?

2016-04-19 Thread Thomas Broyer
On Tuesday, April 19, 2016 at 2:37:03 PM UTC+2, Colin Alworth wrote: > > As an aside, the fix for the original bug seems to be very easy - the code > that is building the JMethod instances seems to be overzealous about what > modifiers must be applied. Aside from the constant failures, all

[gwt-contrib] Re: dev tests failing locally?

2016-04-19 Thread Colin Alworth
As an aside, the fix for the original bug seems to be very easy - the code that is building the JMethod instances seems to be overzealous about what modifiers must be applied. Aside from the constant failures, all compiler tests pass with this change: diff --git