On 14/05/2018 16:43, Peter Levart wrote:
:
Is it really that important to allow users to modify static final
fields that way? As such fields are normally constant folded by JIT, I
doubt that anybody is doing it nowadays. Doing it is bound to
unpredictable program behavior, as JVM is free to never reload such
field's value.
Sadly, there are still a number of libraries using this hack to change
static final fields. They should be seeing an "Illegal reflective ..."
warning today and the hack will break once the encapsulation is dialed
up. The torch and pitchfork crowd will be out when that happens and I
read Mandy's approach as just leaving the issue to that day.
-Alan
