RFR(S): 8136469: OptimizeStringConcat fails on pre-sized StringBuilder shapes
tobias.hartmann at oracle.com
Fri Sep 18 11:00:08 UTC 2015
thanks for having a look.
On 18.09.2015 12:38, Aleksey Shipilev wrote:
> On 09/18/2015 12:57 PM, Tobias Hartmann wrote:
> Fix itself looks good to me, but why don't we add a proper
> PrintOptimizeStringConcat debug block instead of the assert?
I had that at first but then thought that with the additional check for a dead test, the isa_Bool() should not fail.
> I think OptoStringConcat should fail gracefully on graph mismatch
> anyway, not crash the fastdebug build. Is it possible to have anything
> else besides Bool or Con as an input for If?
IfNode::Ideal() suggests that we could also have TOP but I don't think this should be possible right after parsing.
To be save, we can also go with this fix:
More information about the hotspot-compiler-dev