RFR(XS) 8239854: Cannot build gtest without PCH after 8239235. Missing include in test/hotspot/gtest/runtime/test_signatureStream.cpp

Claes Redestad claes.redestad at oracle.com
Mon Feb 24 16:25:03 UTC 2020


Yes, I've been asking myself this, and turns out it was all my fault.

I fixed the test last Friday (was missing a ResourceMark), tested 
locally, then fired off a huge test job (tier4-8) over the weekend
thinking I'd be covered without re-running the lowest tiers. I wasn't.

Not running PCH on my local environment would have saved us some trouble
- and me some embarassment - and I've taken steps to ensure this never 
happens.

/Claes

On 2020-02-24 17:15, Aleksey Shipilev wrote:
> On 2/24/20 12:13 PM, Claes Redestad wrote:
>> and thanks for the periodic reminder that PCH should be an opt-in.
> I actually see that jdk-submit runs no-PCH configurations, as it failed some of my jobs before this
> fix was pushed:
>    linux-x64-debug-nopch-linux-x64-build-5 error while building, return value: 2
> 
> So I wonder how did this non-PCH failure manage to slip into jdk/jdk?
> 


More information about the hotspot-runtime-dev mailing list