RFR: 8297606: [AIX] Broken build after JDK-8295475
Stefan Karlsson
stefank at openjdk.org
Thu Dec 1 08:18:46 UTC 2022
On Wed, 30 Nov 2022 21:19:35 GMT, Tyler Steele <tsteele at openjdk.org> wrote:
> It is my understanding that future PRs will migrate many instances of ResourceObj to AnyObj (or other classes), so it's possible this work-around could be removed in the future.
No, that was not the intention with #11443. It's almost the opposite. There were a large number of classes that inherited from ResourceObj, that only used the "resource area" allocation strategy, and never needed the other allocation strategies provided by ResourceObj. Only a small number of ResourceObj classes used the other allocation strategies. So, to make that clear we split the to types into ResourceObj, which is only used for "resource area" dynamic allocations. And AnyObj, which is used when the class needs multiple dynamic allocation strategies.
Note that we got similar error messages when compiling risc-v, and there the problem was that some classes had a virtual destructor, which, apparently, requires the "operator delete" to be available. See:
https://github.com/openjdk/jdk/pull/11443#issuecomment-1286952346
I guess that requirement has to do with exception handling, but I'm not sure. Do you compile with exceptions on?
-------------
PR: https://git.openjdk.org/jdk/pull/11443
More information about the hotspot-runtime-dev
mailing list