RFR (XS) 8233698: GCC 4.8.5 build failure after JDK-8233530
Aleksey Shipilev
shade at redhat.com
Wed Nov 6 14:07:47 UTC 2019
On 11/6/19 2:17 PM, David Holmes wrote:
> On 6/11/2019 11:00 pm, Aleksey Shipilev wrote:
>> On 11/6/19 12:33 PM, Aleksey Shipilev wrote:
>>> Bug:
>>> https://bugs.openjdk.java.net/browse/JDK-8233698
>>>
>>> Our current RHEL-based CIs fail to compile jdk/jdk. That C++14 compat is the gift that keeps on
>>> giving! The fix is to get even deeper into the warning disabling story:
>>>
>>> diff -r bb2a436e616c src/hotspot/share/memory/operator_new.cpp
>>> --- a/src/hotspot/share/memory/operator_new.cpp Wed Nov 06 13:43:25 2019 +0800
>>> +++ b/src/hotspot/share/memory/operator_new.cpp Wed Nov 06 12:31:23 2019 +0100
>>> @@ -89,11 +89,13 @@
>>> fatal("Should not call global delete []");
>>> }
>>>
>>> #ifdef __GNUG__
>>> // Warning disabled for gcc 5.4
>>> +// Warning for unknown warning disabled for gcc 4.8.5
>>> PRAGMA_DIAG_PUSH
>>> +PRAGMA_DISABLE_GCC_WARNING("-Wpragmas")
>>> PRAGMA_DISABLE_GCC_WARNING("-Wc++14-compat")
>>> #endif // __GNUG__
>>>
>>> void operator delete(void* p, size_t size) throw() {
>>> fatal("Should not call global sized delete");
>>>
>>> Testing: gcc 4.8.5 build
>>
>> This passes jdk-submit too, so Oracle CI is not affected by this change.
>>
>> Trivial, right?
>
> I've given up trying to guess what might be trivial when it comes to these interactions with
> compilers :(
>
> The proof of this change is in the building and you have done that so it seems to be good.
I am pushing it then to unbreak our CIs. OK?
--
Thanks,
-Aleksey
More information about the hotspot-runtime-dev
mailing list