RFR: JDK-8319704: LogTagSet::set_output_level() should not accept NULL as LogOutput

Thomas Stuefe stuefe at openjdk.org
Fri Nov 10 12:12:01 UTC 2023


On Wed, 8 Nov 2023 11:41:54 GMT, Johan Sjölen <jsjolen at openjdk.org> wrote:

>> Trivial change to add an assert.
>> 
>> See [JDK-8319104](https://bugs.openjdk.org/browse/JDK-8319104)
>> 
>> `LogTagSet::set_output_level()` and `LogOutputList::set_output_level` should not accept NULL as LogOutput* argument.
>> 
>> They do, currently, which leads to delayed crashes later since the NULL output is registered in the output list.
>> 
>> That triggered [JDK-8319104](https://bugs.openjdk.org/browse/JDK-8319104), where - due to a fluke in C++ initialization order and a mislabeling of LogTagSet tests as "TEST", without "_VM" - `LogTagSet::set_output_level()` was called with `LogConfiguration::StdoutLog` which was still uninitialized (null). 
>> 
>> An assert would have saved investigation time.
>
> Oops, apparently this is a draft, sorry about that.

@jdksjolen would you agree this is Trivial?

-------------

PR Comment: https://git.openjdk.org/jdk/pull/16555#issuecomment-1805627195


More information about the hotspot-runtime-dev mailing list