RFR: 8345883: Relax system property "stackwalk.debug" in StackStreamFactory to be case insensitive
Eirik Bjørsnøs
eirbjo at openjdk.org
Tue Dec 10 16:35:37 UTC 2024
On Tue, 10 Dec 2024 16:01:13 GMT, Roger Riggs <rriggs at openjdk.org> wrote:
> But there is no need for boolean properties to be case insensitive.
Indeed. Seems like a matter of style choice, or perhaps often just determined a bit by chance or by the APIs picked to perform the evaluation. Some like their tabs, other spaces :-)
> This feels like a bit of high overhead thrash for the sake of consistency.
Isolated, this PR certainly does not carry its weight. I picked this property as an inital warm-up exercise because it is undocumented, for debug purposes only and thus is a low risk, trivial change.
If you think we should not try to reach for consistency in case insensitive interpretation, then I don't think it makes sense to go through with this isolated change.
For context, the remaining four boolean system properties with a case sensitive interpretation are:
InetAddress:
"java.net.preferIPv4Stack"
ReflectionFactory:
"jdk.reflect.useNativeAccessorOnly"
"jdk.disableSerialConstructorChecks"
Win32ErrorMode:
"sun.io.allowCriticalErrorMessageBox"
-------------
PR Comment: https://git.openjdk.org/jdk/pull/22659#issuecomment-2532228621
More information about the core-libs-dev
mailing list