On Tue, 10 Dec 2024 11:16:55 GMT, Eirik Bjørsnøs <eir...@openjdk.org> wrote:
> Please review this PR which relaxes the interpretation of the system property > `stackwalk.debug` in `java.lang.StackStreamFactory` to be case insensitive. > > Motivation: > Only 5 of 83 boolean system properties in `java.base` have a case sensitive > interpretation. Relaxing these to be case insensitive will improve > consistency across the code base and simplifies our reasoning about such > system properties. > > The `stackwalk.debug` property was introduced when JEP-259 was integrated in > JDK-8143911. The property was not mentioned in the CCC for this change, > neither in the JEP. Seems like an undocumented, internal debug facility for > development and maintenance purposes. > > Risk: > This should be a low risk change. The property is used for debugging, and the > change will shift the interpretation of "TRUE" from false to true. Any user > specifying `-Dstackwalk.debug=TRUE` probably meant to enable it. > > Verification: > Manually verified that `-Dstackwalk.debug=TRUE` and `-Dstackwalk.debug=true` > enables debugging and that `-Dstackwalk.debug=false`, `-Dstackwalk.debug` and > `-Dstackwalk.debug=abc` does not enable it. I agree with Roger. The benefit of making this debug property case insensitive is so minor. For boolean property with false as the default, I think a good improvement would be to accept empty value (i.e. `-Dfoo`) as it's just a switch to turn it on. For example, `-Djdk.net.URLClassPath.disableClassPathURLCheck`, `-Djdk.net.URLClassPath.showIgnoredClassPathEntries`. I'm not suggesting this PR to do it as this debug property is only used during debugging and consistency argument does not really stand. ------------- PR Comment: https://git.openjdk.org/jdk/pull/22659#issuecomment-2532558476
