On Fri, 5 Mar 2021 07:36:26 GMT, Yasumasa Suenaga <ysuen...@openjdk.org> wrote:
>> test/hotspot/jtreg/serviceability/sa/sadebugd/ClhsdbAttachToDebugServer.java
>> line 95:
>>
>>> 93: out.shouldMatch("^java/lang/Object @0x[0-9a-f]+$"); // for
>>> "class java.lang.Object"
>>> 94: out.shouldMatch("^java/lang/String @0x[0-9a-f]+$"); // for
>>> "class java.lang.String"
>>> 95: out.shouldHaveExitValue(0);
>>
>> I think you should do more error checking here. If you look at
>> ClhsdbLauncher you will see:
>> // This will detect most SA failures, including during the attach.
>> oa.shouldNotMatch("^sun.jvm.hotspot.debugger.DebuggerException:.*$");
>> // This will detect unexpected exceptions, like NPEs and asserts,
>> that are caught
>> // by sun.jvm.hotspot.CommandProcessor.
>> oa.shouldNotMatch("^Error: .*$");
>> I think you could use something similar here.
>
> I added them in new commit.
> BTW I found out to check warnings which relate to `-Xcheck:jni` at
> ClhsdbLauncher. I did not add them to ClhsdbAttachToDebugServer because they
> are not the issue which relates to debug server. Is it ok?
No need for the `-Xcheck:jni` checks. I added that recently because we did have
a bug, but I think we get more than enough testing with it already when running
the ClhsdbLauncher tests.
-------------
PR: https://git.openjdk.java.net/jdk/pull/2773
