RFR: JDK-8282593: JDK-8281472 breaks 32-bit builds and gtests

[Thomas Stuefe]

On Thu, 3 Mar 2022 12:55:23 GMT, David Holmes <dholmes at openjdk.org> wrote:

>> Hi, may I please have reviews for this trivial fix.
>> 
>> JDK-8281472 added a test_largeOptions.cpp gtest. Parts of it don't build for 32-bits, parts build but don't run. 
>> 
>> I propose to exclude the failing parts. AFAICS they check that numerical arguments are handled correctly when going beyond INT_MAX and INT_MIN, which is not an issue on 32-bit with 32-bit ints.
>> 
>> Thanks, Thomas
>
> BTW the test is checking that the option parsing code, which takes strings, correctly handles potential overflow/underflow so this is still an issue on 32-bit.

@dholmes-ora

One question. Do these tests actually modify JVM parameters, or do they just test the parser? To me, it looks like the former. If yes, then this should be changed, since individual gtests are supposed to be side effect free. In that case, TEST_OTHER_VM may be more suitable.

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

PR: https://git.openjdk.java.net/jdk/pull/7672