On 5/17/19 4:25 PM, Jie Fu wrote: > JBS: https://bugs.openjdk.java.net/browse/JDK-8224112 > Webrev: http://cr.openjdk.java.net/~jiefu/8224112/webrev.00/ > > java/util/Base64/TestEncodingDecodingLength.java may fail due to insufficient > memory when our test > machine is busy. > It might be better to skip it when the available memory is insufficient. > Could you please review it and give me some advice?
I am against special casing like that: it may give you the false sense of passing the test when it is actually broken. In other words, having the infrastructure-induced test failure is better than having the test that allows bugs to creep in. Get more memory on a test machine, and maybe lobby for jtreg to unhoard some [1]. Thanks, -Aleksey [1] https://bugs.openjdk.java.net/browse/CODETOOLS-7902454
