On Tue, 23 Jan 2024 13:15:46 GMT, Daniel Jeliński <[email protected]> wrote:
> The test was occasionally failing with TestNG timeout. When a TestNG test > times out, TestNG interrupts the test thread, and starts another test case in > a new thread. The timeout is invisible to JTReg, and the timeout failure > handlers are not run. > > This PR removes the TestNG timeout. It also removes `-ea -esa` from the test > command line (the test does not use `assert`), and reduces the amount of > synchronization objects used. > > I verified that the timeouts are now handled by JTReg. The test still passes. test/jdk/java/net/httpclient/MaxStreams.java line 220: > 218: // client issues MAX_STREAMS + 3 requests in total > 219: // but server should only see MAX_STREAMS + 2 in > total. One is rejected by client > 220: // counter c captured before increment so final > value is MAX_STREAMS + 1 Can we keep/add a trace here to say that counter is reset to 0? Otherwise looks good. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/17536#discussion_r1463453547
