Closed mkarg closed 2 years ago
This seems to be unnecessary pushing the TCK on an edge. It has been reported to be an issue and unless the original reporter (@alwin-joseph) confirms there is no issue, I'd stay safe here.
This seems to be unnecessary pushing the TCK on an edge. It has been reported to be an issue and unless the original reporter (@alwin-joseph) confirms there is no issue, I'd stay safe here.
There was no evidence in the issue https://github.com/jakartaee/rest/issues/1088 that ServerSocket(0)
was returning port < 1024. The only two failing tests in the issue were fixed by the Jersey upgrade to M3 , those tests were not using the someFreeIpPort
method modified here.
This PR removes an unnecessary workaround: There is not need to check the port number, as nobody ever has seen any JRE on any OS that returned a non-ephemerial port. Oracle's OpenJDK team even disputes that this could happen ever, and there is no proof that there would be a bug. Hence we should not fix something that is not broken.
As these are no API changes, this is a fast-track review period of just one day as per our committer rules.