On Thu, 29 Jul 2021 at 14:22, sebb <seb...@gmail.com> wrote: > > On Thu, 29 Jul 2021 at 13:43, Gary Gregory <garydgreg...@gmail.com> wrote: > > > > On Tue, Jul 27, 2021 at 7:38 PM sebb <seb...@gmail.com> wrote: > > > > > > On Tue, 27 Jul 2021 at 16:05, sebb <seb...@gmail.com> wrote: > > > > > > > > Something strange is going on with the FTPSClientTest. > > > > > > > > One my macOS system, the original jks file works with all my Java > > > > installations: > > > > Java 8 (Oracle and AdoptOpenJDK) > > > > Java 11 (Oracle and AdoptOpenJDK) > > > > Java 15 (Oracle) > > > > Java 16 (AdoptOpenJDK) > > > > > > > > With the jre8.jks, only Java8 works; the others all fail with: > > > > java.io.EOFException: SSL peer shut down incorrectly > > > > > > > > With jre16.jks and java 16, I get > > > > java.io.EOFException: SSL peer shut down incorrectly > > > > > > > > However, when GitHub builds using the original JKS file, only Java 8 > > > > works. > > > > Java 11, 16 and 17 all hang when running FTPSClientTest > > > > > > Lots of debugging later, and it looks like the MINA server does not > > > always see the USER command, so that test is timing out. > > > However it does work sometimes (in all the JDK versions), so it does > > > not seem to be an issue with the JKS file (currently using the > > > original one). > > > > > > The code works OK on Jenkins (currently using JDK 16). > > > > > > Why is only GitHub failing? > > > > Could there be a subtle difference in OS and JRE version? > > It's not just the OS, because Java 8 works on GitHub with the same OS. > > Obviously something is different, otherwise we would not see a > consistent difference in behaviour. > > Question is: how to proceed? > > It seems to be associated with MINA embedded FTP server: is there an > alternative that could be used for the test?
I've found a work-round: add a short sleep (200 ms) before the USER command. I added it because I noticed that the the MINA ftpserver did not appear to receive the USER command. So perhaps the previous commands sometimes need a bit of time to catch up? I'm just guessing. > BTW, we still need to document how the original JKS file was created. This is still the case. > The new ones don't work reliably. I have deleted them. > > Gary > > > > > > > > > Any ideas? > > > > > > > > Sebb. > > > > > > --------------------------------------------------------------------- > > > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > > > For additional commands, e-mail: dev-h...@commons.apache.org > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > > For additional commands, e-mail: dev-h...@commons.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
