Although 1.6.5 has flown, please do follow up with your observations. There aren't a lot of Sparc machines between the committers, and I'd even thrown my x86-solaris VM's to archive, but I can restore them if we can narrow down what you are seeing.
It sounds from Yann's feedback that repeated tests are necessary to decide if the pass or failure is a fluke. On Tue, Sep 11, 2018 at 5:26 PM Dennis Clarke <dcla...@blastwave.org> wrote: > On 09/11/2018 05:48 PM, Yann Ylavic wrote: > > On Tue, Sep 11, 2018 at 11:33 PM Dennis Clarke <dcla...@blastwave.org> > wrote: > >> > >> testsock : Line 376: Cannot test if connect completes > >> synchronously > > > > This one really depends on the time taken by the system to connect > > localhost non-blocking. > > If it's immediate (that happens), the test is not relevant and this is > > what the above is saying. > > So this message may or not be printed on different runs. > > > >> testsockets : SUCCESS > >> testsockopt : Line 84: TCP isn't corkable > > > > Is TCP_CORK or TCP_NOPUSH defined on Solaris (in "netinet/tcp.h")? > > nope. > > $ grep -E "TCP_CORK|TCP_NOPUSH" /usr/include/netinet/tcp.h > > > Same Solaris version for 1.6.x and 1.5.x tests? > > yep. > > Lemme dig around here a bit. Something is fishy about these results. > > dc >
