On Fri, Mar 29, 2019 at 09:53:51AM +0000, Daniel Gustafsson wrote:
> On Saturday, March 9, 2019 8:16 AM, Noah Misch <noah@leadboat.com> wrote:
> > I tested on Red Hat and on Windows Server 2016; I won't be shocked
> > if the test (not the code under test) breaks on other Windows configurations.
>
> IIRC there are Windows versions where Win32::Process::KillProcess is required
> for this, but thats admittedly somewhat dated knowledge. If the buildfarm goes
> red on older Windows animals it might be something to look at perhaps.
Since my second attempt at committing this (commit c098509), I fixed these
bugs in the new test file:
- MSYS-orchestrated mingw32 (buildfarm member jacana) failed the Perl kill(9,
...) calls[1]. For HEAD and v11, using "pg_ctl kill KILL <pid>" fixed that.
For v10 and v9.6, I disabled the test under msys. I can reproduce this with
Perl 5.28 from msys2. Its kill(9, ...) fails for any non-msys2 process (any
ordinary Windows process). [commits 947a350, 2bc0474]
- The regress.dll path needed MSYS-to-w32 path translation. [commit 9daefff]
- The changes to port selection in get_new_node() made Linux-specific
assumptions, so Windows builds had much less protection against port
conflict. [commit 4ab02e8]
- Got EPIPE when writing to stdin of a child process that exited too quickly.
[commit e12a472]
Things have been stable on the buildfarm for the last twelve days, so I think
this one is over.
[1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=jacana&dt=2019-04-13%2014%3A39%3A17