Re: Regression tests failures on Windows Server 2019 - on master at commit # d816f366b

Hi,

On 2022-03-04 09:57:35 -0800, Andres Freund wrote:
> On 2022-03-04 09:46:44 -0800, Andres Freund wrote:
> > On 2022-03-04 09:30:37 -0800, Andres Freund wrote:
> > > I wonder if we're missing some steps, at least on windows, to make pg_ctl
> > > start independent of the starting shell?
> >
> > Sure looks that way. On windows, if I do pg_ctl start, then hit ctrl-c, the
> > server shuts down.
>
> Short term the easiest fix might be to start postgres for those tests as a
> service. But it seems we should fix whatever the cause of that
> terminal-connectedness behaviour is.
>
> I'm out for ~2-3h. I started a test run with using a service just now:
> https://cirrus-ci.com/task/5519573792325632 but I very well might have typoed
> something...

Seems to have worked for the first few tests at least. Unless somebody wants
to clean up that commit and push it, I'll do so once I'm back.

Perhaps pg_ctl needs to call FreeConsole() or such?

https://docs.microsoft.com/en-us/windows/console/closing-a-console

Or perhaps pg_ctl ought to pass CREATE_NEW_PROCESS_GROUP to CreateProcess()?
The lack of a process group would explain why we're getting signalled on
ctrl-c...

Greetings,

Andres Freund