"GOT", but the "O" is a cute, smiling pufferfish. Index | Thread | Search

From:
Omar Polo <op@omarpolo.com>
Subject:
Re: got-build-regress.sh regress failure
To:
Stefan Sperling <stsp@stsp.name>
Cc:
Mark Jamsek <mark@jamsek.com>, gameoftrees@openbsd.org
Date:
Fri, 21 Apr 2023 13:25:49 +0200

Download raw body.

Thread
On 2023/04/21 12:59:33 +0200, Stefan Sperling <stsp@stsp.name> wrote:
> On Fri, Apr 21, 2023 at 12:40:33PM +0200, Omar Polo wrote:
> > : ===> regress/tog
> > : ==== log ====
> > : ./log.sh -q -r "/home/chiaki/tmp/"
> > : tog: standard input is not a tty
> > : diff: /home/chiaki/tmp/tog-test-log_hsplit_diff-6oaPkUmX/view: No such file or directory
> > : test failed; leaving test data in /home/chiaki/tmp/tog-test-log_hsplit_diff-6oaPkUmX
> > : tog: standard input is not a tty
> > 
> > I missed that now if ran from cron it dies because stdin is not a
> > tty.  ok?
> 
> Seems ok to me. I hope we are now really avoiding all tty-related ioctls
> in test mode (I believe we do) since pledge would terminate the process
> otherwise, enforcing that file descriptors point to an actual tty.
> But I assume tests are passing for you and so it will be fine.

I believe it was already the case.  This latest one in particular was
a sefeguard we added some time ago to prevent from things like `tog <foo'
since it is supposed to be used interactively.