On Sun, Aug 23, 2020 at 04:00:07PM -0300, Fabio Augusto De Muzio Tobich wrote:
> Em 23/08/2020 10:19, Adam Borowski escreveu:
> > On Fri, Aug 21, 2020 at 02:48:16PM -0300, Fabio Augusto De Muzio Tobich
> > wrote:
> >> Em 21/08/2020 05:54, Adam Borowski escreveu:
> >>> On Mon, Aug 17, 2020 at 09:12:51PM -0300, Fabio Augusto De Muzio Tobich
> >>> wrote:
> >>>> * Package name : eqonomize
> >>>> Version : 1.5.0-1
> >
> >>> Hi, it fails autopkgtest:
> Got the fail with sbuild + schroot and I believe your bet is correct.
> I've modified the test to fix this and get it running right, no matter
> the tool, and I believe it's good now.
It does pass now, but...
.--====[ debian/tests/openfile ]
export XDG_RUNTIME_DIR=$AUTOPKGTEST_TMP/tmp
cp debian/examples/housekeeping_book.eqz $AUTOPKGTEST_TMP
xvfb-run -a eqonomize $AUTOPKGTEST_TMP/housekeeping_book.eqz &
EQOPID=`ps ax | grep -e 'eqonomize.*housekeeping_book.eqz' | grep -v grep | awk
'{print
if [ -z $EQOPID ]; then
exit 1
else
kill -9 $EQOPID
fi
exit 0
`----
It's racey: will fail if "ps" runs before xvfb-run progresses past the point
of spawning the child process.
And then, what is this test even supposed to check? The only thing is does
is checking that eqonomize doesn't exit immediately -- which, for a GUI
program with a ton of library dependencies, probably means it won't even
enter main() by the time "ps" does its thing, still being in the process of
loading libraries.
So this test is pointless even if it passes, and I'd recommend either:
* adding a few seconds of sleep between xvfb-run and ps (so it at least has
a chance to pop up a window / crash), or
* dropping this test altogether
Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁
⢿⡄⠘⠷⠚⠋⠀ It's time to migrate your Imaginary Protocol from version 4i to 6i.
⠈⠳⣄⠀⠀⠀⠀
