On 05/18/10 08:29 PM, Brock Pytlik wrote:
On 05/18/10 06:22 PM, Shawn Walker wrote:
On 05/18/10 08:16 PM, Brock Pytlik wrote:
On 05/18/10 06:10 PM, [email protected] wrote:
On Tue, May 18, 2010 at 05:56:13PM -0700, Brock Pytlik wrote:
Webrev:
http://cr.opensolaris.org/~bpytlik/ips-16001-v1/
Bug:
16001 gui test suite broken again
Instead of catching every exception on import, would it be possible to
do something like:
try:
import XXX
except (EnvironmentError, ImportError):
This would catch the cases that might cause the module to fail to
import, but shouldn't catch _everything_. I'm assuming, of course, that
we really don't want to catch everything, but I've been wrong before...
-j
I do think we want to catch everything. I don't really care WHY the gui
tests failed to run this particular time, I just want the rest of the
test suite to drive along after telling me that the gui tests couldn't
be run. For example, in this case it doesn't appear that either an
EnvironmentError or ImportError was raised. Instead, it seems like a
gtk.GtkWarning was raised. I think we've tried finesse approaches
several times now. Since this continues to be an issue, I'm for taking a
sledge hammer to the problem.
I'm perfectly fine with the sledgehammer, although I'm still confused
as to how rev 1890 would have caused an issue.
Turns out I was wrong about when this went bad. I looked at the log for
gui/testutils and the most recent putback matched with when I started
seeing this problem happen. But, apparently I didn't remember things
right or got something else wrong. Turns out the actual change was
either 1715 or 1716. I have no idea why I only started seeing this
recently.
+1 from me then.
Cheers,
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
