On Fri, 2015-10-09 at 13:05 -0700, Adam Williamson wrote:
> How are the screenshots rebased? Is it all manual work?
More or less. the openQA web UI has an interactive needle editor
which
lets you basically run a test and pause after each failed match, then
you can fiddle with the needle and re-run the match. I haven't used
it
much and never for doing a big screenshot rebase, so I can't say
whether it's more or less convenient than the way I did the last one,
which was to reproduce the tests by hand in a VM, take the
screenshots
from virt-manager, copy the json file and make any necessary
adjustments to the match area with the 'crop.py' needle editor (which
is an alternative tool upstream provides in one of their repos).
For the record I'm now doing the rebase using the interactive editor
and it's WAY better, wish I'd done the last one this way. =) It lets
you easily just apply the areas and tags from the existing needle to
the new screenshot, tweak a bit if you like, save under a new name, and
move on to the next one. Handy.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net