On Wed, Mar 9, 2016 at 4:18 PM, Marc-André Lureau <mlureau@xxxxxxxxxx> wrote: > > Hi > > ----- Original Message ----- >> On Wed, 2016-03-09 at 16:02 +0100, Fabiano Fidêncio wrote: >> > Marc-André, >> > >> > On Tue, Mar 8, 2016 at 11:33 PM, Marc-André Lureau <mlureau@xxxxxxxxxx> >> > wrote: >> > > Hi >> > > >> > > ----- Original Message ----- >> > > > This just adds a few basic tests to verify that we handle the usb-acl >> > > > helper >> > > > properly. The first patch is actually a bugfix for an issue that was >> > > > discovered >> > > > while writing the tests. >> > > > >> > > > The second patch provides some infrastructure for the tests. I'm not >> > > > sure >> > > > that >> > > > adding this acl-helper-path property is the most elegant approach here, >> > > > but >> > > > it is fairly simple. My initial approach was to re-compile >> > > >> > > Even simpler and without having to add a property in API etc, use an >> > > environment variable? >> > >> > And then we would have to setup the env-var when running make check? :-\ >> >> >> I think you could set the env variable from the test itself. I already do >> this >> for several env variables to control the behavior of the mock acl helper. >> >> The reason that I didn't use an env variable was because I thought it would >> be a >> little more reliable if the behavior couldn't be affected by the user's >> environment. But I can certainly use that approach if people prefer it. > > That is a limited risk, and could be considered a feature in some case (for study/debugging etc). So I'd use an env. Okay, Jonathon, I will be waiting for your patches using env var then. Please, feel free to push the first patch of this series if you want to. Best Regards, -- Fabiano Fidêncio _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel