Closed NfNitLoop closed 8 years ago
So, it looks like the ruby 1.x and 2.2 testers failed to install some dependencies.
The ruby2 tests are failing because the specs are trying to fake out the old xvfb_running?
by mocking CliUtil methods. But the updated version of xvfb_running?
doesn't even get that far. Since the spec is mocking launch_xvfb, @pid
never gets set, and xvfb_running?
will always return false.
I'm scratching my head a bit at the easiest way to update the tests to handle this.
Ooh, I see. I accidentally changed the signature of xvfb_running?
. It used to return true if there's an Xvfb running for display
, even if we didn't start it. Which is the desired behavior if we want to re-use that display. Fixing!
Yay! Ruby2 tests are passing.
The other 2 environments still seem to be having problems installing their dependencies. I don't think that's due to anything I changed, but please poke me if there's anything I can do to help get this ready for merge!
Thanks for writing Headless!
Anything possible to do here to help with getting this accepted/rejected?
I'm still around and willing to make any changes necessary to get this merged. (In the meantime, we're just using the fixed version over in https://github.com/NfNitLoop/headless/tree/fix-launch-race.)
Good to know, we're also gonna need to either use a workaround or a custom patch as we're seeing this (or something related) in the wild, so wanted to add an data-point and offer help in any way to get to an resolution.
@NfNitLoop @erkki I'll fix the tests and merge it this weekend.
@leonid-shevtsov much appreciated!
Any update here?
As it turned out, bundler on Travis was causing issues and that's the only reason the build was failing.
Merged and released in version 2.2.3.
This resolves #75