Hey anyone here? Using chromedriver extension by DMore, having a few issues:
Works fine on local (mac os) if i run chrome in none-headless mode
If i use on local in headless mode then my first scenario always fails bu then all of the others pass
When I use on CI (circle CI using 7.2.17node-browsers image all scenarios fail on same issue as first on local which is "StreamReadException" no other info
(Session info: chrome=74.0.3729.131)
(Driver info: chromedriver=2.43.600210 (68dcf5eebde37173d4027fa8635e332711d2874a),platform=Windows NT 10.0.17763 x86_64) (WebDriver\Exception\UnknownError)
Foo service, that you can just do behat --init FooService (or w/e it is)
Unable to find a matching set of capabilities
Build info: version: '3.141.59', revision: 'e82be7d358', time: '2018-11-14T08:25:53'
System info: host: 'QC06', ip: '172.16.9.15', os.name: 'Windows 10', os.arch: 'amd64', os.version: '10.0', java.version: '1.8.0_211'
Driver info: driver.version: unknown
remote stacktrace:
I have the latest version of selenium and firefox... What am I missing here?
I have a few of Ciaran McNulty's videos and for the most part I think I understand that tests should be "declarative", not "imperative". However, I am trying to test an API which uses events to drive an infrastructure service, but I can't really think of good tests. The best I can think of are tests like
Given I have payload ...
When I POST ...or
Given I have a machine
When I set the machine to process
Then it will be "processing"The problem with the second one is that it's a clear attempt to transcribe a state machine to a gherkin feature. I once replied to a state-machine question on SO with the example of an oven where in the idea was that an oven is HEATING when the oven is ON and below a temperature. With that simple scenario you can describe your tests with business rules and naturally come up with a state-machine as an after-thought. But in this kind if situation, the criteria for actions such as moving the machine to a processing state is just "I want the machine to start processing" ... so their aren't really any business rules around it. It is a requisite that you move the machine through states via the API, but that's about it.
Idk, maybe it's foolish to avoid trying to write a test that just checks state. There is that at least:
Given the machine is in state A
And I move the machine to state B
When I move the machine to state C
Then it will be in state CIt just doesn't seem like a super productive ruleset.
fillField fails because Behat\Mink\Element::findAll returns null, and I get an ElementNotFoundException. I took a look at the generated xPath, and it looked really over-complicated(but might still have been valid).I opened an issue on GitHub so that the solution to this doesn't end up lost in chat: Behat/MinkExtension#345
fillField failing - it's because ChromeDriver is defaulting to W3C mode, which isn't supported yet. Need to disable W3C mode in the behat.yml file (or however you're configuring your behat instance)
ArtifactExceptionInterface are thrown. I have afterStep/afterScenario hook that can take screenshots, but at that point it's too late - formatter already has gotten it's exception and I cannot access it from AfterStepScope (it's Behat 3). I could also throw my own exceptions, but most basic steps are from MinkContext and RawMinkContext which use WebAssert class that throws it's own exceptions (in private methods).All that said, what would be the best way to handle this? Make edits to WebAssert and PR it into MinkExtension project? Somehow alter formatter and make a PR there?
