removed headless: false from a test
#193
Conversation
The test 'should return actual viewport size' was using `headless: false` which doesn't mean anything as it was always running on docker which is headless by default. If you'd try to run this test as-is without docker it'll fail. Removing this feature fixes this issue.
|
@itaibh Your understanding of the test is wrong. It has to be headful, because in headless mode it will be possible to actually resize the browser to 5000x5000 size, and we wouldn't know if the value we got is an actual size or just the value provided by the user. |
|
The fact that it fails for you without docker should make you think about what is wrong with the SDK, not try to fix the test. I am not sure what you meant when you said that in docker it is headless by default, with this flag it shouldn't be headless this is the whole point. |
|
Ok. We had a call, where Itai shoved that this test doesn't do what I thought it does. Probably it happens since we changed the behavior of open, to actually throw when the viewport size couldn't be set to the required size. So since this change, this test doesn't tests anything. |
The test 'should return actual viewport size' was using
headless: falsewhich doesn't mean anything as it was always running on docker which is headless by default. If you'd try to run this test as-is without docker it'll fail. Removing this feature fixes this issue.