seeking advice on why the page under test fails to...
# helpp
plain-garden-5374
07/25/2022, 4:23 PMseeking advice on why the page under test fails to load in headless chrome, but succeeds in electron and headed chrome: https://github.com/constantlyk/cypress-program-guide-demo
plain-garden-5374
07/25/2022, 4:24 PM(i put as much information about the issue as i could in the readme, but let me know if i can clarify anything else)
plain-garden-5374
07/25/2022, 11:07 PMupdated readme to include debug logs, which i'll also post here
* headless: https://pastebin.com/HSJJjxDi
* headed: https://pastebin.com/HnAiRbqf
n
nutritious-honey-65632
07/26/2022, 1:16 PMHi, I tried it but didn't manage to find anything useful. I just notice that I'm getting the same whitescreen running in cypress open if I block
api.asco.orgp
plain-garden-5374
07/26/2022, 4:32 PM@nutritious-honey-65632 thanks for taking a look! when you say "block", do you mean blocking that domain/url in the browser, or intercepting it through cypress, or something else?
n
nutritious-honey-65632
07/26/2022, 5:23 PMI blocked the domain using /etc/hosts file on Linux. You could probably do the same with cypress
nutritious-honey-65632
07/26/2022, 5:24 PMI would double check can I get the response from that request using headless browser
p
plain-garden-5374
07/26/2022, 5:37 PMand also to confirm, you are able to replicate the original bug when you clone the repo? it's not just on my machine?
n
nutritious-honey-65632
07/26/2022, 8:16 PMyes, I see white screen when running in run mode and flaky test in open mode, but it pass most of the time
p
plain-garden-5374
07/26/2022, 8:19 PMflaky in open mode -- i hadn't been seeing it flake, but if true then that suggests a dreaded race condition, especially considering one of the few differences in headless mode is that it's faster..... i did try increasing the timeout before, but that didn't resolve the white screen
n
nutritious-honey-65632
07/26/2022, 8:24 PMI tried the same, increasing timeout and still failed in run mode
nutritious-honey-65632
07/26/2022, 8:28 PMI tried to wait for
graphqlp
plain-garden-5374
07/26/2022, 8:29 PMyes, in fact, the original version of this test was actually waiting on the graphql request (i changed it to a dom element check for this example)
n
nutritious-honey-65632
07/26/2022, 8:29 PMit would be nice if you could see the network log somehow, then you will see the response code from that request
nutritious-honey-65632
07/26/2022, 8:30 PMlet me see something
nutritious-honey-65632
07/26/2022, 8:35 PMThis is the flakines that I can see sometimes in open mode
p
plain-garden-5374
07/26/2022, 8:37 PMah yes this thing
plain-garden-5374
07/26/2022, 8:37 PMit causes a very specific type of flake: it seems that the first test that is run in open mode will always fail due to that, but subsequent tests always succeed
plain-garden-5374
07/26/2022, 8:40 PMi believe that setting chromeWebSecurity: false gets around it, but in my experience setting that had no effect on the current white screen issue. perhaps they are related, perhaps not
plain-garden-5374
07/26/2022, 8:42 PMthere may also be a MIME-related error which at least in headed/open mode could be surmounted by setting modifyObstructiveCode: false, but again that didn't seem to fix the headless/run white screen
n
nutritious-honey-65632
07/26/2022, 8:59 PMcould be, everything else is beyond my domain of knowledge. I would try to find the tool that can capture full network log coming from the headless browser and check the responses. Then you can isolate which request is not succeeding
p
plain-garden-5374
07/26/2022, 9:03 PMthanks so much for investigating with me! good to know i'm not alone
2 Views