add simple integration test with jsdom#169
Conversation
|
I can look into having mochify skip the test same as node 4 if you want to keep the test. |
|
|
|
See my change here for the |
|
@mantoni Yes, that part is included in the The issue now (at least in this PR) is that |
|
A sorry. The silly app I was viewing this in shows it as an issue instead of a PR 🙄 |
460795c to
89c4524
Compare
|
@fatso83 any way of triggering the integration tests from a PR? I wonder if the last commit fixes the issue |
|
@SimenB I don't think we have a way of triggering the full suite of Saucelabs tests on the PRs. It's been a long time, but I think the reason why we chose not to do it was that it would open us up to exposing the secret key (or something like that) by aggressors adding PRs that would read out environment variables. Or something like that ... @mantoni might remember it better 😊 Anyway, I see you have your email listed on your profile, so I'll send you what you need so you can do it. In other news ™️ the last commit failed on PhantomJS. |
|
This is a Travis security restriction, unfortunately. I have a general question regarding this RP: Wouldn't switching to the latest Mochify with headless Chrome make a JSDom test superfluous? Or is JSDom something we want to have explicitly tested? |
|
@mantoni Given the current work that is done on Jest to try and integrate lolex, which would touch a lot of users, I think it makes sense to support such a big target. |
|
@fatso83 @benjamingr this now passes CI, and |
|
👍 Looks good to me. |
| @@ -0,0 +1,57 @@ | |||
| /** | |||
There was a problem hiding this comment.
Sorry for forgetting this, but could you just skip the comment here about copyright and all that?
|
A release of this would be awesome 😀 |
Saw this was missing when reviewing DefinitelyTyped/DefinitelyTyped#56961 Bit funny that it has not been added until now, given it was added way back in #106 😄 (And later #169 #248 and others).
Includes both #162 and #168, plus an integration test using JSDOM.
This doesn't really work though, aswindow.performanceisreadonlyin JSDOM. This matches the spec (or, at least it matches the text on MDN), but it does not match the behavior of Chrome.The second commit skips fakingperformance, and the test passes. I don't like that, though.window.performanceisreadonly, so instead of using simple assignment, this PR usesObject.defineProperty.