25 Apr 2024 |
diemol | How can I update the code in the tests? I thought I did the right thing, but it did not work. | 15:24:35 |
titusfortner | I can't look at anything until this evening | 15:26:10 |
diemol | I will merge the code samples and we fix it later, then | 15:29:15 |
p0deje | diemol What did you do? | 15:47:32 |
p0deje | I could probably help if you share your changes | 15:47:45 |
diemol | I merged titusfortner PR 😄 | 15:48:02 |
diemol | So, this is the PR https://github.com/SeleniumHQ/selenium/pull/13386 | 15:48:16 |
diemol | And this is the change I did in the examples https://github.com/SeleniumHQ/seleniumhq.github.io/commit/936f95e7fed141b49c34d695c81fc6b649df127a#diff-5b4f16008b88f9f27e[…]ffa57598e503497058571af | 15:49:13 |
diemol | Based on this https://github.com/SeleniumHQ/selenium/pull/13386/files#diff-3dd830abe159e616a7e2ef063429e960febd5fec88c9fc4616871cf637468b9d | 15:49:41 |
diemol | And that change errors in the same error the Rails folks get | 15:50:10 |
diemol | (btw, I spent two days enabling tests on our nightly builds, so this type of thing should be caught early) | 15:51:20 |
titusfortner | I should have put all my PRs into draft before I left :) | 16:54:01 |
diemol | All the other languages worked. I guess you need to improve in Ruby :trollface: | 17:05:25 |
titusfortner | Well, I was working to make them all behave more similarly when I left | 17:11:17 |
p0deje | Shall we return DriverFinder.path ? It’s breaking Rails and RubyGems | 17:32:22 |
p0deje | And then properly deprecate it because it’s in fact public API 🙂 | 17:32:42 |
titusfortner | I really should put everything I do in draft until I'm sure it's ready to merge. I'll jump back into looking through everything next week, but star east is going to eat some of my time | 17:46:00 |
diemol | Those PRs were already reviewed and ready to merge. The piece missing these days is test Selenium Manager in our CI. Now we have that with the nightly builds. | 18:03:12 |
diemol | https://github.com/SeleniumHQ/selenium/issues/13876 | 20:01:50 |
p0deje | titusfortner Did you want to have DriverFinder.new accept service class or service instance? | 21:09:38 |
jimevans | Well, that was kinda fun. WebDriverBiDi.NET now has a somewhat palatable way to surface exceptions to the user (not a straightforward feat with .NET async code). This will be especially useful for users who have async event handlers that might throw an exception, where the default behavior would be to silently swallow the exception*.
*Well, to be strictly precise, the exception would be thrown on the synchronization context of the thread running the event handler, which, because the signature is async void , doesn't have anything to await on, so the exception never gets propagated. | 21:17:30 |
p0deje | Anyway, I’ve prepared a fix but let me know if you’d like that to be different https://github.com/SeleniumHQ/selenium/pull/13877 | 21:19:58 |
diemol | If you want to test it, you can generate a nightly from that branch and then kick the Ruby tests in the examples repo | 21:30:50 |
p0deje | I tested it locally, should be good | 21:47:54 |
p0deje | I’m just not sure what API should look like | 21:48:23 |
titusfortner | I'll look tonight | 21:49:52 |
titusfortner | Thanks | 21:49:55 |
diemol | https://www.selenium.dev/blog/2024/selenium-4-20-released/ | 21:51:35 |
p0deje | Ok, I think we should use service instance, that’s what you have in specs. I’ll go ahead and merge it | 21:57:45 |
p0deje | I’ll then prepare 4.20.1 release and update examples too | 21:58:05 |