See pull request for details. Summary:
* Reduce dependencies between CI jobs, allowing more of them to run in parallel.
* Use Linux for as many jobs as possible.
* Merge matching "-ndll" and "-samples" jobs. This way, the latter doesn't have to waste time recreating the former's environment, and the workflow summary is simplified. Downside: this delays package-haxelib.
* Update to a newer macOS version, as the old version massively slowed down Homebrew.
* Skip the HelloWorld test for Mac and iOS, as those two can build slowly under some circumstances, delaying package-haxelib. These could be restored later, if performance seems good enough.
* Add `HAXE_VERSION` environment variable, so future updates will only have to make one change. Unfortunately, doing the same for the Windows/Mac/Linux versions would severely hurt performance, so those remain hard-coded.
* Split up some steps within jobs for better organization.
Code intelligence should always use the newest hxml content, so the fallback mode where the hxml content is generated, instead of loaded from an existing .hxml file, should be used when the project file is newer.
For instance, if the user changes any file/dir paths in their project file, continuing to use the existing .hxml file could lead to confusing error messages that still reference the old/cached file paths. It should always use the latest paths or other values from the project file. It should be considered a bug to use the old cached paths.
Previously, as a workaround, the user would need to clean or build their project again to get updated .hxml files. It might also require restarting their editor/IDE too. Bad developer experience when we can detect this case automatically.
Now if you need to add something to the `<application />` or `<activity />` tags, you can do it from project.xml. This will require new documentation, however.
Fixes the following exception when we don't have permissions:
> Failed to execute 'getGamepads' on 'Navigator': Access to the feature "gamepad" is disallowed by permissions policy.
By catching the exception, it should now behave the same as older browsers, where navigator.getGamepads() doesn't exist at all.
In the future, it might make sense to set a flag if navigator.getGamepads() throws, and skip calling it more than once. However, we may want to listen for some kind of browser event that indicates that permission was granted later, and clear the flag when appropriate. Perhaps the gamepadconnected event?
---------
Co-authored-by: Josh Tynjala <joshtynjala@bowlerhat.dev>
