Should continue the project

[DiMiTriFrog]

You should continue to update the project. This project is too good to let it die. I have achieved time reduction of almost half compared to Selenium. The potential of asynchronous browser automator is brutal.

[dimaqq]

Please contribute :)

[DiMiTriFrog]

Please contribute :)

Any roadmap?

[ojii]

Any roadmap?

there's no formal roadmap, since the project does what we need it to do. however, there are a few things (some of this might be quite large projects) that I'd like to eventually see happen, in no particular order:

• Change the way requests are sent to use a queue, since webdriver is a sequential protocol and doing two simultaneous requests doesn't make sense.
• Drop IE11 support
• More comprehensive and less fickle test suite (this is probably the biggest annoyance I have with this, the tests are just not at a point where they fill me with confidence)
• Better "capabilities" support. Make strongly typed Python objects for capabilities for the different browsers (since a lot of issues opened are about "how do I pass this option/capability to this browser"). This still would need some sort of escape-hatch on top of the strongly typed objects for non-standard options and to remain forward-compatible.
• Related to the last one, does the service/browser split even make sense in the code? Wouldn't it make more sense for the main API to be webdriver(driver, capabilities) where the driver describes how to start/stop the driver (either a local chromedriver/geckodriver/... or a remote-driver) and capabilities is the capabilities (including which browser) that we want?
• Change logging. I personally really like structlog but looking at the issue tracker and talking to people, it seems I'm the only one actually thinking it's a good logging library. I won't remove it, but if someone were to open a PR that replaces it (probably with stdlib logging?) comprehensively, I'd be okay with that.
• Just look at the open issues and fix those