Closed joankaradimov closed 7 years ago
Nice find! I wonder if we should just do this in the WebPageManager
constructor, instead of calling back from WebPage
? We do other initialization and resetting from the WebPageManager
, so that seems appropriate -- so long as it can be set up before the first WebPage
is created.
I wonder if we should just do this in the WebPageManager constructor, instead of calling back from WebPage?
That was my initial thought too. Unfortunately I could not getting it working.
This test:
Capybara::Webkit::Driver localStorage works clears the message after a driver reset!
... is failing unless the cache initialization happens between the connect
s and the createWindow
call in the WebPage constructor. The first call to the constructor to be precise. I could not determine the exact cause.
Ok, I moved some code around. The cache initialization still happens outside the WebPageManager
constructor. Bit it is slightly better now, I think.
Thanks!
I am trying to get Capybara Webkit to run on my environment (Windows 10, 64 bit with JRuby 9.1.7.0). I use MSYS2 which [among other things] provides a MinGW environment with QT5 binaries and build tools. Everything is 64 bit.
Getting the code to build is fairly straightforward once the environment is all set (with the exception of this issue in Capybara Webkit). But running
bundle exec rake
results in multiple test failures.I was able to isolate a simple scenario that fails. The steps are:
I found this issue in the webkit bugtracker in which it is stated that:
... which is exactly what is happening for me. The build that I produced is a debug one. So I moved the cache initialization away from the WebPage constructor.
I suspect that Linux/macOS builds are release ones and are not affected by this. Still if this is an issue as the guys over bugs.webkit.org claim, this best be fixed everywhere.