Driver path is saved in SeleniumWebdrivers.json relative to Solution location but restored in path relative to project file.

How to reproduce

1. Install Tellurium.SeleniumDriverInstaller nuget package
2. Invoke in nuget packge manager console

Install-SeleniumWebDriver -Browser Chrome

3. Remove Driver directory from your project
4. Invoke command

Restore-SeleniumWebDriver

Result: Restored webdriver is not in the same directory as the one which was deleted

Source of SeleniumDriverInstaller script are available here https://github.com/cezarypiatek/Tellurium/blob/master/Src/Tools/SeleniumDriverInstaller.psm1

There is a need to make Dashboard as a self-hosted website. Migrate VisualAssertion.Dashboard to ASP.NET Core. It should work as a window service https://codeopinion.com/self-host-asp-net-web-api-as-a-windows-service/

Create browser extension that allows to stream page content during test session
https://www.html5rocks.com/en/tutorials/streaming/screenshare/

https://code.google.com/archive/p/aost/

Now Drivers are *.exe files and when I ran Sample for first time, test failed cause driver was out of date.
With having them in Nuget we can easily update them by updating nuget package version to newest.

Dashboard -> patterns

https://stackoverflow.com/questions/43541925/how-to-set-the-browser-window-size-when-using-google-chrome-headless

Consider if Chrome headless should be added as a separated browser to available browsers list.

https://gist.github.com/addyosmani/5336747

1. Discover all available controllers and actions
2. Extract from driver all request addresses

@cezarypiatek What do you think about this idea?
Latest builds are failing if we want to fix them then why not to keep everything in one place (code + build pipelines in github)?

https://www.w3.org/WAI/intro/aria

1. Tree https://www.w3.org/TR/wai-aria-1.1/#tree
2. List https://www.w3.org/TR/wai-aria-1.1/#list
3. Table https://www.w3.org/TR/wai-aria-1.1/#table

Take account aria-labelledby attribute

The following command should install the newest version of webdriver but does not (random version is selected). The same for others drivers.

Install-SeleniumWebDriver  -Browser Firefox -Version v0.19.0 -Platform win64

However this command list all driver versions in proper order

Get-SeleniumWebDriverVersions -Browser Firefox

Driver  Version     Platform
------  -------     --------
Firefox v0.19.0     arm7hf  
Firefox v0.19.0     linux32 
Firefox v0.19.0     linux64 
Firefox v0.19.0     macos   
Firefox v0.19.0     win64   
Firefox v0.18.0     arm7hf  
Firefox v0.18.0     linux32 
Firefox v0.18.0     linux64 
Firefox v0.18.0     macos   
Firefox v0.18.0     win32   
Firefox v0.18.0     win64   
Firefox v0.17.0     arm7hf  
Firefox v0.17.0     linux32 
Firefox v0.17.0     linux64 
Firefox v0.17.0     macos   
Firefox v0.17.0     win32   
Firefox v0.17.0     win64   
Firefox v0.16.1     arm7hf  

Source of SeleniumDriverInstaller script are available here https://github.com/cezarypiatek/Tellurium/blob/master/Src/Tools/SeleniumDriverInstaller.psm1

Something like "mark as pattern" for given region

Create fullpage screenshots in Chrome >59 according to this solution implemented in Java:

https://stackoverflow.com/questions/45199076/take-full-page-screen-shot-in-chrome-with-selenium

Please add it to wiki

http://selenium-python.readthedocs.io/faq.html#how-to-auto-save-files-using-custom-firefox-profile
https://stackoverflow.com/questions/27824124/how-to-change-file-download-location-in-webdriver-while-using-chrome-driver-fire

Map chromePrefs = new HashMap();
chromePrefs.put(“profile.default_content_settings.popups”, 2);
chromePrefs.put(“download.default_directory”, getFilePath());
ChromeOptions options = new ChromeOptions();
//Map chromeOptionsMap = new HashMap();
options.setExperimentalOption(“prefs”, chromePrefs);
options.addArguments(“–test-type”);
capability = DesiredCapabilities.chrome();
capability.setCapability(“chrome.switches”, Arrays.asList(“–start-maximized,–ignore-certificate-error”));
capability.setCapability(ChromeOptions.CAPABILITY, options);
//capability.setCapability(ChromeOptions.CAPABILITY, chromeOptionsMap);

Now it does not work for SampleUITest for Firefox.

There is need to switch to geckodriver:

http://stackoverflow.com/questions/38916650/what-are-the-benefits-of-using-marionette-firefoxdriver-instead-of-the-old-selen/38917100#38917100

I have tried but had not luck, and my time run out.

@cezarypiatek btw which version of FF (x86, x64) are you using? There are different drivers and I think we should stick to the one that you have.

Things to accomplish this issue:

1. Create a separated nuget package with DownloadDrivers.psm1 (name proposal: Tellurium.SeleniumDriversInstaller)
2. Add parameter that allows to select driver architecture (x86 and x64)
3. Add parameter that allows to select driver version
4. Add command that allows to list all available versions of given driver
5. Configure package deployment

The purpose of this issue is to simplify driver installation from the official sources.

VisualAssertion should be able to detect that more than screenshot is affected with the same change (ex. layout change). It could be very helpful to mark as pattern all test result affected with the same change.

Cannot create driver for Firefox 47.0.1

System.NullReferenceException : Object reference not set to an instance of an object.
   at Tellurium.MvcPages.BrowserCamera.Lens.BrowserCameraLensFactory.IsHeadlessFirefox(RemoteWebDriver webDriver) in C:\projects\maintainableselenium\src\MvcPages\BrowserCamera\Lens\BrowserCameraLensFactory.cs:line 45
   at Tellurium.MvcPages.BrowserCamera.Lens.BrowserCameraLensFactory.CreateLensForAutoMode(RemoteWebDriver webDriver) in C:\projects\maintainableselenium\src\MvcPages\BrowserCamera\Lens\BrowserCameraLensFactory.cs:line 40
   at Tellurium.MvcPages.BrowserCamera.BrowserCameraFactory.CreateNew(RemoteWebDriver driver, BrowserCameraConfig cameraConfig) in C:\projects\maintainableselenium\src\MvcPages\BrowserCamera\BrowserCameraFactory.cs:line 10
   at Tellurium.MvcPages.BrowserAdapter.Create(BrowserAdapterConfig config, RemoteWebDriver driver) in C:\projects\maintainableselenium\src\MvcPages\BrowserAdapter.cs:line 58
   at Ailleron.EFL.UITest.UITests.RunTestForCategory(String categoryName, Action`1 testAction) in C:\TeamCityAgents\AgentMoa1\work\eace9c548cfcf1b2\src\EFL\EFL.UITest\UITests.cs:line 55

1. Update nuget package to Selenium 3
2. Update script for downloading webdrivers (add geckodriver)

It could be convenient to install Dashboard application https://github.com/cezarypiatek/Tellurium/tree/master/Src/Dashboard using dedicated installer instead of manually executing the following steps:

1. Extract Dashboard files into installation directory
2. Make modification to configuration files:
   • set port (appsettings.json)
   • set connection string to database (app.config)
3. Register app as windows service

Currently this app is released as zip file. I think about replacing it (or add as and alternative) with installer build with WIX http://wixtoolset.org/documentation/

Registering Dashboard application as a windows service is already implemented with Topshelf .
https://github.com/cezarypiatek/Tellurium/blob/master/Src/Dashboard/Program.cs

TelluriumDashboard.exe install -servicename "TelluriumDashboard"
TelluriumDashboard.exe start

1. Upload Error Screenshot as a build artifact
2. Generate file with report which could be attached as a custom tab
3. Add report tab for VisualAsserions
4. Add Unique id for test result

This file should be saved in project directory instead of solution directory (every project should be able to have own version of SeleniumWebdrivers.json).

Source of SeleniumDriverInstaller script are available here https://github.com/cezarypiatek/Tellurium/blob/master/Src/Tools/SeleniumDriverInstaller.psm1

For Chrome and Opera try to use DOM.setFileInputFiles from RemoteInterface:
https://chromedevtools.github.io/devtools-protocol/tot/DOM/#method-setFileInputFiles

Sample usage could be found here:
cyrus-and/chrome-remote-interface#164

More about Chrome RI:
https://github.com/cyrus-and/chrome-remote-interface
https://github.com/MasterDevs/ChromeDevTools

It could be helpful to mark as pattern all selected tests