Alvin Lai

I signed up for the 280Atlas Beta (one time fee of $20) and one of the first things I did was to create a native Cappuccino based app and take a windowed screenshot of it and guess what I found!

This isn’t really a bug, but just an observation on overflowing borders when taking windowed screenshots.

The border of the windowed screenshot is thicker than usual and overflowing, contrast this with a regular windowed screenshot:

Although the actual windowed screenshot looks fine:

In the meantime, check out how easy it is to create a native browser with Atlas here:

http://dailyjs.com/2009/11/16/atlas/

Wished that git came with colors?

Got this in a recent email from the Singapore Ruby Group:


git config --global color.branch "auto"
git config --global color.status "auto"
git config --global color.diff "auto"


Thanks Dinesh!

You can find out more here:
http://cheat.errtheblog.com/s/git

Nicely attached in Github:

Comment on existing Github issues:

and have them nicely show up too:

Check it out at http://freshlog.com

Enjoy!


./time_consuming_process.rb && say "Your time consuming process is done"


and you will hear Mac OS X tell you when its done while you go do other stuff.

You might notice that Selenium’s rake acceptance test would fail the second time because Firefox restored tabs that belonged to the first time you ran the acceptance test.

If you run Selenium for the first time round, you’ll end with 2 tabs, the second one being the results:

When you run the test again, the first tab (circled in red) which was created in the previous session (also the first Selenium test), restored by Firefox, also runs a parallael Selenium test, resulting test failures:

The way to fix this is to disable session restore in Firefox’s about:config:

But what if you love Firefox’s session restore feature? Create a separate Firefox profile for Selenium testing!

Go to Terminal.app then enter this to run Firefox’s Profile Manager:


/Applications/Firefox.app/Contents/MacOS/firefox-bin -ProfileManager


Then modify /vendor/plugins/selenium-on-rails/lib/selenium-on-rails/acceptance_test_runner.rb:

Under the start_browser method, add a


-P Selenium


An example:


command = "\"#{path}\" -P Selenium \"http://#{HOST}:#{@port}#{TEST_RUNNER_URL}?test=tests&auto=true&baseUrl=#{base_url}&resultsUrl=postResults/#{log}&multiWindow=#{MULTI_WINDOW}\""


That would set Selenium to use Firefox with your newly created Selenium Profile, where you can turn off Session Restore just for that profile and still enjoy it in your default Firefox profile.

Edit .git/info/exclude

and add your directories, i.e.:

/log/*
/db/schema.rb

After a git commit and push, the next command you might run is cap deploy. But how about running your tests?

If you don’t want to run any extra commands beyond git commit/push and cap deploy, you’ll want to look having your Capistrano script run your Rake tests locally too.

The Capistrano command “run” execute stuff remotely on your target deployed server, but since the deploy.rb Capistrano script is just a regular ruby script, you can just use “system()” or ““” (backticks) to execute commands locally.

Here’s an example:


# acceptance test
task :run_acceptance_test do
system("rake test:acceptance") # can use backticks `` as a shorthand also
exit if $? != 0
end
before "deploy:set_comment", :run_acceptance_test


Let me explain.


before "deploy:set_comment", :run_acceptance_test


would tell Capistrano to run your custom named task, in the case here, it’s “:run_acceptance_test”.


exit if $? != 0


Would check the error code of the previously run system command and if the error code (defined by $?) is not 0 (zero), it’ll cause the whole Capistrano script to exit.

So, if rake test:acceptance fails, the Capistrano script will exit and deployment will not happen at all.