I believe the usual behaviour is to make the window the minimum size such that no scrolling is required
Which debs are you referring to? Currently, Sublime Text 2 is distributed as tarballs only.
Do you mean that when you press escape, you're not being returned to where you were?
I like the idea in principle, but I don't think I could ever justify implementing it. A reasonable alternative may be to implement a command line tool that will connect to a remote instance of Sublime Text, so that you can still edit files on a server you're SSHed into, assuming your primary computer has a graphical display.
This may be fixed in 2027, please let me know if it happens again
I don't believe WinSparkle supports 64 bit applications, nor does it support updating portable versions.
Also, it doesn't save as many clicks as you'd hope, as it just downloads the installer automatically, you still have to click through the installer screens.
From other reports of this problem, my understanding is that if you resize the window, the problem goes away. Could you confirm that for me?
Also, if resizing the window does fix the problem, does the window only have to be resized once, and then its still fixed for subsequent layout changes, or does it have to be resized after each layout change?
For what it's worth, Alt+u,u does create a ü for me.
Have you changed any key bindings, or installed any plugins? Does alt+<key> work for you at all for inserting alternate characters?
That is the default behaviour... is it not working for you?
These signals, and more, are taken into account in the cost function: you can see this when filtering with a shorter string, such as 'app', or 'appc', which ranks application_controller first, as it should.
The issue you're seeing is that there are many ways to match 'appcon' against 'app/controllers/application_controller.rb', and the ranking function is giving up before finding the best way. I'll increase the number of iterations it's willing to do before giving up, which will fix the problem on in this instance.
Many thanks for this report, by the way: In the test cases I've got here, the ranking gives better results than any other similar systems, and as much as possible, I'd like to keep it that way.