|
[
Permalink
| « Hide
]
Bert Freudenberg added a comment - 31/Aug/09 02:43 PM
Added changeset with proposed solution.
In addition to Bert's patch, this is required. And I found out that I did almost the same thing 15 months ago when I filed out the patch from my image.
Published as 2259swapCtrlAltAt6-yo and 2260ctrlAltDefault-bf.
Oh, but with these changes swapping control and alt takes effect even in the dev image on Windows (and presumably on Linux), which is absolutely not desirable.
Why would that not be desirable? Ctrl-C is the standard for both Windows and Linux. Experienced developers can switch it all they like, but IMHO we should default to the platform convention. It's been far too long we've been violating them.
One thing is that the swapping is not consistent; (I thought more keys but) apparently only 'm' is the inconsistent one, so maybe somebody can fix it and I have less reason to object, but still I certainly get confused by the change.
Hmm, ^M is 13, maybe something is special-casing a carriage return somewhere? And are you saying it does not get swapped?
I'm not going to spend whole time to try it, but the issue is in the VM an dI think the code around:
case WM_SYSCHAR: /* Note: VK_RETURN is recorded as virtual key ONLY */ if(keyCode == 13) return 1; or: /* Special case: VK_RETURN is handled as virtual key *only* */ if(wParam == 13) return 1; This is not visible to the etoys-user (hopefully) and breaks develop cycle, can we just include mine and fix further in 2010? 2259swapCtrlAltAt6-yo actually breaks the swapping on Linux. Only when I remove the "char ~= (evt at: 6) ifTrue: [^ self]" line it works.
Pushed fix removing the offending line as 2266swapCtrlAlt-bf. Please test on Windows.
Wow, ok. Not sure what happens when the keyboard is configured to enter greek for example...
For windows, I just would change the default setting for the dev with ctrlAltDefault2-yo.2.cs. Shouldn't the dev image be as close as possible to the user image? If developers do not feel the pain, stuff won't get fixed.
> Shouldn't the dev image be as close as possible to the user image?
In a sense, yes. But I'd rather spend my limited time here to fix real user visible bugs but not just feel pain. Fine by me. feel free to push.
Here is story from a person who is running Etoys workshop very actively in Japan;
He is teaching Etoys in elementary school. One of objectives of the course is to gain basic literacy of computer operation; For some kids the course is the first experience to play with coumputer. In that context Etoys is difficult to learn becuase of its UI with unneeded uniqueness. We could make Etoys less worse if we stick to platform standard where possible. |
|||||||||||||||||||||||||||||||||||||