keyDown: in NSMatrix subclass

[matt neuburg]

In <znu-55B7B3.11560309072003@news.fu-berlin.de> ZnU wrote:

> I'm trying to override keyDown: in an NSMatrix subclass so I can
> implement some reasonable behavior for arrow keys, but I never seem to
> receive any key events. The documentation isn't being particularly
> helpful. What's going on here? Are my events being eaten by the
> containing NSScrollView? If so, is there a way to get it to cut that
> out without subclassing it?

I presume you remembered to dick around with acceptsFirstResponder? When
that doesn't work, what I do is override NSWindow's sendEvent:. I test
for an NSKeyDown event and then ask the window's delegate what to do.
The window's delegate can look to see who is firstResponder and
implement special behavior if desired, or just report that we should
keep our hands off, in which case we call the super's functionality.
Perhaps this is overkill but it works great for me; I've used it to mend
many apparent incoherencies in Cocoa's key handling. m.

--
matt neuburg, phd = [email]matt@tidbits.com[/email], [url]http://www.tidbits.com/matt[/url]
REALbasic: The Definitive Guide! 2nd edition!
[url]http://www.amazon.com/exec/obidos/ASIN/0596001770/somethingsbymatt[/url]
Subscribe to TidBITS. It's free and smart.

[matt neuburg]

In <znu-27DE66.09010313072003@news.fu-berlin.de> ZnU wrote:

> I stuck:
>
> [[self window] makeFirstResponder:self];
>
> in acceptsFirstResponder. This seems to work.

That seems really dangerous to me. Are you quite sure that you couldn't
get the desired result by returning YES from both acceptFirstResponder
and becomeFirstResponder, which is the canonical way to do it? m.


--
matt neuburg, phd = [email]matt@tidbits.com[/email], [url]http://www.tidbits.com/matt[/url]
REALbasic: The Definitive Guide! 2nd edition!
[url]http://www.amazon.com/exec/obidos/ASIN/0596001770/somethingsbymatt[/url]
Subscribe to TidBITS. It's free and smart.

[ZnU]

In article <20030713091723225-0700@news.la.sbcglobal.net>,
matt neuburg <matt@tidbits.com> wrote:

> In <znu-27DE66.09010313072003@news.fu-berlin.de> ZnU wrote:

> > I stuck:
> >
> > [[self window] makeFirstResponder:self];
> >
> > in acceptsFirstResponder. This seems to work.

>
> That seems really dangerous to me. Are you quite sure that you couldn't
> get the desired result by returning YES from both acceptFirstResponder
> and becomeFirstResponder, which is the canonical way to do it? m.

Positive. If I simply do that, becomeFirstResponder never even gets
called.

--
"First, let me make it very clear, poor people aren't necessarily killers. Just
because you happen to be not rich doesn't mean you're willing to kill."
-- George W. Bush in Washington, D.C. on May 19, 2003

[ZnU]

In article <20030715082836994-0700@news.la.sbcglobal.net>,
matt neuburg <matt@tidbits.com> wrote:

> In <znu-E93F54.10495515072003@news.fu-berlin.de> ZnU wrote:

> > In article <20030713091723225-0700@news.la.sbcglobal.net>,
> > matt neuburg <matt@tidbits.com> wrote:
> >

> >> In <znu-27DE66.09010313072003@news.fu-berlin.de> ZnU wrote:
> >> > I stuck:
> >> >
> >> > [[self window] makeFirstResponder:self];
> >> >
> >> > in acceptsFirstResponder. This seems to work.
> >>
> >> That seems really dangerous to me. Are you quite sure that you
> >> couldn't get the desired result by returning YES from both
> >> acceptFirstResponder and becomeFirstResponder, which is the
> >> canonical way to do it? m.

> >
> > Positive. If I simply do that, becomeFirstResponder never even gets
> > called.

>
> But that's only because you're not doing anything to cause the NSMatrix
> subclass to be the first responder in the first place! It doesn't happen
> all by itself, you know.
>
> In other words, what I'm saying is, don't call makeFirstResponder inside
> acceptsFirstResponder - that's a sort of circular corruption of the
> framework - but do call it! In your window controller's awakeFromNib,
> for example, say something like this:
>
> [[myMatrix window] makeFirstResponder: myMatrix];

That works initially, of course. But if another view becomes the first
responder, there's no way to make the matrix the first responder again;
when it's clicked, acceptsFirstResponder gets called and then... nothing.
I suppose I could send the makeFirstResponder message from mouseDown
rather than from acceptsFirstResponder, but that seems functionally
identical in this case; there appears to be a one-to-one correspondence
between clicks on the view and calls to acceptsFirstResponder.

[snip]

--
"First, let me make it very clear, poor people aren't necessarily killers. Just
because you happen to be not rich doesn't mean you're willing to kill."
-- George W. Bush in Washington, D.C. on May 19, 2003