I’m trying to replicate the behaviour of the search field in iTunes, for looking up stock symbols and names. Specifically, as you start typing in the search field a popover appears with the filtered items. For the most part I have this working however what I can’t replicate is the way it handles first responder
I have my popover appear after three characters are entered. At this point the NSSearchField would lose first responder status and therefore I could no longer continue typing. The behaviour I would like is
the ability to continue typing after the popover appears
if scrolling through the items with the arrow keys, and then resume typing, you would continue from the last character in the Search field.
What I tried is subclassing NSTextView (use this as the custom field editor for the NSSearchField) and overriding
- (BOOL)resignFirstResponder
By simply returning NO, I can continue typing once the popover appears, but obviously I can’t select any of the items in the popover. So i tried the following, which returns YES if the down arrow or a mousedown event occurs.
#interface SBCustomFieldEditor ()
{
BOOL resignFirstRepond;
}
#end
#implementation SBCustomFieldEditor
- (id)initWithFrame:(NSRect)frame
{
self = [super initWithFrame:frame];
if (self) {
// Initialization code here.
resignFirstRepond = NO;
}
return self;
}
- (BOOL)resignFirstResponder
{
return resignFirstRepond;
}
- (void)keyDown:(NSEvent *)theEvent
{
if ([theEvent keyCode] == 125) {
resignFirstRepond = YES;
[self resignFirstResponder];
}
[super keyDown:theEvent];
}
- (void)mouseDown:(NSEvent *)theEvent
{
resignFirstRepond = YES;
[self resignFirstResponder];
}
This works for the mousedown event, but not the keydown event, furthermore this doesn’t address the issue, when the user resumes typing.
Any suggestions?
In the meantime I found an easy fix. Subclass your text view and implement - (BOOL)canBecomeKeyView. Always return NO there. It will be called only once when the popover is shown. You can work with the text view any time still.
Related
I need a NSButton that gives me 2 events, one when the button is pressed down (NSOnState) and one when the button is released (NSOffState) and so far i've got it working with the mouse (intercepting the mouseDown: event). But using a keyboard shortcut doesn't work, it fires a NSOnState once and then after a delay really often. Is there any way to get a button that fires NSOnState when pressed and NSOffState when released?
My current subclass of NSButton looks like this and unfortunately works using a delegate:
-(void)awakeFromNib {
[self setTarget:self];
[self setAction:#selector(buttonAction:)];
}
-(void)mouseDown:(NSEvent *)theEvent {
[_delegate button:self isPressed:YES];
[super mouseDown:theEvent];
}
-(void)buttonAction:(id)sender {
[_delegate button:self isPressed:NO];
}
The button can be set to send mouse events at both times by using -sendActionOn::
[self.button sendActionOn: NSLeftMouseDownMask | NSLeftMouseUpMask];
Handling keyboard events similarly seems more difficult. If you don't need the event exactly at the same time the highlight is removed from the button, you could override NSButton's -performKeyEquivalent: so that it will e.g. send the action twice.
- (BOOL) performKeyEquivalent: (NSEvent *) anEvent
{
if ([super performKeyEquivalent: anEvent])
{
[self sendAction: self.action to: self.target];
return YES;
}
return NO;
}
If you do need the event at the same time, I think you need to use a custom button cell (by creating a subclass of NSButtonCell and setting the button's cell in the initializer) and override its -highlight:withFrame:inView::
- (void)highlight:(BOOL)flag
withFrame:(NSRect)cellFrame
inView:(NSView *)controlView
{
[super highlight: flag withFrame:cellFrame inView:controlView];
if (flag)
{
// Action hasn't been sent yet.
}
else
{
// Action has been sent.
}
}
I have a list of files. Next to it I have a QLPreviewView which shows the currently selected file.
Unfortunately QLPreviewView loads a web view to preview bookmark files. Some web pages can grab keyboard focus. E.g. the Gmail login form places the insertion point into the user name field.
This breaks the flow of my application. I want to navigate my list using arrow keys. This is disrupted when keyboard focus is taken away from the table view.
So far the best I could come up with is to override - [NSWindow makeFirstResponder:] and not call super for instances of classes named with a QL prefix. Yuck.
Is there a more reasonable way to
Prevent unwanted changes of first responder?
or prevent user interaction on QLPreviewView and its subviews?
I ended up using a NSWindow subclass that allows QLPreviewViews and its private subviews to become first responder on user interaction, but prevents these views from simply stealing focus.
- (BOOL)makeFirstResponder:(NSResponder *)aResponder
{
NSString *classname = NSStringFromClass([aResponder class]);
// This is a hack to prevent Quick Look from stealing first responder
if ([classname hasPrefix:#"QL"]) {
BOOL shouldMakeFirstRespnder = NO;
NSEvent *currentEvent = [[NSApplication sharedApplication] currentEvent] ;
NSEventType eventType = currentEvent.type;
if ((eventType == NSLeftMouseDown) || (eventType == NSRightMouseDown) || (eventType == NSMouseEntered)) {
if ([aResponder isKindOfClass:[NSView class]]) {
NSView *view = (NSView *)aResponder;
NSPoint locationInWindow = currentEvent.locationInWindow;
NSPoint locationInView = [view convertPoint:locationInWindow fromView:nil];
BOOL pointInRect = NSPointInRect(locationInView, [view bounds]);
shouldMakeFirstRespnder = pointInRect;
}
}
if (!shouldMakeFirstRespnder) {
return NO;
}
}
return [super makeFirstResponder:aResponder];
}
Maybe you can subclass QLPreviewView and override its becomeFirstResponder so that you can either enable or disable it when your application should allow it to accept focus.
Header
#interface MyQLPreviewView : QLPreviewView
#end
Implementation
#implementation
- (BOOL)becomeFirstResponder
{
return NO;
}
#end
I have been working on this for hours, have no idea what went wrong. I want a custom cursor for a button which is a subview of NSTextView, I add a tracking area and send the cursorUpdate message when mouse entered button.
The cursorUpdate method is indeed called every time the mouse entered the tracking area. But the cursor stays the IBeamCursor.
Any ideas?
Reference of the Apple Docs: managing cursor-update event
- (void)cursorUpdate:(NSEvent *)event {
[[NSCursor arrowCursor] set];
}
- (void)myAddTrackingArea {
[self myRemoveTrackingArea];
NSTrackingAreaOptions trackingOptions = NSTrackingCursorUpdate | NSTrackingMouseEnteredAndExited | NSTrackingActiveInKeyWindow;
_trackingArea = [[NSTrackingArea alloc] initWithRect: [self bounds] options: trackingOptions owner: self userInfo: nil];
[self addTrackingArea: _trackingArea];
}
- (void)myRemoveTrackingArea {
if (_trackingArea)
{
[self removeTrackingArea: _trackingArea];
_trackingArea = nil;
}
}
I ran into the same problem.
The issue is, that NSTextView updates its cursor every time it receives a mouseMoved: event. The event is triggered by a self updating NSTrackingArea of the NSTextView, which always tracks the visible part of the NSTextView inside the NSScrollView. So there are maybe 2 solutions I can think of.
Override updateTrackingAreas remove the tracking area that is provided by Cocoa and make sure you always create a new one instead that excludes the button. (I would not do this!)
Override mouseMoved: and make sure it doesn't call super when the cursor is over the button.
- (void)mouseMoved:(NSEvent *)theEvent {
NSPoint windowPt = [theEvent locationInWindow];
NSPoint superViewPt = [[self superview]
convertPoint: windowPt fromView: nil];
if ([self hitTest: superViewPt] == self) {
[super mouseMoved:theEvent];
}
}
I had the same issue but using a simple NSView subclass that was a child of the window's contentView and did not reside within an NScrollView.
The documentation for the cursorUpdate flag of NSTrackingArea makes it sound like you only need to handle the mouse entering the tracking area rect. However, I had to manually check the mouse location as the cursorUpdate(event:) method is called both when the mouse enters the tracking area's rect and when it leaves the tracking rect. So if the cursorUpdate(event:) implementation only sets the cursor without checking whether it lies within the tracking area rect, it is set both when it enters and leaves the rect.
The documentation for cursorUpdate(event:) states:
Override this method to set the cursor image. The default
implementation uses cursor rectangles, if cursor rectangles are
currently valid. If they are not, it calls super to send the message
up the responder chain.
If the responder implements this method, but decides not to handle a
particular event, it should invoke the superclass implementation of
this method.
override func cursorUpdate(with event: NSEvent) {
// Convert mouse location to the view coordinates
let mouseLocation = convert(event.locationInWindow, from: nil)
// Check if the mouse location lies within the rect being tracked
if trackingRect.contains(mouseLocation) {
// Set the custom cursor
NSCursor.openHand.set()
} else {
// Reset the cursor
super.cursorUpdate(with: event)
}
}
I just ran across this through a Google search, so I thought I'd post my solution.
Subclass the NSTextView/NSTextField.
Follow the steps in the docs to create an NSTrackingArea. Should look something like the following. Put this code in the subclass's init method (also add the updateTrackingAreas method):
NSTrackingArea *trackingArea = [[NSTrackingArea alloc] initWithRect:self.bounds options:(NSTrackingMouseMoved | NSTrackingActiveInKeyWindow) owner:self userInfo:nil];
[self addTrackingArea:trackingArea];
self.trackingArea = trackingArea;
Now you need to add the mouseMoved: method to the subclass:
- (void)mouseMoved:(NSEvent *)theEvent {
NSPoint point = [self convertPoint:theEvent.locationInWindow fromView:nil];
if (NSPointInRect(point, self.popUpButton.frame)) {
[[NSCursor arrowCursor] set];
} else {
[[NSCursor IBeamCursor] set];
}
}
Note: the self.popUpButton is the button that is a subview of the NSTextView/NSTextField.
That's it! Not too hard it ends up--just had to used mouseMoved: instead of cursorUpdate:. Took me a few hours to figure this out, hopefully someone can use it.
I'm trying to get my program to recognize a double click with an NSCollectionView. I've tried following this guide: http://www.springenwerk.com/2009/12/double-click-and-nscollectionview.html but when I do it, nothing happens because the delegate in IconViewBox is null:
The h file:
#interface IconViewBox : NSBox
{
IBOutlet id delegate;
}
#end
The m file:
#implementation IconViewBox
-(void)mouseDown:(NSEvent *)theEvent {
[super mouseDown:theEvent];
// check for click count above one, which we assume means it's a double click
if([theEvent clickCount] > 1) {
NSLog(#"double click!");
if(delegate && [delegate respondsToSelector:#selector(doubleClick:)]) {
NSLog(#"Runs through here");
[delegate performSelector:#selector(doubleClick:) withObject:self];
}
}
}
The second NSLog never gets printed because delegate is null. I've connected everything in my nib files and followed the instructions. Does anyone know why or an alternate why to do this?
You can capture multiple-clicks within your collection view item by subclassing the collection item's view.
Subclass NSView and add a mouseDown: method to detect multiple-clicks
Change the NSCollectionItem's view in the nib from NSView to MyCollectionView
Implement collectionItemViewDoubleClick: in the associated NSWindowController
This works by having the NSView subclass detect the double-click and it pass up the responder chain. The first object in the responder chain to implement collectionItemViewDoubleClick: is called.
Typically, you should implement collectionItemViewDoubleClick: in the associated NSWindowController, but it can be in any object within the responder chain.
#interface MyCollectionView : NSView
/** Capture double-clicks and pass up responder chain */
-(void)mouseDown:(NSEvent *)theEvent;
#end
#implementation MyCollectionView
-(void)mouseDown:(NSEvent *)theEvent
{
[super mouseDown:theEvent];
if (theEvent.clickCount > 1)
{
[NSApplication.sharedApplication sendAction:#selector(collectionItemViewDoubleClick:) to:nil from:self];
}
}
#end
Another option is to override the NSCollectionViewItem and add an NSClickGestureRecognizer like such:
- (void)viewDidLoad
{
NSClickGestureRecognizer *doubleClickGesture =
[[NSClickGestureRecognizer alloc] initWithTarget:self
action:#selector(onDoubleClick:)];
[doubleClickGesture setNumberOfClicksRequired:2];
// this should be the default, but without setting it, single clicks were delayed until the double click timed-out
[doubleClickGesture setDelaysPrimaryMouseButtonEvents:FALSE];
[self.view addGestureRecognizer:doubleClickGesture];
}
- (void)onDoubleClick:(NSGestureRecognizer *)sender
{
// by sending the action to nil, it is passed through the first responder chain
// to the first object that implements collectionItemViewDoubleClick:
[NSApp sendAction:#selector(collectionItemViewDoubleClick:) to:nil from:self];
}
What you said notwithstanding, you need to be sure you followed step four in the tutorial:
4. Open IconViewPrototype.xib in IB and connect the View's delegate outlet with "File's Owner":
That should do ya, provided you did follow the rest of the steps.
My app has an NSOutlineView and an NSTableView, and I'm having the same problem with both. With a row in either selected, pressing the tab key puts the first column into edit mode instead of making the next key view first responder. To get to the next key view, you need to tab through all of the columns.
Also, shift-tabbing into either view results in the last column going into edit mode, necessitating more shift-tabs to get into its previous key view.
In case it matters, I'm using the autocalculated key view loop, not my own, with my NSWindow set to autorecalculatesKeyViewLoop = YES. I would like tabbing between the columns once the user elects to edit a column, but I don't think it's standard behavior for the tab key to trigger edit mode.
Update
Thanks to the helpful responses below, I worked it out. Basically, I override -keyDown in my custom table view class, which handles tabbing and shift-tabbing out of the table view. It was tougher to solve shift-tabbing into the table view, however. I set a boolean property to YES in the custom table view's -acceptsFirstResponder if it's accepting control from another view.
The delegate's -tableView:shouldEditTableColumn:row checks for that when the current event is a shift-tab keyDown event. -tableView:shouldEditTableColumn:row is called and it's not a shift-tab event, it sets the table view's property back to NO so it can still be edited as usual.
I've pasted the full solution below.
/* CustomTableView.h */
#interface CustomTableView : NSTableView {}
#property (assign) BOOL justFocused;
#end
/* CustomTableView.m */
#implementation CustomTableView
#synthesize justFocused;
- (BOOL)acceptsFirstResponder {
if ([[self window] firstResponder] != self) {
justFocused = YES;
}
return YES;
}
- (void)keyDown:(NSEvent *)theEvent
{
// Handle the Tab key
if ([[theEvent characters] characterAtIndex:0] == NSTabCharacter) {
if (([theEvent modifierFlags] & NSShiftKeyMask) != NSShiftKeyMask) {
[[self window] selectKeyViewFollowingView:self];
} else {
[[self window] selectKeyViewPrecedingView:self];
}
}
else {
[super keyDown:theEvent];
}
}
#end
/* TableViewDelegate.m */
. . .
- (BOOL)tableView:(NSTableView *)tableView shouldEditTableColumn:(NSTableColumn *)tableColumn
row:(NSInteger)row
{
NSEvent *event = [NSApp currentEvent];
BOOL shiftTabbedIn = ([event type] == NSKeyDown
&& [[event characters] characterAtIndex:0] == NSBackTabCharacter);
if (shiftTabbedIn && ((CustomTableView *)tableView).justFocused == YES) {
return NO;
} else {
((CustomTableView *)tableView).justFocused = NO;
}
return YES;
}
. . .
This is the default behavior. If there's no row selected, the table view as a whole has focus, and the Tab key switches to the next key view. If there is a row selected, the table view begins editing or moves to the next cell if already editing.
From AppKit Release Notes:
Tables now support inter-cell
navigation as follows:
Tabbing forward to a table focuses the entire table.
Hitting Space will attempt to 'performClick:' on a NSButtonCell in
the selected row, if there is only one
instance in that row.
Tabbing again focuses the first "focusable" (1) cell, if there is one.
If the newly focused cell can be edited, editing will begin.
Hitting Space calls 'performClick:' on the cell and sets the datasource
value afterwards, if changed. (2)
If a text cell is editing, hitting Enter will commit editing and focus
will be returned to the tableview, and
Tab/Shift-tab will commit the editing
and then perform the new tab-loop
behavior.
Tabbing will only tab through a single row
Once the last cell in a row is reached, tab will take the focus to
the next focusable control.
Back tabbing into a table will select the last focusable cell.
If you want to change this behavior, the delegate method tableView:shouldEditTableColumn:row: may be helpful. You may also have to subclass NSTableView if you really want to affect only the behavior of the Tab key.
The solution using keyDown didn't work for me. Perhaps because it is for cell-based table view.
My solution for a view-based table view, in Swift, looks like this:
extension MyTableView: NSTextFieldDelegate {
func controlTextDidEndEditing(_ obj: Notification) {
guard
let view = obj.object as? NSView,
let textMovementInt = obj.userInfo?["NSTextMovement"] as? Int,
let textMovement = NSTextMovement(rawValue: textMovementInt) else { return }
let columnIndex = column(for: view)
let rowIndex = row(for: view)
let newRowIndex: Int
switch textMovement {
case .tab:
newRowIndex = rowIndex + 1
if newRowIndex >= numberOfRows { return }
case .backtab:
newRowIndex = rowIndex - 1
if newRowIndex < 0 { return }
default: return
}
DispatchQueue.main.async {
self.editColumn(columnIndex, row: newRowIndex, with: nil, select: true)
}
}
}
You also need to set the cell.textField.delegate so that the implementation works.
My blog post on this tricky workaround: https://samwize.com/2018/11/13/how-to-tab-to-next-row-in-nstableview-view-based-solution/
I've had to deal with this before as well. My solution was to subclass NSTableView or NSOutlineView and override keyDown: to catch the tab key presses there, then act on them.
How convenient! I was just looking at this myself yesterday, and it's good to see some confirmation of the approach I took - keyDown: handling.
However, I have one small possible refinement to your approach: I worked out that the method triggering editing on shift-tabbing back to the table was the becomeFirstResponder call. So what I did on a NSTableView subclass was:
Add a synthesized property to control whether tab-editing behaviour was disabled
On keydown, check the first character (also check for [[theEvent characters] length] to avoid exceptions for dead keys!) for tab; if tab editing is disabled, move on to the next/previous view, as per your code sample.
Override becomeFirstResponder:
- (BOOL)becomeFirstResponder {
if (tabEditingDisabled) {
[self display];
return YES;
}
return [super becomeFirstResponder];
}
This keeps all the code in the tableview subclass, keeping the delegate cleaner :)
The only danger is I don't know what else NSTableView does in becomeFirstResponder; I didn't notice anything breaking, but...
This worked for me:
- (BOOL)tableView:(NSTableView *)tableView shouldEditTableColumn:(NSTableColumn *)tableColumn row:(NSInteger)row {
NSEvent *e = [NSApp currentEvent];
if (e.type == NSKeyDown && e.keyCode == 48) return NO;
return YES;
}