357560 - Switch to use native NSSearchField widget in Privacy sheets

Status: RESOLVED FIXED

(Camino Graveyard :: Preferences, defect)

Description

[Stuart Morgan]

This is the baby cousin of bug 330666; it's really easy to do here, and not related to the other uses.

Comment 1

[Stuart Morgan]

Attachment: use NSSearchField

Comment 2

[Stuart Morgan]

Attachment: new nib

Just swaps out the fields.  Tab order should still all be fine.

Comment 3

[Stuart Morgan]

Attachment: with sdk workaround

I forget to make it build against the 10.2 SDK.

Comment 4

[froodian (Ian Leue)]

With this patch/nib combo, hitting return just loses the focus around the search field - it doesn't actually "push" the default button.

Comment 5

[froodian (Ian Leue)]

Comment on attachment 243122 [details] [diff] [review]
with sdk workaround

I talked to smokey about this in channel, and apparently what we said is that this is worth living with, since it's uncommon that users open the sheet, search, and try to close without acting on the actions.  On that premise, r=me.  A followup bug seems like it'd be a good idea though.

Comment 6

[froodian (Ian Leue)]

Comment on attachment 243122 [details] [diff] [review]
with sdk workaround

I talked to smokey about this in channel, and apparently what we said is that this is worth living with, since it's uncommon that users open the sheet, search, and try to close without acting on the actions.  On that premise, r=me.  A followup bug seems like it'd be a good idea though.

Comment 7

[Mike Pinkerton (not reading bugmail)]

i think we should fix the return issue before we land this. i think it'll be more annoying that you think, especially since that field is focused by default.

Comment 8

[Stuart Morgan]

(In reply to comment #7)
> i think we should fix the return issue before we land this.

Unfortunately, there doesn't seem to be a way to do that without subclassing NSSearchField, which I believe means this is blocked on moving to the 10.3 SDK.

Comment 9

[froodian (Ian Leue)]

Comment on attachment 243122 [details] [diff] [review]
with sdk workaround

Removing sr flag.  We're on the 10.3 SDK now, so we can do this properly.

Comment 10

[Stuart Morgan]

Attachment: v3

New version. After all that, it turned out that I didn't need to subclass after all.

Comment 11

[Stuart Morgan]

Attachment: new nib

The change is easy for me to re-make to other nibs, so it's not a big deal that this will rot the other Privacy nib.

Someone should check that IB 3 was behaving when it saved this, and that it's still editable with older tools.

Comment 12

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

FWIW, it takes me two tries to open the nib, and I get this sort of console spew:

2007-06-26 01:23:08.137 Interface Builder[18326] **** verification error ****: <NSCustomObject: 0x58419a0> has an invalid object id -3 (must be < 759)

2007-06-26 01:23:08.469 Interface Builder[18326] *** Assertion failure in -[NSMenu itemAtIndex:], Menus.subproj/NSMenu.m:638

2007-06-26 01:23:08.470 Interface Builder[18326] Invalid parameter not satisfying: (index >= 0) && (index < (_itemArray ? CFArrayGetCount(_itemArray) : 0))

I can open and edit it, but the problem persists on subsequent tries.  Dunno if it's the IB 3 or IB on 10.3.9 showing its age :p

Comment 13

[Sean Murphy]

Comment on attachment 269801 [details] [diff] [review]
v3

Looks A-OK here.  Nice technique allowing the default button to trigger without hard-coding specifically which button.

I just want to point out that there is now a short delay before the search field's filtering action is invoked.  This is the way NSSearchFields work by default (-[NSSearchFieldCell setSendsSearchStringImmediately:] alters the invocation, but is 10.4+ only).  Definitely not a problem and I only chose to mention it for the sake of detail, and the fact that it does differ from previous behavior in this pane.

Also, I experience the same trouble Smokey did in comment 22 when opening up the nib, using IB 2.5.4 on 10.4.  IB3 must have done something bad.

Finally, just keep an eye on pointer placement (Object* vs. Object *).  There are some minor inconstancies within the patch (and the entire file seems divided, as usual!).

Comment 14

[Stuart Morgan]

(I'll fix the nib and the style inconsistencies on checkin if there's nothing else needing a change from sr)

Comment 15

[Mike Pinkerton (not reading bugmail)]

Comment on attachment 269801 [details] [diff] [review]
v3

sr=pink

Comment 16

[Stuart Morgan]

Landed on trunk and MOZILLA_1_8_BRANCH.