Implement drag and drop of downloads for the Mac downloads page

Implement drag and drop of downloads for the Mac downloads page.

BUG=15776
(http://crbug.com/15776)
TEST=Download an item, open the download page, drag the icon of
     the download to the desktop.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=23548

[pink (ping after 24hrs), 2009-08-14 14:43:50]

LGTM with some nits, would be even better with a unit test, per below.

http://codereview.chromium.org/164459/diff/53/55
File chrome/browser/cocoa/download_util_mac.mm (right):

http://codereview.chromium.org/164459/diff/53/55#newcode18
Line 18: void DragDownload(const DownloadItem* download,
is there any way to unit test this? Can you break it up into a couple of
routines internally and at least unit test setting the data on the pasteboard?

http://codereview.chromium.org/164459/diff/53/55#newcode24
Line 24: NSArray* fileList = [NSArray arrayWithObjects:file, nil];
you can just use arrayWithObject: here.

http://codereview.chromium.org/164459/diff/53/55#newcode40
Line 40: double eventTime = (double)(AbsoluteToDuration(UpTime())) / 1000.0;
isn't there a  cocoa way to get what the current time is?

http://codereview.chromium.org/164459/diff/53/55#newcode41
Line 41: NSEvent* dragEvent = [NSEvent mouseEventWithType:NSLeftMouseDragged
in the web_drag_source.mm, we don't bother synthesizing an event. we should
probably be consistent. We may want to do it in both places as there are some
hacks in w_d_s.mm for when sometimes the current event isn't a drag event.

[Paul Godavari, 2009-08-14 17:20:26]

Updated as per your comments.

What do you think the best approach is for getting the NSEvent needed for
dragging? Synthesize or [NSApp currentEvent]?

http://codereview.chromium.org/164459/diff/53/55
File chrome/browser/cocoa/download_util_mac.mm (right):

http://codereview.chromium.org/164459/diff/53/55#newcode18
Line 18: void DragDownload(const DownloadItem* download,
On 2009/08/14 14:43:51, pink wrote:
> is there any way to unit test this? Can you break it up into a couple of
> routines internally and at least unit test setting the data on the pasteboard?

Added a unit test to cover the pasteboard operation.

http://codereview.chromium.org/164459/diff/53/55#newcode24
Line 24: NSArray* fileList = [NSArray arrayWithObjects:file, nil];
On 2009/08/14 14:43:51, pink wrote:
> you can just use arrayWithObject: here.

Done.

http://codereview.chromium.org/164459/diff/53/55#newcode40
Line 40: double eventTime = (double)(AbsoluteToDuration(UpTime())) / 1000.0;
On 2009/08/14 14:43:51, pink wrote:
> isn't there a  cocoa way to get what the current time is?

No longer needed when using [NSApp currentEvent].

http://codereview.chromium.org/164459/diff/53/55#newcode41
Line 41: NSEvent* dragEvent = [NSEvent mouseEventWithType:NSLeftMouseDragged
On 2009/08/14 14:43:51, pink wrote:
> in the web_drag_source.mm, we don't bother synthesizing an event. we should
> probably be consistent. We may want to do it in both places as there are some
> hacks in w_d_s.mm for when sometimes the current event isn't a drag event.

Getting the event via [NSApp currentEvent] works well here, and is less code. If
the current event isn't a drag, how much of a problem would it be?

I can change web_drag_source to synthesize events as well, but it feels a little
hacky (I'm not sure what other options we have).

[pink (ping after 24hrs), 2009-08-17 15:48:53]

LGTM with one fix to unit test.

http://codereview.chromium.org/164459/diff/1071/82
File chrome/browser/cocoa/download_util_mac_unittest.mm (right):

http://codereview.chromium.org/164459/diff/1071/82#newcode24
Line 24: NSPasteboard* pasteboard = [NSPasteboard
pasteboardWithName:NSDragPboard];
i wouldn't use this, there's a method to create a guaranteed unique pasteboard.
This might get disturbed by someone using their machine while the tests were
running.