106689 - [META] Tracking bug for CarbonEvent transition

Status: RESOLVED INVALID

(Core Graveyard :: Widget: Mac, defect)

Description

[Mike Pinkerton (not reading bugmail)]

meta-bug to track the work for CarbonEvents.

Comment 1

[cyrus]

I'm sure this has been covered, but can anybody summarize the cost/
benefits of this?

Comment 2

[Mike Pinkerton (not reading bugmail)]

*** Bug 76901 has been marked as a duplicate of this bug. ***

Comment 3

[strobe anarkhos]

Benefits:

• Faster connections, fewer failed connections (polling the event manager slowly
like Mozilla does often fails to catch the three way handshake).

• Play animated GIFs at the correct speed.

• Faster 'boot-up' time (you can simulate the effect by clicking down on the
splash screen)

• Play Flash movies correctly

• Doesn't pause any tasks when using the menu or title bar.

• Better overall responsiveness

• No more losing up-click when quickly clicking the title bar

• Will no longer use CPU when idle. 


I'm sure there are more, but that's a starter list. 

Comment 4

[Simon Fraser [no longer active]]

> • Faster connections, fewer failed connections (polling the event manager slowly
> like Mozilla does often fails to catch the three way handshake).

Can you explain this more? I'm not aware of any issues with lost connections.

Comment 5

[Mike Pinkerton (not reading bugmail)]

-> 098

Comment 6

[Mike Pinkerton (not reading bugmail)]

not gonna happen in 098

Comment 7

[Mike Pinkerton (not reading bugmail)]

over to dagley, i probably won't have time for these by 099 with the embedding
work i'm getting pulled into.

Comment 8

[Steve Dagley]

-> mozilla1.0

Comment 9

[saari (gone)]

not going to do full transition to carbon events for mozilla 1.0

Comment 10

[Greg K.]

*** Bug 151921 has been marked as a duplicate of this bug. ***

Comment 11

[Ari]

Are we ever going to do this? In my opinion, it would be a waste to do both this
and to rewrite widget in cocoa (already partially done and necessary for
chimera) because cocoa offers all of the benefits of CarbonEvents and more.
After all, doing this right would probably require a rewrite of a large portion
of widget anyway.

Comment 12

[Frankie]

Whether it gets fixed through cocoa-izing the widgets or by writing
CarbonEvents,  either way it would be nice to stop wasting so many CPU cycles. A
shame this set of bugs have been at a standstill so long.

Mozilla (any OSX non-mach build) typically sucks up 4% of my CPU while it's
inactive with no windows open, and much more if I leave an open window with a
few tabs in the background. Chimera exhibits similar behavior, BTW.

Comment 13

[Paul Bergsagel]

Is bug 168611 a duplicate of this bug or a seperate bug?

Comment 14

[Steve Dagley]

Since I don't report into Internet Technologies anymore this bug needs a new
owner -> saari

Comment 15

[Mark Mentovai]

See also bug 332579.

Comment 16

[Mark Mentovai]

We're part of the way there.  EventRecord-type events are now handled as Carbon events outside and don't use WaitNextEvent, but are still converted to EventRecords and processed through the old dispatcher.  Reference bug 332579 (1.8 branch), bug 326273 (trunk), bug 338153 (consolidating 1.8 branch and trunk implementations).

Comment 17

[Frankie]

RIP Carbon