From: Sean McBride <sean@ro...> - 2010-03-24 17:39:35
On Wed, 24 Mar 2010 10:42:33 -0600, Nathan Hjelm said:
>Looks good. A couple of things:
Thanks for looking!
>1) why are you storing the current runloop in event_thread_main? Run
>loops are associated with threads so the (current) two calls to
>CFRunLoopGetCurrent return the same value. There is no need to keep
>track of the current run loop.
The change solves a warning about unbalanced retain/release. The clang
static analyzer is not smart enough to realise that CFRunLoopGetCurrent
() returns the same thing both times.
>2) Apple uses a void * to pass the transfer size so it is probably safer
>to use 64 bits for io_size (certainly doesn't hurt).
That change was from Mike Colagrosso actually, see:
I think perhaps using uintptr_t would be better still.
>I have attached a
>patch making this change. Regenerate your patches versus the current
>tree patched with my change.
I'll try. :) I'm new to git...
Sean McBride, B. Eng sean@...
Rogue Research http://www.rogue-research.com
Mac Software Developer Montréal, Québec, Canada
From: Mike Colagrosso <mikec@la...> - 2010-03-24 18:12:44
>>2) Apple uses a void * to pass the transfer size so it is probably safer
>>to use 64 bits for io_size (certainly doesn't hurt). I have attached a
>>patch making this change.
> That change was from Mike Colagrosso actually, see:
Nathan's patch is an improvement. My 2/19/2010 patch made
darwin_async_io_callback() consistent with darwin_handle_callback()
(using a 32-bit io_size). Nathan's patch changes both functions to use
a 64-bit io_size.
Get latest updates about Open Source Projects, Conferences and News.