Hi All,
Here's a second version of my patch, updated for various comments.
It seems to pass the regression tests, and E still starts up.
This version has lots more #ifdefs (sorry Vincent), to address Gustavo's
comment that the #ifdefs should be inside functions, not outside them.
I think this m
Gustavo Sverzut Barbieri wrote:
>> +void _ecore_main_loop_init(void)
>> +{
>> + epoll_fd = epoll_create(1);
>> + if (epoll_fd < 0)
>> + ERR("Failed to create epoll fd!");
>
> make it a CRIT(), so it will abort when eina is set to abort on
> critical. This will make the application not wo
On Mon, May 17, 2010 at 2:16 AM, Mike McCormack
wrote:
> Hey Guys,
>
> Here's a patch to support epoll in ecore.
>
> Probably needs a bit more testing, but let me know what you think.
Hi Mike, code looks great. I don't agree with Vincent about having yet
another file for it, but I'll suggest some
Hi folks!
I know raster is working on rewriting lua, but still the timer API had
a minor usability issue that I would like to see solved.
Upon creating a timer, it's callback must return an int, otherwise it
would panic. The python bindings do the nicety of, if the callback
function returns nothin
On Mon, 2010-05-17 at 18:18 +0900, Mike McCormack wrote:
> Dejun.Liu wrote:
> >
> > On Mon, 2010-05-17 at 14:16 +0900, Mike McCormack wrote:
> >> Hey Guys,
> >>
> >> Here's a patch to support epoll in ecore.
> >>
> >> Probably needs a bit more testing, but let me know what you think.
> >>
> >
> >
On Mon, 17 May 2010, Mike McCormack wrote:
> Vincent Torri wrote:
>
>> * add a file, named ecore_epoll.c instead of having lots of #ifdef
>
> I added 2 ifdefs vs. 11 for _WIN32. I could get it down to one if you would
> insist.
it's cleaner imho. The _WIN32 are necessary, unfortunately.
>
Dejun.Liu wrote:
>
> On Mon, 2010-05-17 at 14:16 +0900, Mike McCormack wrote:
>> Hey Guys,
>>
>> Here's a patch to support epoll in ecore.
>>
>> Probably needs a bit more testing, but let me know what you think.
>>
>
> why trigger method not be considered? sample code only uses level
> trigger.
Vincent Torri wrote:
> * add a file, named ecore_epoll.c instead of having lots of #ifdef
I added 2 ifdefs vs. 11 for _WIN32. I could get it down to one if you would
insist. Adding a separate file would mean no inlining.
> * don't add an _init() in ecore_main.c, but in the above file
epoll
On Fri, 14 May 2010, Mike McCormack wrote:
>
> "cd ecore && ./autogen.sh --enable-tests &&make && make check-local" was
> failing.
why doing make check-local and not just make check ?
Vincent
> Make sure to build ecore_suite before trying to run it.
>
> Makefile.am |2 ++
> 1 file changed
On Mon, 2010-05-17 at 14:16 +0900, Mike McCormack wrote:
> Hey Guys,
>
> Here's a patch to support epoll in ecore.
>
> Probably needs a bit more testing, but let me know what you think.
>
why trigger method not be considered? sample code only uses level
trigger.
Steven
> plain text documen
On Mon, 17 May 2010, Mike McCormack wrote:
> Hey Guys,
>
> Here's a patch to support epoll in ecore.
>
> Probably needs a bit more testing, but let me know what you think.
* add a file, named ecore_epoll.c instead of having lots of #ifdef
* don't add an _init() in ecore_main.c, but in the a
11 matches
Mail list logo
