Callback API Change: Allow Bind() to take a Callback<> and bind all its free parameters.

Allow Bind() to take a Callback<> and bind all its free parameters.

Basically, turns base::Callback<void(...)> to base::Closure.

It turns out there are a number of use caess where an API takes a callback, and then wants to invoke the Callback on another thread. This piece of syntactic sugar removes the need for custom helper functions.

This isn't quite full currying, however it is much simpler to implement.  

BUG=87287
TEST=none


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=103446

[awong, 2011-09-29 00:58:52]

Changing subject to mark this is an API extension.

As for concrete use cases, media wants this.  My fix for cancelable_request also
wants this.

[darin (slow to review), 2011-09-29 05:48:33]

To make sure I understand, this does not allow you to translate a
Callback<void(a,b)> into a Callback<void(b)> simply by binding a value for the
"a" slot, right?  It looks like you can only use this to bind all of the free
arguments to produce a Closure.

[awong, 2011-09-29 07:05:43]

Exactly.  You can only bind to a closure.  Doing the full implementation of
currying would require a parallel path through large chunks of the bind
internal templates which doesn't seem worth it.
On Sep 28, 2011 10:48 PM, <darin@chromium.org> wrote:
> To make sure I understand, this does not allow you to translate a
> Callback<void(a,b)> into a Callback<void(b)> simply by binding a value for

> the
> "a" slot, right? It looks like you can only use this to bind all of the
> free
> arguments to produce a Closure.
>
> http://codereview.chromium.org/8073012/

[awong, 2011-09-29 17:55:33]

Are there concerns about this usage?  I know it's last day of perf, but if
people don't have any hesitations, I'd love to get this checked in since the big
quota CL I'm reviewing could be simplified with this.

If there are reservations though, happy to wait a few days.

Thanks,
Albert

On 2011/09/29 07:05:43, awong wrote:
> Exactly.  You can only bind to a closure.  Doing the full implementation of
> currying would require a parallel path through large chunks of the bind
> internal templates which doesn't seem worth it.
> On Sep 28, 2011 10:48 PM, <mailto:darin@chromium.org> wrote:
> > To make sure I understand, this does not allow you to translate a
> > Callback<void(a,b)> into a Callback<void(b)> simply by binding a value for
> 
> > the
> > "a" slot, right? It looks like you can only use this to bind all of the
> > free
> > arguments to produce a Closure.
> >
> > http://codereview.chromium.org/8073012/

[darin (slow to review), 2011-09-30 06:39:00]

LGTM

Yeah, perf got the best of me today.  This looks great!

On Thu, Sep 29, 2011 at 10:55 AM, <ajwong@chromium.org> wrote:

> Are there concerns about this usage?  I know it's last day of perf, but if
> people don't have any hesitations, I'd love to get this checked in since
> the big
> quota CL I'm reviewing could be simplified with this.
>
> If there are reservations though, happy to wait a few days.
>
> Thanks,
> Albert
>
>
> On 2011/09/29 07:05:43, awong wrote:
>
>> Exactly.  You can only bind to a closure.  Doing the full implementation
>> of
>> currying would require a parallel path through large chunks of the bind
>> internal templates which doesn't seem worth it.
>> On Sep 28, 2011 10:48 PM, <mailto:darin@chromium.org> wrote:
>> > To make sure I understand, this does not allow you to translate a
>> > Callback<void(a,b)> into a Callback<void(b)> simply by binding a value
>> for
>>
>
>  > the
>> > "a" slot, right? It looks like you can only use this to bind all of the
>> > free
>> > arguments to produce a Closure.
>> >
>> >
http://codereview.chromium.**org/8073012/<http://codereview.chromium.org/8073...
>>
>
>
>
>
http://codereview.chromium.**org/8073012/<http://codereview.chromium.org/8073...
>

[commit-bot: I haz the power, 2011-09-30 07:34:28]

CQ is trying tha patch. Follow status at
https://chromium-status.appspot.com/cq/ajwong@chromium.org/8073012/1

[awong, 2011-09-30 07:35:08]

Yeah...I'm still doing some of my perf :(

Thanks for the review!  committing.

On 2011/09/30 06:39:00, darin wrote:
> LGTM
> 
> Yeah, perf got the best of me today.  This looks great!
> 
> On Thu, Sep 29, 2011 at 10:55 AM, <mailto:ajwong@chromium.org> wrote:
> 
> > Are there concerns about this usage?  I know it's last day of perf, but if
> > people don't have any hesitations, I'd love to get this checked in since
> > the big
> > quota CL I'm reviewing could be simplified with this.
> >
> > If there are reservations though, happy to wait a few days.
> >
> > Thanks,
> > Albert
> >
> >
> > On 2011/09/29 07:05:43, awong wrote:
> >
> >> Exactly.  You can only bind to a closure.  Doing the full implementation
> >> of
> >> currying would require a parallel path through large chunks of the bind
> >> internal templates which doesn't seem worth it.
> >> On Sep 28, 2011 10:48 PM, <mailto:darin@chromium.org> wrote:
> >> > To make sure I understand, this does not allow you to translate a
> >> > Callback<void(a,b)> into a Callback<void(b)> simply by binding a value
> >> for
> >>
> >
> >  > the
> >> > "a" slot, right? It looks like you can only use this to bind all of the
> >> > free
> >> > arguments to produce a Closure.
> >> >
> >> >
>
http://codereview.chromium.**org/8073012/%3Chttp://codereview.chromium.org/80...>
> >>
> >
> >
> >
> >
>
http://codereview.chromium.**org/8073012/%3Chttp://codereview.chromium.org/80...>
> >

[commit-bot: I haz the power, 2011-09-30 09:09:36]

Change committed as 103446