Add RuntimeEnabled info to InterfaceInfo.pickle.

Add RuntimeEnabled info to InterfaceInfo.pickle.

Adding RuntimeEnabled info so that changes to RuntimeFeature information
properly triggers regeneration of bindings code that depends on them.
This was discovered by removing the VideoTrack feature flag from
TrackEvent.idl and Event.cpp was not getting properly regenerated.

[acolwell GONE FROM CHROMIUM, 2014-03-21 22:14:12]

This patch was created to fix build errors uncovered by the removal of a
RuntimeEnabled flag in https://codereview.chromium.org/205583009/ . All the try
jobs were failing because Events.cpp was not getting regenerated.

[haraken, 2014-03-22 05:27:01]

LGTM

[Nils Barth (inactive), 2014-03-24 02:05:52]

This is ok as a quick fix (if you could add a comment to explain it),
though it's fragile due to the dependency being implicit.
(generate_event_interfaces only implicitly depends on the actual event IDL
files)
This can be made more robust by making this dependency explicit;
I'll post a separate CL and see if that's a better solution.

[Nils Barth (inactive), 2014-03-24 04:05:55]

On 2014/03/24 02:05:52, Nils Barth wrote:
> This is ok as a quick fix (if you could add a comment to explain it),
> though it's fragile due to the dependency being implicit.
> (generate_event_interfaces only implicitly depends on the actual event IDL
> files)
> This can be made more robust by making this dependency explicit;
> I'll post a separate CL and see if that's a better solution.

Actually, not LGTM, after thinking about it.

I've posted a correct fix at:
Fix dependencies for EventInterfaces.in
https://codereview.chromium.org/209723002/
...and it's currently landing, so should be fixed soon!

(This CL adds cruft, is fragile, and doesn't solve the underlying problem,
which is that we need to depend on the actual IDL files, which is a GYP change.
I actually have a FIXME to fix this; it's a 1-line GYP fix, which I've done.)

Thanks for bringing this up, and of course it's fine to have quick fixes
to fix build, but this was easy enough to fix properly!

[acolwell GONE FROM CHROMIUM, 2014-03-25 16:35:31]

On 2014/03/24 04:05:55, Nils Barth wrote:
> On 2014/03/24 02:05:52, Nils Barth wrote:
> > This is ok as a quick fix (if you could add a comment to explain it),
> > though it's fragile due to the dependency being implicit.
> > (generate_event_interfaces only implicitly depends on the actual event IDL
> > files)
> > This can be made more robust by making this dependency explicit;
> > I'll post a separate CL and see if that's a better solution.
> 
> Actually, not LGTM, after thinking about it.
> 
> I've posted a correct fix at:
> Fix dependencies for EventInterfaces.in
> https://codereview.chromium.org/209723002/
> ...and it's currently landing, so should be fixed soon!
> 
> (This CL adds cruft, is fragile, and doesn't solve the underlying problem,
> which is that we need to depend on the actual IDL files, which is a GYP
change.
> I actually have a FIXME to fix this; it's a 1-line GYP fix, which I've done.)
> 
> Thanks for bringing this up, and of course it's fine to have quick fixes
> to fix build, but this was easy enough to fix properly!

Thanks for providing a better fix. I didn't really know what the right fix would
be so I just tried adding the info to the InterfaceInfo.pickle since that seemed
to be what all the other code appeared to be getting its info from. I'm not very
familiar w/ the build system, but I figured I'd try my hand at an improvement
instead of just saying "it's broken!"  :)

[Nils Barth (inactive), 2014-03-26 04:20:08]

On 2014/03/25 16:35:31, acolwell wrote:
> Thanks for providing a better fix. I didn't really know what the right fix
would
> be so I just tried adding the info to the InterfaceInfo.pickle since that
seemed
> to be what all the other code appeared to be getting its info from. I'm not
very
> familiar w/ the build system, but I figured I'd try my hand at an improvement
> instead of just saying "it's broken!"  :)

No problem, and thanks for the help!
(Ideally of course it should "just work" ;)

I've documented the build in the GYP file and somewhat further at:
http://www.chromium.org/developers/design-documents/idl-build
...but the docs need improvement (notably an outline of the build steps is
missing!).

EventInterfaces.in is actually pretty separate from bindings,
so I'll look into factoring it out to simplify the build and avoid these
problems in future.