Removing #ifdefs for WEB_AUDIO compile time flag.

Removing #ifdefs for WEB_AUDIO compile time flag.

There is already a runtime flag for Web audio feature
so removing the compile time flag WEB_AUDIO.

BUG=417628

[RashmiVijay, 2014-09-25 11:38:06]

rashmi.vijay@samsung.com changed reviewers:
+ kbr@chromium.org

[RashmiVijay, 2014-09-25 11:38:34]

rashmi.vijay@samsung.com changed reviewers:
+ rtoy@chromium.org
- kbr@chromium.org

[RashmiVijay, 2014-09-25 11:39:18]

rashmi.vijay@samsung.com changed reviewers:
+ kbr@chromium.org

[RashmiVijay, 2014-09-25 13:28:53]

This patch needs clobber build as the bindings files(which are
auto-generated)are not getting regenerated after applying the patch on different
bots.Please help so that all the files which are generated will have my changes.

[Mike West, 2014-09-25 13:40:03]

mkwst@chromium.org changed reviewers:
+ mkwst@chromium.org

[Mike West, 2014-09-25 13:40:04]

We don't compile this feature on some Android platforms due to hardware
limitations. Have you evaluated whether or not that restriction is outdated with
the API's owners?

https://codereview.chromium.org/604753003/diff/1/LayoutTests/fast/events/even...
File LayoutTests/fast/events/event-creation.html (left):

https://codereview.chromium.org/604753003/diff/1/LayoutTests/fast/events/even...
LayoutTests/fast/events/event-creation.html:199: //
shouldBeTrue("document.createEvent('AudioProcessingEvent') instanceof
window.AudioProcessingEvent");
You should be able to uncomment these tests, correct?

https://codereview.chromium.org/604753003/diff/1/Source/build/features.gypi
File Source/build/features.gypi (left):

https://codereview.chromium.org/604753003/diff/1/Source/build/features.gypi#o...
Source/build/features.gypi:71: 'ENABLE_WEB_AUDIO=1',
Given the comments here, it's not at all clear that we can remove this
compile-time flag. kbr@, WDYT?

[dcheng, 2014-09-25 16:28:16]

On 2014/09/25 at 13:28:53, rashmi.vijay wrote:
> This patch needs clobber build as the bindings files(which are
auto-generated)are not getting regenerated after applying the patch on different
bots.Please help so that all the files which are generated will have my changes.

Independent of whether or not this patch lands, we should not land patches that
require manual clobbers ever.

[Raymond Toy, 2014-09-25 17:43:18]

https://codereview.chromium.org/604753003/diff/1/Source/build/features.gypi
File Source/build/features.gypi (left):

https://codereview.chromium.org/604753003/diff/1/Source/build/features.gypi#o...
Source/build/features.gypi:71: 'ENABLE_WEB_AUDIO=1',
On 2014/09/25 13:40:04, Mike West wrote:
> Given the comments here, it's not at all clear that we can remove this
> compile-time flag. kbr@, WDYT?

The comments should probably be updated.

I have personally tested webaudio on Android arm, arm64, x86, and mips. These
all build and run as expected. (I have not tested x86 or mips recently,
however.)

https://codereview.chromium.org/604753003/diff/1/Source/modules/webaudio/Medi...
File Source/modules/webaudio/MediaStreamAudioDestinationNode.h (right):

https://codereview.chromium.org/604753003/diff/1/Source/modules/webaudio/Medi...
Source/modules/webaudio/MediaStreamAudioDestinationNode.h:65: #endif //
MediaStreamAudioDestinationNode_h
Isn't this in the wrong place now?

[Ken Russell (switch to Gerrit), 2014-09-25 18:21:39]

I don't see a good reason to land this patch. It will require a flag day and
reduces flexibility in case the code needs to be compiled out for some reason.

Agree with @dcheng that we should not land CLs that require clobber builds.
Sometimes it's inevitable, and that's why the landmines mechanism is present,
but doing so has high costs for all Chromium developers and should be avoided.

[haraken, 2015-04-24 05:33:23]

On 2014/09/25 18:21:39, Ken Russell wrote:
> I don't see a good reason to land this patch. It will require a flag day and
> reduces flexibility in case the code needs to be compiled out for some reason.
> 
> Agree with @dcheng that we should not land CLs that require clobber builds.
> Sometimes it's inevitable, and that's why the landmines mechanism is present,
> but doing so has high costs for all Chromium developers and should be avoided.

We want to make this change at some point somehow because WEB_AUDIO is the last
remaining compile time flag, and by removing the flag, we can remove lots of
code about [Conditional] from the IDL compiler.

Would there be any more modest way to land this CL?

[Mike West, 2015-04-24 12:34:40]

On 2015/04/24 at 05:33:23, haraken wrote:
> On 2014/09/25 18:21:39, Ken Russell wrote:
> > I don't see a good reason to land this patch. It will require a flag day and
> > reduces flexibility in case the code needs to be compiled out for some
reason.
> > 
> > Agree with @dcheng that we should not land CLs that require clobber builds.
> > Sometimes it's inevitable, and that's why the landmines mechanism is
present,
> > but doing so has high costs for all Chromium developers and should be
avoided.
> 
> We want to make this change at some point somehow because WEB_AUDIO is the
last remaining compile time flag, and by removing the flag, we can remove lots
of code about [Conditional] from the IDL compiler.
> 
> Would there be any more modest way to land this CL?

I'll defer to krb@ as he had stronger opinions about the topic than I, but this
seems to me to be a good enough reason to override the general concern about
forcing clobber builds (via landmines).

[haraken, 2015-05-20 04:57:34]

On 2015/04/24 12:34:40, Mike West (holiday in DE) wrote:
> On 2015/04/24 at 05:33:23, haraken wrote:
> > On 2014/09/25 18:21:39, Ken Russell wrote:
> > > I don't see a good reason to land this patch. It will require a flag day
and
> > > reduces flexibility in case the code needs to be compiled out for some
> reason.
> > > 
> > > Agree with @dcheng that we should not land CLs that require clobber
builds.
> > > Sometimes it's inevitable, and that's why the landmines mechanism is
> present,
> > > but doing so has high costs for all Chromium developers and should be
> avoided.
> > 
> > We want to make this change at some point somehow because WEB_AUDIO is the
> last remaining compile time flag, and by removing the flag, we can remove lots
> of code about [Conditional] from the IDL compiler.
> > 
> > Would there be any more modest way to land this CL?
> 
> I'll defer to krb@ as he had stronger opinions about the topic than I, but
this
> seems to me to be a good enough reason to override the general concern about
> forcing clobber builds (via landmines).

Ping on this.

I'm not in a hurry but it's great if we can remove the WEB_AUDIO #ifdefs.

[Ken Russell (switch to Gerrit), 2015-05-20 18:27:03]

On 2015/05/20 04:57:34, haraken wrote:
> On 2015/04/24 12:34:40, Mike West (holiday in DE) wrote:
> > On 2015/04/24 at 05:33:23, haraken wrote:
> > > On 2014/09/25 18:21:39, Ken Russell wrote:
> > > > I don't see a good reason to land this patch. It will require a flag day
> and
> > > > reduces flexibility in case the code needs to be compiled out for some
> > reason.
> > > > 
> > > > Agree with @dcheng that we should not land CLs that require clobber
> builds.
> > > > Sometimes it's inevitable, and that's why the landmines mechanism is
> > present,
> > > > but doing so has high costs for all Chromium developers and should be
> > avoided.
> > > 
> > > We want to make this change at some point somehow because WEB_AUDIO is the
> > last remaining compile time flag, and by removing the flag, we can remove
lots
> > of code about [Conditional] from the IDL compiler.
> > > 
> > > Would there be any more modest way to land this CL?
> > 
> > I'll defer to krb@ as he had stronger opinions about the topic than I, but
> this
> > seems to me to be a good enough reason to override the general concern about
> > forcing clobber builds (via landmines).
> 
> Ping on this.
> 
> I'm not in a hurry but it's great if we can remove the WEB_AUDIO #ifdefs.

Sorry for the delay. If this will help eliminate a lot of code from the IDL
compiler, then making this change sounds good. However, avoiding clobber builds
is essential. Landmines can only be updated on the Chromium side, so landing
this in Blink will require two landmine updates: one to clear up the Blink-side
canary bots when this lands, and another for the next Blink roll. haraken, can
you figure out how to modify the bindings' dependencies to avoid the need for
these landmines?

[haraken, 2015-12-22 07:30:29]

On 2015/05/20 18:27:03, Ken Russell wrote:
> On 2015/05/20 04:57:34, haraken wrote:
> > On 2015/04/24 12:34:40, Mike West (holiday in DE) wrote:
> > > On 2015/04/24 at 05:33:23, haraken wrote:
> > > > On 2014/09/25 18:21:39, Ken Russell wrote:
> > > > > I don't see a good reason to land this patch. It will require a flag
day
> > and
> > > > > reduces flexibility in case the code needs to be compiled out for some
> > > reason.
> > > > > 
> > > > > Agree with @dcheng that we should not land CLs that require clobber
> > builds.
> > > > > Sometimes it's inevitable, and that's why the landmines mechanism is
> > > present,
> > > > > but doing so has high costs for all Chromium developers and should be
> > > avoided.
> > > > 
> > > > We want to make this change at some point somehow because WEB_AUDIO is
the
> > > last remaining compile time flag, and by removing the flag, we can remove
> lots
> > > of code about [Conditional] from the IDL compiler.
> > > > 
> > > > Would there be any more modest way to land this CL?
> > > 
> > > I'll defer to krb@ as he had stronger opinions about the topic than I, but
> > this
> > > seems to me to be a good enough reason to override the general concern
about
> > > forcing clobber builds (via landmines).
> > 
> > Ping on this.
> > 
> > I'm not in a hurry but it's great if we can remove the WEB_AUDIO #ifdefs.
> 
> Sorry for the delay. If this will help eliminate a lot of code from the IDL
> compiler, then making this change sounds good. However, avoiding clobber
builds
> is essential. Landmines can only be updated on the Chromium side, so landing
> this in Blink will require two landmine updates: one to clear up the
Blink-side
> canary bots when this lands, and another for the next Blink roll. haraken, can
> you figure out how to modify the bindings' dependencies to avoid the need for
> these landmines?

Let me revive this super old CL.

RashmiVijay@: Would you help me understand why this CL needs a clobber build?
Given that you're touching the IDL files, the auto-generated files will be just
auto-generated and re-compiled, won't it?

[Ken Russell (switch to Gerrit), 2015-12-22 19:04:55]

On 2015/12/22 07:30:29, haraken wrote:
> On 2015/05/20 18:27:03, Ken Russell wrote:
> > On 2015/05/20 04:57:34, haraken wrote:
> > > On 2015/04/24 12:34:40, Mike West (holiday in DE) wrote:
> > > > On 2015/04/24 at 05:33:23, haraken wrote:
> > > > > On 2014/09/25 18:21:39, Ken Russell wrote:
> > > > > > I don't see a good reason to land this patch. It will require a flag
> day
> > > and
> > > > > > reduces flexibility in case the code needs to be compiled out for
some
> > > > reason.
> > > > > > 
> > > > > > Agree with @dcheng that we should not land CLs that require clobber
> > > builds.
> > > > > > Sometimes it's inevitable, and that's why the landmines mechanism is
> > > > present,
> > > > > > but doing so has high costs for all Chromium developers and should
be
> > > > avoided.
> > > > > 
> > > > > We want to make this change at some point somehow because WEB_AUDIO is
> the
> > > > last remaining compile time flag, and by removing the flag, we can
remove
> > lots
> > > > of code about [Conditional] from the IDL compiler.
> > > > > 
> > > > > Would there be any more modest way to land this CL?
> > > > 
> > > > I'll defer to krb@ as he had stronger opinions about the topic than I,
but
> > > this
> > > > seems to me to be a good enough reason to override the general concern
> about
> > > > forcing clobber builds (via landmines).
> > > 
> > > Ping on this.
> > > 
> > > I'm not in a hurry but it's great if we can remove the WEB_AUDIO #ifdefs.
> > 
> > Sorry for the delay. If this will help eliminate a lot of code from the IDL
> > compiler, then making this change sounds good. However, avoiding clobber
> builds
> > is essential. Landmines can only be updated on the Chromium side, so landing
> > this in Blink will require two landmine updates: one to clear up the
> Blink-side
> > canary bots when this lands, and another for the next Blink roll. haraken,
can
> > you figure out how to modify the bindings' dependencies to avoid the need
for
> > these landmines?
> 
> Let me revive this super old CL.
> 
> RashmiVijay@: Would you help me understand why this CL needs a clobber build?
> Given that you're touching the IDL files, the auto-generated files will be
just
> auto-generated and re-compiled, won't it?

This CL will need to be regenerated. The base URL is now obsolete.

[haraken, 2016-01-05 11:59:55]

On 2015/12/22 19:04:55, Ken Russell wrote:
> On 2015/12/22 07:30:29, haraken wrote:
> > On 2015/05/20 18:27:03, Ken Russell wrote:
> > > On 2015/05/20 04:57:34, haraken wrote:
> > > > On 2015/04/24 12:34:40, Mike West (holiday in DE) wrote:
> > > > > On 2015/04/24 at 05:33:23, haraken wrote:
> > > > > > On 2014/09/25 18:21:39, Ken Russell wrote:
> > > > > > > I don't see a good reason to land this patch. It will require a
flag
> > day
> > > > and
> > > > > > > reduces flexibility in case the code needs to be compiled out for
> some
> > > > > reason.
> > > > > > > 
> > > > > > > Agree with @dcheng that we should not land CLs that require
clobber
> > > > builds.
> > > > > > > Sometimes it's inevitable, and that's why the landmines mechanism
is
> > > > > present,
> > > > > > > but doing so has high costs for all Chromium developers and should
> be
> > > > > avoided.
> > > > > > 
> > > > > > We want to make this change at some point somehow because WEB_AUDIO
is
> > the
> > > > > last remaining compile time flag, and by removing the flag, we can
> remove
> > > lots
> > > > > of code about [Conditional] from the IDL compiler.
> > > > > > 
> > > > > > Would there be any more modest way to land this CL?
> > > > > 
> > > > > I'll defer to krb@ as he had stronger opinions about the topic than I,
> but
> > > > this
> > > > > seems to me to be a good enough reason to override the general concern
> > about
> > > > > forcing clobber builds (via landmines).
> > > > 
> > > > Ping on this.
> > > > 
> > > > I'm not in a hurry but it's great if we can remove the WEB_AUDIO
#ifdefs.
> > > 
> > > Sorry for the delay. If this will help eliminate a lot of code from the
IDL
> > > compiler, then making this change sounds good. However, avoiding clobber
> > builds
> > > is essential. Landmines can only be updated on the Chromium side, so
landing
> > > this in Blink will require two landmine updates: one to clear up the
> > Blink-side
> > > canary bots when this lands, and another for the next Blink roll. haraken,
> can
> > > you figure out how to modify the bindings' dependencies to avoid the need
> for
> > > these landmines?
> > 
> > Let me revive this super old CL.
> > 
> > RashmiVijay@: Would you help me understand why this CL needs a clobber
build?
> > Given that you're touching the IDL files, the auto-generated files will be
> just
> > auto-generated and re-compiled, won't it?
> 
> This CL will need to be regenerated. The base URL is now obsolete.

I updated the CL in https://codereview.chromium.org/1557363002/.

[tommyw, 2016-01-07 12:05:35]

tommyw@google.com changed reviewers:
+ tommyw@google.com

[tommyw, 2016-01-07 12:05:36]

LGTM, Anything that simplifies the bindings generation gets my approval. And
Both run-time and compile-time flags seems excessive.

[Raymond Toy, 2016-01-12 22:21:13]

haraken's CL https://codereview.chromium.org/1557363002/ has landed.