[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

This is to match the expected framework behavior for O low end devices
(see internal bug go/gries).

BUG=711079
Review-Url: https://codereview.chromium.org/2814323002
Cr-Commit-Position: refs/heads/master@{#464503}
Committed: https://chromium.googlesource.com/chromium/src/+/16616a21c2ed290d72aba2e90cabf1c50799cebd

[Maria, 2017-04-13 00:02:00]

mariakhomenko@chromium.org changed reviewers:
+ primiano@chromium.org, yfriedman@chromium.org

[Maria, 2017-04-13 00:02:00]

Yaron, I will need you for OWNERs approval.

Primiano, I am making this change now to make the testing easier without having
to set flags manually.

[Primiano Tucci (use gerrit), 2017-04-13 08:58:06]

Yeah I took a look to the internal bug (go/gries) and seems
internally-controversial.
I wonder if the right thing to do here in the long term is:

if (isAtLeastO()) 
   return the_other_flag_if_they_introduce_it

anyways, this for now LGTM. Let's keep an eye on that space.

Maybe in the commit message you can mention something like: This is to match the
expected framework behavior for O low end devices (see internal bug go/gries).

[Maria, 2017-04-13 17:56:18]

Description was changed from

==========
[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

BUG=711079
==========

to

==========
[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

This is to match the expected framework behavior for O low end devices
(see internal bug go/gries).

BUG=711079
==========

[Maria, 2017-04-13 17:57:32]

On 2017/04/13 08:58:06, Primiano Tucci wrote:
> Yeah I took a look to the internal bug (go/gries) and seems
> internally-controversial.
> I wonder if the right thing to do here in the long term is:
> 
> if (isAtLeastO()) 
>    return the_other_flag_if_they_introduce_it
> 
> anyways, this for now LGTM. Let's keep an eye on that space.
> 
> Maybe in the commit message you can mention something like: This is to match
the
> expected framework behavior for O low end devices (see internal bug go/gries).

Yep. Done. And will keep an eye on that discussion.

[Yaron, 2017-04-13 18:06:49]

On 2017/04/13 17:57:32, Maria wrote:
> On 2017/04/13 08:58:06, Primiano Tucci wrote:
> > Yeah I took a look to the internal bug (go/gries) and seems
> > internally-controversial.
> > I wonder if the right thing to do here in the long term is:
> > 
> > if (isAtLeastO()) 
> >    return the_other_flag_if_they_introduce_it
> > 
> > anyways, this for now LGTM. Let's keep an eye on that space.
> > 
> > Maybe in the commit message you can mention something like: This is to match
> the
> > expected framework behavior for O low end devices (see internal bug
go/gries).
> 
> Yep. Done. And will keep an eye on that discussion.

ugh. I don't like that this varies by platform. Perhaps we should treat everyone
<1GB as low-ram but that would change the experience for current users. Still
it'd be weird to be using a 1GB device on N, then get an OTA and suddenly we
treat you as low end and get a different experience :/

Isn't this an argument for just using the Android API rather than deciding for
ourselves?

[Maria, 2017-04-13 18:11:58]

On 2017/04/13 18:06:49, Yaron (limited availability) wrote:
> On 2017/04/13 17:57:32, Maria wrote:
> > On 2017/04/13 08:58:06, Primiano Tucci wrote:
> > > Yeah I took a look to the internal bug (go/gries) and seems
> > > internally-controversial.
> > > I wonder if the right thing to do here in the long term is:
> > > 
> > > if (isAtLeastO()) 
> > >    return the_other_flag_if_they_introduce_it
> > > 
> > > anyways, this for now LGTM. Let's keep an eye on that space.
> > > 
> > > Maybe in the commit message you can mention something like: This is to
match
> > the
> > > expected framework behavior for O low end devices (see internal bug
> go/gries).
> > 
> > Yep. Done. And will keep an eye on that discussion.
> 
> ugh. I don't like that this varies by platform. Perhaps we should treat
everyone
> <1GB as low-ram but that would change the experience for current users. Still
> it'd be weird to be using a 1GB device on N, then get an OTA and suddenly we
> treat you as low end and get a different experience :/
> 
> Isn't this an argument for just using the Android API rather than deciding for
> ourselves?

The problem with using Android API is that some larger tablets have flipped the
flag in the OS that turns on the low end mode and end up with horrible 16-bit
graphics Chrome experience. Sure we can say they brought it on themselves, but
that's not helpful to the users. I think longer term, we may want to decouple
decisions about graphics resolution from isLowEndDevice() bit and base it on RAM
directly, but keep the features on/off toggles tied to the Android bit. However,
we are not there yet, and I definitely don't want to change the existing users
to a different version of Chrome.

[Yaron, 2017-04-13 18:17:44]

On 2017/04/13 18:11:58, Maria wrote:
> On 2017/04/13 18:06:49, Yaron (limited availability) wrote:
> > On 2017/04/13 17:57:32, Maria wrote:
> > > On 2017/04/13 08:58:06, Primiano Tucci wrote:
> > > > Yeah I took a look to the internal bug (go/gries) and seems
> > > > internally-controversial.
> > > > I wonder if the right thing to do here in the long term is:
> > > > 
> > > > if (isAtLeastO()) 
> > > >    return the_other_flag_if_they_introduce_it
> > > > 
> > > > anyways, this for now LGTM. Let's keep an eye on that space.
> > > > 
> > > > Maybe in the commit message you can mention something like: This is to
> match
> > > the
> > > > expected framework behavior for O low end devices (see internal bug
> > go/gries).
> > > 
> > > Yep. Done. And will keep an eye on that discussion.
> > 
> > ugh. I don't like that this varies by platform. Perhaps we should treat
> everyone
> > <1GB as low-ram but that would change the experience for current users.
Still
> > it'd be weird to be using a 1GB device on N, then get an OTA and suddenly we
> > treat you as low end and get a different experience :/
> > 
> > Isn't this an argument for just using the Android API rather than deciding
for
> > ourselves?
> 
> The problem with using Android API is that some larger tablets have flipped
the
> flag in the OS that turns on the low end mode and end up with horrible 16-bit
> graphics Chrome experience. Sure we can say they brought it on themselves, but
> that's not helpful to the users. I think longer term, we may want to decouple
> decisions about graphics resolution from isLowEndDevice() bit and base it on
RAM
> directly, but keep the features on/off toggles tied to the Android bit.

Sounds reasonable to me.

 However,
> we are not there yet, and I definitely don't want to change the existing users
> to a different version of Chrome.

ok lgtm as a short-term solution

[Maria, 2017-04-13 18:29:47]

The CQ bit was checked by mariakhomenko@chromium.org

[Maria, 2017-04-13 18:29:48]

The patchset sent to the CQ was uploaded after l-g-t-m from
primiano@chromium.org
Link to the patchset: https://codereview.chromium.org/2814323002/#ps20001
(title: "Fixed description")

[commit-bot: I haz the power, 2017-04-13 18:30:20]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-04-13 19:29:56]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1492108187963890,
"parent_rev": "2f2991ce896d93583dc3dab2ba6466fcae3d23b3", "commit_rev":
"16616a21c2ed290d72aba2e90cabf1c50799cebd"}

[commit-bot: I haz the power, 2017-04-13 19:30:51]

Description was changed from

==========
[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

This is to match the expected framework behavior for O low end devices
(see internal bug go/gries).

BUG=711079
==========

to

==========
[LowMemory] Increase IsLowEndDevice() triggering to 1GB devices in O.

This is to match the expected framework behavior for O low end devices
(see internal bug go/gries).

BUG=711079

Review-Url: https://codereview.chromium.org/2814323002
Cr-Commit-Position: refs/heads/master@{#464503}
Committed:
https://chromium.googlesource.com/chromium/src/+/16616a21c2ed290d72aba2e90cab...
==========

[commit-bot: I haz the power, 2017-04-13 19:30:52]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/16616a21c2ed290d72aba2e90cab...