Slimming Paint phase 1 launch for Chrome 45

[Chris Harrelson]

Hi blink-dev, gardeners and perf sheriffs,

The Slimming Paint phase 1 launch is arriving soon!

The immediate impact is that wkorman@ is turning on the SlimmingPaintEnabled flag with status=stable as soon as we can get it through the CQ.

Please let paint-dev know if you observe regressions or failures, we will respond quickly. Or if the problem is bad feel free to flip the flag back off while we debug it.

More info about the rollout plan is available here:

https://docs.google.com/document/d/1CrASjQ76y-aTBil8dYvE-iBJq2uqNYWcZUJovr6UbFY/edit#

Thanks,

Chris

[Levi Weintraub]

http://weknowgifs.com/wp-content/uploads/2013/04/its-happening-ron-paul-gif.gif

--
You received this message because you are subscribed to the Google Groups "paint-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to paint-dev+...@chromium.org.
To post to this group, send email to pain...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/paint-dev/CAOMQ%2Bw8pCUCTpxpaR%3D3Zxtn-XR4WLJ6UXSDO61J_2V6KEWyuVg%40mail.gmail.com.

[Dimitri Glazkov]

Drats. Gardening Wednesday and Thursday.

:DG<

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/paint-dev/CAMLeo%2BGnO4H%2BKdNrQd95ke1CCJSA-eVf%3Dt4F5P2xY5v4O%3DJNew%40mail.gmail.com.

[Dimitri Glazkov]

I mean. YAYAYAY!!!

:DG<

[Dominic Cooney]

I don't see any plan to monitor crashes. Is there one? I'd like to look over the queries you are planning to use. I understand if those unfortunately have to be Google-only.

PSA: Non-Google contributors, I'm happy to help monitor crashes if you're turning on a big feature, just get in touch with me--and blink-dev, I'm sure other Google Chromium contributors would be happy to pitch in--a week or so ahead of time to strategize.

Dominic, today's Chrome Stability Sheriff

[Chris Harrelson]

On Mon, Jun 8, 2015 at 6:09 PM, Dominic Cooney <domi...@chromium.org> wrote:

I don't see any plan to monitor crashes. Is there one? I'd like to look over the queries you are planning to use. I understand if those unfortunately have to be Google-only.

You're right that we have no specific plan. To be honest, I am also not so expert at crafting queries for crashes, any advice appreciated.

[Dominic Cooney]

I'll follow up with you directly.

[Chris Harrelson]

Update: as of a few minutes ago, slimming paint is now on by default. Use --disable-slimming-paint to turn it off when testing.

Blink layout tests are still run without slimming paint. We will switch them over once we are sure the Chromium setting will stick.

On Mon, Jun 8, 2015 at 2:35 PM, Chris Harrelson <chri...@chromium.org> wrote:

[c.el...@method.me]

In Chrome 46 it appears as though the paint slimming flag is either defaulting to on, or that there is a bug in the toggle. Sites that use multiple iframes become increasingly less responsive from a UI standpoint. Enabling the "disable paint slimming" flag solves the problem entirely for those who encounter it. Is there a way to bypass this feature through code?

[Jeremy Roman]

No, web content cannot opt out of slimming paint. If you are experiencing a noticeable performance degradation, that is a bug. Please file it at http://crbug.com/new (ideally with a reliable reproduction case).

--
You received this message because you are subscribed to the Google Groups "paint-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to paint-dev+...@chromium.org.
To post to this group, send email to pain...@chromium.org.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/paint-dev/8a2f71f1-bdd7-4ef4-a6f7-435f8a299380%40chromium.org.

[Carl Elliott]

Thanks Jeremy, I'll log the defect as soon as we can put together a simpler reproduction case than the environment we are currently seeing it on.

Interestingly, enabling the "Disable" flag for this appears to be a permanent fix, hence my reference to a potential problem with the toggle itself. Once we enable this we are unable to reproduce the slowdowns, even if we flip it back to disable. Is there a way to verify if slimming paint is enabled/disabled other than looking at the flags?

[c.el...@method.me]

Thanks Jeremy, I'll log the defect as soon as we can put together a simpler reproduction case than the environment we are currently seeing it on.

Interestingly, enabling the "Disable" flag for this appears to be a permanent fix, hence my reference to a potential problem with the toggle itself: Once we enable this we are unable to reproduce the slowdowns, even if we flip it back to disable. Is there a way to verify if slimming paint is enabled/disabled other than looking at the flags?

[Walter Korman]

Did you restart the browser after flipping disable back to off? There is no explicit way to check whether it's on that I know of.

You can also pass '--disable-slimming-paint' or '--enable-slimming-paint' on the command line, and the latter will take precedence over disable if present. So if you're worried that SP is disabled, then if you start chrome with '--enable-slimming-paint' that should force it to on.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/paint-dev/5c27fd1b-47c8-4e27-ae87-8d4cf71ebaa5%40chromium.org.

[Jeremy Roman]

The logic for turning SP on/off is regrettably complicated, and I don't think we have an easy way to tell whether it's on. Currently it seems it's on if "Enable Slimming Paint" is on, or if "Disable Slimming Paint" is off and you haven't been selected for the hold-back experiment group.

You can see what the full switch list used by Chrome is in chrome://version; if either --enable-slimming-paint or --disable-slimming-paint is provided (via chrome://flags or the command line), it takes precedence over the experiment group.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/paint-dev/5c27fd1b-47c8-4e27-ae87-8d4cf71ebaa5%40chromium.org.

[wko...@chromium.org]

Hi Carl, just checking to see whether you've got a sample url or similar we can take a look at to investigate the iframe performance issue. If you've filed a bug with info can you reply with link? Thanks.

[Carl Elliott]

We've created a reproduction case and submitted a defect at the following URL: https://code.google.com/p/chromium/issues/detail?id=525656&thanks=525656&ts=1440687352

Thanks for looking into it.

Carl