Bug 88121 – [chromium] Support high DPI pixel tests with DumpRenderTree

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Terry Anderson]

Add the flag --compositor-device-scale-factor=n to DRT for Chromium.  This will set the device scale factor to be the integer n and will also have the same effect as using the existing flag --force-compositing-mode.

[Dirk Pranke]

You can already pass arbitrary flags to DRT with --additional-drt-flag "--compositor-device-scale-factor=n"; will that work for this?

[Dana Jansens]

"have the same effect as using the existing flag --force-compositing-mode" -> "also implies the use of --force-compositing-mode" ?

[Terry Anderson]

Created attachment 145359 [details]
Patch

[WebKit Review Bot]

Please wait for approval from [e-mail], [e-mail], [e-mail], [e-mail] or [e-mail] before submitting, as this patch contains changes to the Chromium public API. See also https://trac.webkit.org/wiki/ChromiumWebKitAPI.

[Terry Anderson]

(In reply to comment #2)
> "have the same effect as using the existing flag --force-compositing-mode" -> "also implies the use of --force-compositing-mode" ?

Yes, that is much better wording. I've updated this in the ChangeLog description.

[Dirk Pranke]

Comment on attachment 145359 [details]
Patch

I probably didn't phrase my previous comment clearly enough. We shouldn't be adding more flags to NRWT unless we have to; there's too many already. Given that I believe you can already get the behavior you want using --additional-drt-flag, I don't think you need this.

[Terry Anderson]

Created attachment 147400 [details]
Patch

[Terry Anderson]

(In reply to comment #7)
> Created an attachment (id=147400) [details]
> Patch

There were some additional changes made in light of https://bugs.webkit.org/show_bug.cgi?id=88375 landing since I uploaded my first patch.

[James Robinson]

Comment on attachment 147400 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=147400&action=review

> Source/WebKit/chromium/public/WebWidget.h:99
> +    virtual void paint(WebCanvas*, const WebRect& viewPort, const int scaleFactor = 1) { }

this seems misguided.  We already have a deviceScaleFactor and pageScaleFactor setter on WebView, why are they not sufficient?

[James Robinson]

We also need to support non-integer device scales in general.  IMO it's OK to start with only handling integer scales at first, but we're going to have non-integer scales very soon so we need to make sure the design and API are amenable.

[Terry Anderson]

*** Bug 70066 has been marked as a duplicate of this bug. ***