1607027 - css @media device-pixel-ratio and RFP

Status: RESOLVED INVALID

(Core :: CSS Parsing and Computation, enhancement)

Description

[Thorin [:thorin]]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:68.0) Gecko/20100101 Firefox/68.0

Steps to reproduce:

Note: For my test, my device/monitor has a device pixel ratio of 1. I am not sure what happens on HiDPI devices like a Mac with Retina.

control test

• disable RFP
• load test page: see [1]
• devicePixelRatio = 1 + green RFP notation
• [matchMedia -webkit- | -moz-] device-pixel-ratio = 1 | 1 + green RFP
• zoom, e.g to 120%
• devicePixelRatio = 1.2 + red RFP
• [matchMedia -webkit- | -moz-] device-pixel-ratio = 1.2 | 1.2 (rounded) + red RFP

the test

• enable RFP
• reload the test page at 100% zoom
• devicePixelRatio = 1 + green RFP
• [matchMedia -webkit- | -moz-] device-pixel-ratio = 1 | 1 + green RFP
• zoom, e.g to 120%
• devicePixelRatio = 1 + green RFP = spoofed
• [matchMedia -webkit- | -moz-] device-pixel-ratio = 1.2 | 1.2 + red RFP = not spoofed

** the problem**

• window.devicePixelRatio is protected, but -webkit-{min/max}-device-pixel-ratio and {min/max-}-moz-device-pixel-ratio are not

references

• [1] Test page: https://ghacksuserjs.github.io/TorZillaPrint/TorZillaPrint.html (at the end of the first section titled Screen)
• Bug 418986
• https://dxr.mozilla.org/mozilla-central/source/dom/base/test/chrome/bug418986-1.js

---

android

• this is the only device I have that has a dpr != 1
• The resize event on the test page (which auto recalculates the values) is only triggered on non-android devices
• I can pinch-to-zoom and use the three dot menu and refresh (or I can just hit the re-reun button) and all three values are always the same. With RFP on they are always 1. With RFP off they report the correct device pixel ratio (for me that's 2.608695...)
• I'm sure I'm misinterpreting how android + pinch-to-zoom + dpr work

Comment 1

[Thorin [:thorin]]

Attachment: at 110% zoom: top RFP is off, bottom RFP is on

Comment 2

[Tom Ritter [:tjr]]

You reopened the other two, is this supposed to be reopened as well?

Comment 3

[Thorin [:thorin]]

no, this is invalid, we do protect the value - https://searchfox.org/mozilla-central/source/layout/style/nsMediaFeatures.cpp#197 (it still has issues, but not at 100% zoom which we set on new tabs/eTLDs) - I inadvertently closed the wrong one Bug 1672093 because the titles were similar