1863613 - RFP no longer protects pdfViewerEnabled, plugins or mimeTypes

Status: RESOLVED INVALID

(Core :: DOM: Core & HTML, defect)

Description

[Thorin [:thorin]]

Bug 1756280 when RFP is enabled, always return hardcoded values for plugins and mimeTypes, and pdfViewerEnabled as true

STR

• test: https://arkenfox.github.io/TZP/tzp.html#devices
• enable RFP
• set pdfjs.disabled = true
• ESR115: test returns expected hash (pdfjs enabled etc)
• FF116+: test returns unexpected hash (pdfjs disabled etc)

Regression: Bug 1838415

expected hash: 91073152

{
  "mimeTypes": [
    "application/pdf: application/pdf: pdf",
    "text/pdf: text/pdf: pdf"
  ],
  "pdfViewerEnabled": true,
  "plugins": [
    "PDF Viewer: internal-pdf-viewer: Portable Document Format",
    "Chrome PDF Viewer: internal-pdf-viewer: Portable Document Format",
    "Chromium PDF Viewer: internal-pdf-viewer: Portable Document Format",
    "Microsoft Edge PDF Viewer: internal-pdf-viewer: Portable Document Format",
    "WebKit built-in PDF: internal-pdf-viewer: Portable Document Format"
  ]
}

unexpected hash: beccb452

{
  "mimeTypes": "none",
  "pdfViewerEnabled": false,
  "plugins": "none"
}

Comment 1

[Thorin [:thorin]]

FYI: moz-regression specifically output the following

Bug 1838415 - Introduce RFPTarget::MouseEventScreenPoint. r=tjr

edit: this is more likely

• https://phabricator.services.mozilla.com/D180938 Don't spoof explicitly disabled pdfJS
• If this is an intended change in behavior for Firefox users, then close as invalid and we will lock the pref in Tor Browser

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1838415

Comment 3

[Tom Ritter [:tjr]]

If this is an intended change in behavior for Firefox users, then close as invalid and we will lock the pref in Tor Browser

Yes, this is intentional, I don't recall explicitly what prompted us to revert this, but it aligns with this norm: https://gitlab.torproject.org/tpo/applications/tor-browser-spec/-/wikis/RFP-Behavior#do-we-override-settings-and-preferences