Register about:pdf as a privileged-about page pointing at the landing
page, and add the AboutPDF JSWindowActor pair that bridges the sandboxed
content to chrome. The parent validates a dropped/picked file (file
scheme, .pdf extension, existence and %PDF- header) before navigating
the top browsing context to it, and, on Windows, exposes whether Firefox
can be set as the default PDF handler and performs that change.
Allow the page to call the RPM helpers it needs and to read/write the
promo dismissal pref.
Differential Revision: https://phabricator.services.mozilla.com/D303421
This patch does not make some header files stop including the inline
header because fixing them require more changes.
Note that `#include "nsComponentManagerUtils.h"` is fixing a bustage
caused by "use of undeclared identifier 'do_CreateInstance'" error.
Differential Revision: https://phabricator.services.mozilla.com/D304432
It's a 1:1 mapping for the origin that we cache in the
StyleSheetContents, so use that.
PreloadHashKey only cares about author loads, within a document, so
doesn't really need to deal with user / user-agent ones.
Differential Revision: https://phabricator.services.mozilla.com/D304400
This bug's other patch is about doing s/new/MakeRefPtr/ for reference-counted
variables; this patch here is just a drive-by to do s/new/MakeUnique/ for some
uniquely-owned variables.
Differential Revision: https://phabricator.services.mozilla.com/D302916
This reverts commit 36fc8f5360.
Revert "Bug 2039866 - FindAdjacentEntryFor should never return its input. r=dom-core,smaug"
This reverts commit 465532bd55.
This improves the naming of some of the various (ID -> Name) to reflect what they actually represent.
Additionally, this splits up the one mega run_test in test_URIFixup_info.js into multiple tasks, so that it is easier to see which stage of the test that a test failure occurs in.
Differential Revision: https://phabricator.services.mozilla.com/D301721
This removes the legacy telemetry mirroring for a large set of glean probes. All of these have their data readily available through glean, this means if there are queries depending on this some small changes maybe needed in order to make those work. This also updates a variety of tests that depended on the legacy telemetry mirrors.
Differential Revision: https://phabricator.services.mozilla.com/D301085
This removes the legacy telemetry mirroring for a large set of glean probes. All of these have their data readily available through glean, this means if there are queries depending on this some small changes maybe needed in order to make those work. This also updates a variety of tests that depended on the legacy telemetry mirrors.
Differential Revision: https://phabricator.services.mozilla.com/D301085
This removes the legacy telemetry mirroring for a large set of glean probes. All of these have their data readily available through glean, this means if there are queries depending on this some small changes maybe needed in order to make those work. This also updates a variety of tests that depended on the legacy telemetry mirrors.
Differential Revision: https://phabricator.services.mozilla.com/D301085
This removes the legacy telemetry mirroring for a large set of glean probes. All of these have their data readily available through glean, this means if there are queries depending on this some small changes maybe needed in order to make those work. This also updates a variety of tests that depended on the legacy telemetry mirrors.
Differential Revision: https://phabricator.services.mozilla.com/D301085
Some document.write() WPTs made the assumption that a setTimeout() in
the parent window will run after the load event in the child frame, and
there's no such guarantee per spec.
The focus fixup rule test similarly relied on the rendering not being
suppressed by the time the test run. Do that by awaiting an animation
frame, not just a timeout.
Differential Revision: https://phabricator.services.mozilla.com/D301259
Unlike VS, we're not constrained by legacy behavior and should always
turn this attribute on where it's needed. There's no VS flag for it, but
we can use clang-tidy to detect when it's needed.
It's okay to use that attribute when it has no effect, so don't bother
trying to be smart in dom/bindings/Codegen.py and always add it.
Differential Revision: https://phabricator.services.mozilla.com/D300299
Since bug 2037186 we already required the same allocation kind, but that can change
when we make object layout or GC changes so isn't ideal.
This change also simplifies later patches.
Differential Revision: https://phabricator.services.mozilla.com/D300252
velzie