There are all sorts of lifecycle issues which arise from making DocShell
responsible for discarding BrowsingContexts. In this particular bug, we tend
to run into them in cases where we create a BrowsingContext for a FrameLoader,
and then never create a DocShell for it, leading to it never being destroyed.
But there are myriad other issues as well.
This patch moves the responsibility for BrowsingContext lifecycle management
to the FrameLoader/FrameLoaderOwner, rather than the DocShell, which makes
things more consistent, and more closely aligns with spec-defined behavior.
Differential Revision: https://phabricator.services.mozilla.com/D59008
--HG--
extra : moz-landing-system : lando
See bug 1613275 and bug 1607845. In bug 1607845, the aim was to regenerate all
test certificates that would be expiring. Unfortunately, a few were missed:
* build/pgo/certs/ certificate DBs and mochitest.client are regenerated in a
different way than the rest of the certificates in bug 1607845. These would
probably best be addressed by formally documenting the process of
re-generating all of the certificates.
* security/manager/ssl/tests/unit/test_certDB_import/ certificates were
missed by mistake. It's unclear how this happened.
* security/manager/ssl/tests/unit/test_intermediate_preloads/ were missed
because there was no test_intermediate_preloads entry in the TEST_DIRS
section of security/manager/ssl/tests/unit/moz.build, which means that the
build system never knew to re-generate those certificates, even after
un-commenting-out the contents of
security/manager/ssl/tests/unit/test_intermediate_preloads/moz.build
* security/manager/ssl/tests/unit/test_missing_intermediate/missing-intermediate.der
was DER, not PEM, and we don't have a way to automatically re-generate DER
certificates in the same way. However, it didn't even need to be DER.
Differential Revision: https://phabricator.services.mozilla.com/D61712
--HG--
extra : moz-landing-system : lando
Otherwise we're using state from the pre-open document for whatever content is
being written, which is not likely to be right.
Differential Revision: https://phabricator.services.mozilla.com/D61865
--HG--
extra : moz-landing-system : lando
Otherwise we're using state from the pre-open document for whatever content is
being written, which is not likely to be right.
Differential Revision: https://phabricator.services.mozilla.com/D61865
--HG--
extra : moz-landing-system : lando
Every day clang's warnings continue to amaze me. As does our ability to trigger them.
I'm assuming that this line was meant to go within the scope of the `if`, because if it were outside the scope, then we'd be duplicating the previous `fprintf`.
Differential Revision: https://phabricator.services.mozilla.com/D61835
--HG--
extra : moz-landing-system : lando
Upstream patch:
6bd491daaf%5E%21/#F0
_sifields is a glibc-internal field, and is not available on musl
libc. Instead, use the public-facing fields si_call_addr, si_syscall,
and si_arch, if they are available.
Differential Revision: https://phabricator.services.mozilla.com/D61051
--HG--
extra : moz-landing-system : lando
`nsIDocShellTreeItem.rootTreeItem` has been deprecated.
We should use an approach based on BrowsingContext.
Differential Revision: https://phabricator.services.mozilla.com/D61496
--HG--
extra : moz-landing-system : lando
The original issue in bug 1565902 which prompted this investigation
seems to have gone away with a Windows OS update. We saw mixed results
in telemetry over this anyway, and no clear and obvious path for
getting the benefits without the drawbacks, so removing this seems
to be the best solution right now.
Differential Revision: https://phabricator.services.mozilla.com/D61720
--HG--
extra : moz-landing-system : lando
This change improves the intermittency of the test, but does not provably
eliminate it. The code comment in the relevant section gives the rationale.
Differential Revision: https://phabricator.services.mozilla.com/D61631
--HG--
extra : moz-landing-system : lando
Implement WaylandDMABUFSurfaceImage which is backed by dma buf memory on Wayland and it holds
decoded video images produced by ffmpeg va-api decoder.
Differential Revision: https://phabricator.services.mozilla.com/D61678
--HG--
extra : moz-landing-system : lando
Wires up the addon manager update process to the GekcoView API. Adds
several basic tests for verifying the udpate steps. Adds a new way to
automatically bundle new test addons into xpi files and registers them
with the example.com server.
Differential Revision: https://phabricator.services.mozilla.com/D61501
--HG--
extra : moz-landing-system : lando
I'm taking the opportunity to remove isDefaultHandlerApp messaging as the
content-side calls were removed in bug 1353029.
Depends on D59788
Differential Revision: https://phabricator.services.mozilla.com/D59789
--HG--
extra : moz-landing-system : lando
While using always the underline thickness seemed ok when I reviewed the
regressing bug, it seems it's not always fine to use the font-provided value.
At least I see the same thing as
https://bugzilla.mozilla.org/show_bug.cgi?id=1613285#c0 it in two different
(though similarly configured, in fairness) Linux machines.
Keep `auto` meaning the slightly different thing that it used to mean, and use
`from-font` for the underline thickness (though there's at least one case where
`from-font` could maybe be the `strikeOutSize`... wdyt?).
In particular, for spell-checking markers, we have a somewhat bespoke
`ComputeSelectionUnderlineHeight`...
Differential Revision: https://phabricator.services.mozilla.com/D61781
--HG--
extra : moz-landing-system : lando
Having test files in moz.build meant we were shipping those in Firefox.
The better way of doing this is to add those files as support files,
and then load them through a chrome:// url.
The catch is that support files from another folder than the ini files
they are referenced in need to be referenced in an ini file in their
own folder.
And since we can't have a ini file without tests, we need to add a dummy
test in devtools/client/webconsole/test/node/fixtures/stubs/ . The "test"
is always skipped so it shouldn't have any impact.
Doing this work revealed an issue with NetworkEvent stubs, which were
out of sync, and not properly writter since Bug 1579090.
This patch fixes this as well, and we're lucky the node tests are still
passing fine.
Differential Revision: https://phabricator.services.mozilla.com/D61696
--HG--
extra : moz-landing-system : lando
