Kathleen Brade authored 5 years ago and Pier Angelo Vendrame committed 1 year ago

When Tor informs the browser that client authentication is needed,
temporarily load about:blank instead of about:neterror and prompt
for the user's key.

If a correctly formatted key is entered, use Tor's ONION_CLIENT_AUTH_ADD
control port command to add the key (via Torbutton's control port
module) and reload the page.

If the user cancels the prompt, display the standard about:neterror
"Unable to connect" page. This requires a small change to
browser/actors/NetErrorChild.jsm to account for the fact that the
docShell no longer has the failedChannel information. The failedChannel
is used to extract TLS-related error info, which is not applicable
in the case of a canceled .onion authentication prompt.

Add a leaveOpen option to PopupNotifications.show so we can display
error messages within the popup notification doorhanger without
closing the prompt.

Add support for onion services strings to the TorStrings module.

Add support for Tor extended SOCKS errors (Tor proposal 304) to the
socket transport and SOCKS layers. Improved display of all of these
errors will be implemented as part of bug 30025.

Also fixes bug 19757:
 Add a "Remember this key" checkbox to the client auth prompt.

 Add an "Onion Services Authentication" section within the
 about:preferences "Privacy & Security section" to allow
 viewing and removal of v3 onion client auth keys that have
 been stored on disk.

Also fixes bug 19251: use enhanced error pages for onion service errors.

Bug 1752703 - Move the pthread_thread_create() interposer under mozglue and prepare for having a single place where we place interposer functions r=glandium a=RyanVM

Differential Revision: https://phabricator.services.mozilla.com/D164470

Differential Revision: https://phabricator.services.mozilla.com/D156605

Differential Revision: https://phabricator.services.mozilla.com/D152822

Differential Revision: https://phabricator.services.mozilla.com/D147305

Differential Revision: https://phabricator.services.mozilla.com/D147276

This removes the API and skips the tests. The implementation in
mozJSComponentLoader is not removed.

This patch will be backed out after the next merge.

Differential Revision: https://phabricator.services.mozilla.com/D147278

Differential Revision: https://phabricator.services.mozilla.com/D146347

Bug 1768871 - Remove Marionette test js/xpconnect/tests/marionette/test_preloader_telemetry.py r=webdriver-reviewers,dthayer,jdescottes

Turns out that this test didn't provide any value since it was
added. As such lets get it removed.

Differential Revision: https://phabricator.services.mozilla.com/D146628

Bug 1514936, part 3 - Remove NS_ERROR_NO_AGGREGATION which is no longer used. r=xpcom-reviewers,nika

Differential Revision: https://phabricator.services.mozilla.com/D146458

Bug 1514936, part 1 - Remove the outer argument to nsIFactory::createInstance. r=xpcom-reviewers,preferences-reviewers,nika,Gijs

This patch won't actually build, because a few bits of code are used
for both nsIFactory::createInstance and static components, and static
components are not fixed until the next patch.

The first place is nsLoadGroupConstructor, which uses an nsIFactory
macro to create a static component constructor. (This could be worked
around by expanding the macro to the state before this patch.)

The other issue is that nsAppShellConstructor is used in an nsIFactory
on OSX, but as a static component on all other platforms. This could
be worked around by wrapping nsAppShellConstructor in an adaptor that
passes in the extra null argument to nsAppShellConstructor.

Differential Revision: https://phabricator.services.mozilla.com/D146456

Differential Revision: https://phabricator.services.mozilla.com/D146169

Differential Revision: https://phabricator.services.mozilla.com/D146166

Differential Revision: https://phabricator.services.mozilla.com/D145219

Bug 1768060 - Add a wrapper for Cu.import return value that supports lexical variable. r=jonco,Standard8

This wrapper exposes all lexical variables in Cu.import return value,
that allows us removing `this.foo = foo;` hack, in bug 1610653 patches,
without affecting the not-in-tree consumer.

Differential Revision: https://phabricator.services.mozilla.com/D145938

Bug 1766761 - Part 4: Add a shim for Cu.import/ChromeUtils.import with ESM-ified module. r=jonco,Standard8

This shim redirects Cu.import("foo.jsm") to ChromeUtils.importModule("foo.mjs")
if foo.jsm is already renamed to foo.mjs.

This de-optimizes ESM component's global `var`s, in order to expose all
variables in `ModuleEnvironmentObject`, and return those variables from
`Cu.improt`'s return value.

In contrast to normal Cu.import, the return value of shim version exposes
all non-exported lexical variables.
Bug 1768060 patch aligns the normal Cu.import to expose them as well,
so that `this.foo = foo;` hack can be removed in bug 1610653.

Once not-in-tree `Cu.import` consumers are gone and `Cu.import` is removed,
this shim and the de-optimization for the `var` can be removed.

Differential Revision: https://phabricator.services.mozilla.com/D145784

Differential Revision: https://phabricator.services.mozilla.com/D145810

Backed out 8 changesets (bug 1766976, bug 1768922, bug 1769029, bug 1766761, bug 1768060) for causing spidermonkey failures on Modules.cpp CLOSED TREE

Backed out changeset da102935dfd2 (bug 1769029)
Backed out changeset e080258f38a7 (bug 1768922)
Backed out changeset 36e9e0301a24 (bug 1766976)
Backed out changeset dd28b0eb27ef (bug 1768060)
Backed out changeset 682c4afbcfe9 (bug 1766761)
Backed out changeset a7a37e912b90 (bug 1766761)
Backed out changeset d8099b6d970b (bug 1766761)
Backed out changeset 435acc6d6abd (bug 1766761)

Differential Revision: https://phabricator.services.mozilla.com/D146169

Differential Revision: https://phabricator.services.mozilla.com/D146166

Differential Revision: https://phabricator.services.mozilla.com/D145219

Bug 1768060 - Add a wrapper for Cu.import return value that supports lexical variable. r=jonco,Standard8

This wrapper exposes all lexical variables in Cu.import return value,
that allows us removing `this.foo = foo;` hack, in bug 1610653 patches,
without affecting the not-in-tree consumer.

Differential Revision: https://phabricator.services.mozilla.com/D145938

Bug 1766761 - Part 4: Add a shim for Cu.import/ChromeUtils.import with ESM-ified module. r=jonco,Standard8

This shim redirects Cu.import("foo.jsm") to ChromeUtils.importModule("foo.mjs")
if foo.jsm is already renamed to foo.mjs.

This de-optimizes ESM component's global `var`s, in order to expose all
variables in `ModuleEnvironmentObject`, and return those variables from
`Cu.improt`'s return value.

In contrast to normal Cu.import, the return value of shim version exposes
all non-exported lexical variables.
Bug 1768060 patch aligns the normal Cu.import to expose them as well,
so that `this.foo = foo;` hack can be removed in bug 1610653.

Once not-in-tree `Cu.import` consumers are gone and `Cu.import` is removed,
this shim and the de-optimization for the `var` can be removed.

Differential Revision: https://phabricator.services.mozilla.com/D145784

The value is never used.

Differential Revision: https://phabricator.services.mozilla.com/D146150

Backed out 3 changesets (bug 1758745, bug 1758740, bug 1762462) for causing mochitest failures on test_storagePermissionsAccept.html CLOSED TREE

Backed out changeset fa222c0a9a26 (bug 1762462)
Backed out changeset 9daa04b239f9 (bug 1758745)
Backed out changeset c18aecb431c0 (bug 1758740)

Bug 1758745 - Refactor EffectiveStoragePrincipal to EffectiveCookiePrincipal and create new function EffectiveStoragePrincipal that is never unpartitioned; r=twisniewski

Depends on D143626

Differential Revision: https://phabricator.services.mozilla.com/D143627

Backed out 3 changesets (bug 1758745, bug 1758740, bug 1762462) for causing failures at browser_storage_dfpi.js. CLOSED TREE

Backed out changeset a3024e8b2a58 (bug 1762462)
Backed out changeset 8bc1291b9bdb (bug 1758745)
Backed out changeset c7b693b2fd05 (bug 1758740)

Differential Revision: https://phabricator.services.mozilla.com/D145568

Since mozJSComponentLoader is destroyed after the cycle collector, we need to
break cycles manually before the final cycle collection otherwise we get a
memory leak. Unload() is called at an appropriate earlier point in shutdown.

Differential Revision: https://phabricator.services.mozilla.com/D145567

Bug 1432901 - Part 10: Implement mozJSComponentLoader ImportModule method to synchronously import an ES6 module r=yulia

This is the main patch that finally implements the import.

The derived module loader uses a list of pending load requests to drive the system.

Differential Revision: https://phabricator.services.mozilla.com/D145565

Differential Revision: https://phabricator.services.mozilla.com/D145564

Bug 1432901 - Part 8: Add option to compile source to module stencil in mozJSComponentLoader r=yulia

This adds a flag to ComponentLoaderInfo to say whether it represents a module
and calls the appriate JS APIs based on this.

Differential Revision: https://phabricator.services.mozilla.com/D145562

This will be used to hold compilation results before they are passed to the base class.

Differential Revision: https://phabricator.services.mozilla.com/D145561

Bug 1432901 - Part 6: Give our new module loader an event target that dispatches synchronously r=yulia

As explained in the bug, we need to dispatch without using the event loop so we
can complete the import synchronously.

Differential Revision: https://phabricator.services.mozilla.com/D145560

Differential Revision: https://phabricator.services.mozilla.com/D145558

This is called ComponentModuleLoader rather than ModuleLoader as memory leak
detection doesn't like it if we have more than one class with the same name,
even if they're in different namespaces. Other name suggestions welcome.

Differential Revision: https://phabricator.services.mozilla.com/D145557

Bug 1758745 - Refactor EffectiveStoragePrincipal to EffectiveCookiePrincipal and create new function EffectiveStoragePrincipal that is never unpartitioned; r=twisniewski

Depends on D143626

Differential Revision: https://phabricator.services.mozilla.com/D143627

Differential Revision: https://phabricator.services.mozilla.com/D145473

Differential Revision: https://phabricator.services.mozilla.com/D145472

Bug 1767800 - Part 2: Factor out a method to determine if a URL is a real file from PrepareObjectForLocation r=mccr8

We won't call PrepareObjectForLocation when loading ES6 modules.

Differential Revision: https://phabricator.services.mozilla.com/D145471