Commit graph

3,196 commits

Author SHA1 Message Date
wolfbeast
84818be343 No issue - Remove unused network.http.bypass-cachelock-threshold pref. 2020-01-17 09:17:18 +08:00
wolfbeast
540e9c1500 No issue - Don't define gamepad prefs when gamepad support isn't built.
This also sets sane defaults when present. See also #1319.
2020-01-17 09:17:15 +08:00
wolfbeast
fe9d8da299 Issue #1319 - Enable promise debugging only in DEBUG builds. 2020-01-17 09:17:13 +08:00
wolfbeast
c9a887d46b Issue #1319 - Disable MOZ_LOGGING in production builds.
This makes it only enabled in debug builds, instead, where logging
would be most useful.
2020-01-17 09:17:10 +08:00
wolfbeast
d28c0b15a6 Issue #1319 - Remove stderr_to_file.
This local-debug function is only useful for Android to work around
issues with logspewing and should never be in production builds anyway.
2020-01-17 09:17:08 +08:00
wolfbeast
47e85541f1 Fix comments for NSS PBKDF setup (no code change) 2020-01-17 09:17:04 +08:00
Kai Engert
3d75257e8d Issue #1338: Follow-up: Cache the most recent PBKDF2 password hash,
to speed up repeated SDR operations.

Landed on NSS-3.48 for Bug 1606992
2020-01-17 09:15:04 +08:00
wolfbeast
ae6de70624 Reinstate the java->c++ source, generator code + documentation.
We've kept the java source up-to-date until its removal, so there should
be very little additional java mangling needed to have it back up to
speed and usable again.

This reverts commit c6446f1126232935c85397aac493113dd38496cd.
2020-01-17 09:15:02 +08:00
wolfbeast
844d2dd0ff Simplify some alias sets in IonMonkey. 2020-01-17 09:15:00 +08:00
Daiki Ueno
75fdf9c3b0 Issue #1338 - Followup: certdb: propagate trust information if trust
module is loaded afterwards,

Summary: When the builtin trust module is loaded after some temp certs
being created, these temp certs are usually not accompanied by trust
information. This causes a problem in UXP as it loads the module from a
separate thread while accessing the network cache which populates temp
certs.

This change makes it properly roll up the trust information, if a temp
cert doesn't have trust information.
2020-01-11 06:47:38 +08:00
wolfbeast
01ba6b6d82 Update GTK clipboard handling
- Store the clipboard even if it was set in a GTK dialog.
- Fix a GtkTargetList leak.
- Notify GTK that the data is no longer available for
clipboard_get_cb(), so that GTK will no longer advertise nor attempt to
store the data.
2020-01-11 06:47:36 +08:00
wolfbeast
b299b34492 Issue #1338 - Un-bust building of NSS after update to 3.48 on Linux. 2020-01-11 06:47:34 +08:00
wolfbeast
d43f7275d0 Handle missing base64 challenge in NegotiateAuth and NTLMAuth. 2020-01-10 17:07:02 +08:00
Yaron Tausky
a0c6c89208 Make copy of list before iterating over it. 2020-01-10 17:07:00 +08:00
wolfbeast
d12d260c6a Be more consistent about decoding IP addresses in PSM. 2020-01-10 17:06:58 +08:00
wolfbeast
647d6033e0 Simplify value setting.
This gets rid of unused boolean return values on setters and a level of
indirection for calls to set values.
2020-01-10 17:06:55 +08:00
Pale Moon
0aef7dd2f9 Block Noveau NV96 mesa driver layers acceleration. 2020-01-09 21:39:07 +08:00
Pale Moon
53db2dd6ea Issue #1703 - Update UA overrides for Google and YouTube 2020-01-09 21:38:49 +08:00
Gaming4JC
8b32eff646 Bug 1379728 part 1. Remove the double-definition of the 'close' event from EventNameList.h.
Tag #1343
2020-01-09 21:28:04 +08:00
Gaming4JC
3b33af54af Bug 1322938 - Make the HTML tree builder aware of <dialog>.
Original patch by :hsivonen
Tag #1343
2020-01-09 21:28:02 +08:00
Gaming4JC
a77fa3ec13 Bug 1322938 - Put <dialog> element behind preference.
Tag #1343
2020-01-09 21:28:00 +08:00
Gaming4JC
7186de99e7 Bug 1322938 - Update <dialog> element Web Platform Tests expected results.
Tag #1343
2020-01-09 21:27:58 +08:00
Gaming4JC
c39b0e4fd7 Bug 1322938 - Emit close event when HTMLDialogElement.prototype.close() is called.
Tag #1343
2020-01-09 21:27:56 +08:00
Gaming4JC
900df90dae Bug 1322938 - Basic implementation of HTMLDialogElement.
Tag #1343
2020-01-09 21:27:54 +08:00
wolfbeast
1803a30a06 Issue #1348 - Part 3: Set IS_PRIVATE input scope in private browsing.
Microsoft IME on Windows 10 20H1 (build 19025+) supports IME private
mode by input scope. Although previous Windows versions use an
undocumented API for Edge and IE only, the next Windows 10 release will
use a public API for it.
We pre-empt this potential privacy concern by setting the IS_PRIVATE
flag when in PB mode.
2020-01-09 21:26:59 +08:00
wolfbeast
9d2e51ac97 Issue #1348 - Part 2: Teach IMEState about Private Browsing mode. 2020-01-09 21:26:56 +08:00
Makoto Kato
8972304bba Issue #1348 - Part 1: Clean up input scope support for IMM32.
Use AutoTArray to set input scope.
2020-01-09 21:26:54 +08:00
wolfbeast
e82bb23d7e Reject sample rates that are out-of-range for libsoundtouch.
We never reach this with our normal use of this lib but adding the
sanity check just in case. (ported from upstream)
2020-01-09 21:23:49 +08:00
JustOff
13aff00f59 Issue #1345 - Implement non-standard legacy CSSStyleSheet rules 2020-01-09 21:23:47 +08:00
wolfbeast
10fff089dc Issue #1338 - Part 3: Update NSS symbols 2020-01-05 11:42:53 +08:00
wolfbeast
c57cac24e8 Issue #1338 - Part 2: Update NSS to 3.48-RTM 2020-01-05 11:42:29 +08:00
wolfbeast
1a92143e68 Issue #1338 - Part 1: Update NSPR to 4.24 2020-01-05 11:38:02 +08:00
Gaming4JC
f6096004d7 Bug 1599054 - allow callers to ommit sending OAuth2 client_secret parameter. 2020-01-03 13:44:14 +08:00
Gaming4JC
1bc5f20c3a Bug 1597933 - Use URLSearchParams for setting params for OAuth2 authorization request. 2020-01-03 13:44:13 +08:00
Gaming4JC
30b347303b Bug 1597933 - use fetch + URLSearchParms instead of Http.jsm to request OAuth2 access token. 2020-01-03 13:44:11 +08:00
Gaming4JC
a8ac34b90b Bug 1597933 - don't pass string constants to determine OAuth refresh token or not. 2020-01-03 13:44:10 +08:00
Gaming4JC
b037870a86 Bug 1597933 - improve OAuth2 params parsing. 2020-01-03 13:44:09 +08:00
Gaming4JC
9839bd3335 Bug 1597933 - clean up OAuth2 code: remove responseType which is always code.
Response type token is part of the OAuth 2.0 Implicit Flow which is not used in Mail Applications, but also discouraged by the OAuth Working Group: https://developer.okta.com/blog/2019/05/01/is-the-oauth-implicit-flow-dead
2020-01-03 13:44:08 +08:00
Gaming4JC
671b973fe5 Bug 1453643 - Enable proper retry on oauth2 authenication failure.
This prevents mail applications from  attempting to use an unauthenticated connection to mailbox(s) and avoid unexpected deletion of local mbox files and subsequent re-download of mailbox content over imap.
2020-01-03 13:44:07 +08:00
Gaming4JC
f7a24c87f6 Bug 1176399 - Multiple requests for master password when GMail OAuth2 is enabled. 2020-01-03 13:44:05 +08:00
wolfbeast
4103320dde Issue #1118 - Part 9: Don't allow reuse of unlinked CallbackObjectHolder. 2019-12-28 21:20:08 +08:00
wolfbeast
f255a1e606 Issue #1118 - Part 8: Remove no-longer-used mDidDocumentOpen and
CheckForOutdatedParent()

This was only used to check for cases when document.open() changed the
global, and elements being inserted into the document needing a new
reflector as a result.

Since document.open() no longer changes the global, this code is no
longer needed.
2019-12-28 21:20:07 +08:00
wolfbeast
10e59cb5a9 Issue #1118 - Part 7: Remove no-longer-used mWillReparent debug code. 2019-12-28 21:20:06 +08:00
wolfbeast
b753aec164 Issue #1118 - Part 6: Fix various tests that are no longer correct.
The behavior change of document.open() requires these tests to be
changed to account for the new spec behavior.
2019-12-28 21:20:04 +08:00
wolfbeast
fc5d61608c Issue #1118 - Part 5: Change the way document.open() works
This changes the work we do for document.open() in the following ways:
- We no longer create a new Window when doing document.open().
  We use the same Window but remove all the event listeners on the
  existing DOM tree and Window before removing the document's existing
  children to provide a clean slate document to use for .write().
- We no longer create a session history entry (previously would be a
  wyciwyg URI). We now replace the current one, effectively losing the
  entry for the original document.
- We now support document.open() on windowless documents.
2019-12-28 21:20:03 +08:00
wolfbeast
ab6131d1e6 Issue #1118 Part 4: Allow UpdateURLAndHistory to work even if there's no
root session history.

This can happen when someone does a document.open() on a document that
has no session history.
2019-12-28 21:19:48 +08:00
wolfbeast
08c26a9adf Issue #1118 Part 3: Add a public RemoveAllListeners method on
EventListenerManager.
2019-12-28 21:19:46 +08:00
wolfbeast
7d52521f95 Issue #1118 Part 2: Allow UpdateURLAndHistory to work even if mOSHE is
null, if we're doing a replace.

We're going to end up hitting this if someone does a document.open()
before mOSHE has been set. We shouldn't need to worry about mLSHE,
because the document.open() will cancel the corresponding load.
2019-12-28 21:19:45 +08:00
wolfbeast
c6694a0c4c Issue #1118 Part 1: Split out part of nsDocShell::AddState into a
separate method.

This implements the "URL and history update steps" from the HTML spec.
See https://html.spec.whatwg.org/multipage/history.html
2019-12-28 21:19:44 +08:00
athenian200
42659b995e Fix typo. 2019-12-28 21:19:27 +08:00