nsXPCWrappedJSClass now only adds indirections and dynamic
allocations, so eliminate it. The class is kept around as a collection
of static helper functions to avoid needing to move around all of this
code and messing with history. In total, these patches save around 2kb
of dynamic allocations per process.
The major parts of this patch are:
* Dropping indirection for accessing things on nsXPTInterfaceInfo and
renaming things from class to info.
* Removing the stuff related to instances of nsXPCWrappedJSClass
existing (ctor, dtor, field, ISupports implementation, getter methods).
* Removing IID2WrappedJSClassMap, because we only need the map from IIDs
to info.
I dropped the null check in TraverseNative because mInfo is never
cleared, while mClass was.
I dropped the forward declaration of nsXPCWrappedJSClass because no
instances of that class exist, so no function will take or return a
pointer to one.
Differential Revision: https://phabricator.services.mozilla.com/D26218
--HG--
extra : moz-landing-system : lando
Ultimately, this method is really about dumping XPConnect-ish
information about an nsXPTInterfaceInfo, so change it into a static
method that takes an info directly. This loses logging of the
nsXPCWrappedJSClass's ref count, but that is going away in the next
patch anyways.
Differential Revision: https://phabricator.services.mozilla.com/D26217
--HG--
extra : moz-landing-system : lando
This interface serves no real purpose, aside from some weird XPConnect
debugging function. If you are in a debugger already, you can just
call the nsXPCWrappedJSClass DebugDump method directly.
For now, I changed nsXPCWrappedJSClass to just implement nsISupports,
but this will go away later.
I also devirtualized DebugDump(), because it is no longer an XPCOM
method.
Differential Revision: https://phabricator.services.mozilla.com/D26216
--HG--
extra : moz-landing-system : lando
The first idea here is that |this| is actually the GetClass() of the
|wrapper| argument (the one call site looks like
"GetClass()->CallMethod(this, ...)"), so we can locally reconstruct it
when CallMethod is a static method.
The second idea here is that the only real use of the
nsXPCWrappedJSClass is to grab some data from the nsXPTInterfaceInfo
in a few places. This means that we can take a pointer to the info
early on in the function and use that rather than go through the
nsXPCWrappedJSClass. This in turn means that because the info is
statically allocated we no longer need to do a kungFuDeathGrip on the
wrapper's nsXPCWrappedJSClass.
Differential Revision: https://phabricator.services.mozilla.com/D26215
--HG--
extra : moz-landing-system : lando
There are a number of nsXPCWrappedJSClass methods that don't use any
data from |this|, so go ahead and make them static. This is one step
towards eliminating nsXPCWrappedJSClass entirely.
In addition, devirtualize a few methods, because they are going to
have to get devirtualized anyways, and there's no need for them to be
virtual.
Differential Revision: https://phabricator.services.mozilla.com/D26214
--HG--
extra : moz-landing-system : lando
This field now just caches the IsReflectable() field from the method
info, so get rid of it.
Differential Revision: https://phabricator.services.mozilla.com/D26071
--HG--
extra : moz-landing-system : lando
XPCConvert::IsMethodReflectable() is derived entirely from
nsXPTMethodInfo, so we can compute it at build time and include it in
nsXPTMethodInfo. It is the only use of mNotXPCOM and mHidden, so we
can get rid of those fields at the same time.
This paves the way for getting rid of XPCWrappedJSClass::mDescriptors
in the next patch.
Differential Revision: https://phabricator.services.mozilla.com/D26070
--HG--
extra : moz-landing-system : lando
There is only a single XPC JS runtime now, so there's no need to keep a
special pointer around.
Differential Revision: https://phabricator.services.mozilla.com/D26068
--HG--
extra : moz-landing-system : lando
The change to the second FindTearOff call in XPCWrappedNative::GetNewOrUsed is
fixing a longstanding bug that was introduced in bug 903891 when the sense of
the check was incorrectly reversed. Luckily that code is unreached in
practice, because the two PreCreate hooks we have left never create the wrapper.
Differential Revision: https://phabricator.services.mozilla.com/D26007
--HG--
extra : moz-landing-system : lando
This field now just caches the IsReflectable() field from the method
info, so get rid of it.
Differential Revision: https://phabricator.services.mozilla.com/D26071
--HG--
extra : moz-landing-system : lando
XPCConvert::IsMethodReflectable() is derived entirely from
nsXPTMethodInfo, so we can compute it at build time and include it in
nsXPTMethodInfo. It is the only use of mNotXPCOM and mHidden, so we
can get rid of those fields at the same time.
This paves the way for getting rid of XPCWrappedJSClass::mDescriptors
in the next patch.
Differential Revision: https://phabricator.services.mozilla.com/D26070
--HG--
extra : moz-landing-system : lando
There is only a single XPC JS runtime now, so there's no need to keep a
special pointer around.
Differential Revision: https://phabricator.services.mozilla.com/D26068
--HG--
extra : moz-landing-system : lando
1. Adding a new attribute chromeContext in ConsoleEvent
2. Adding a new boolean attribute isFromChromeContext in nsIConsoleMessage
3. Sending IsFromChromeContext to the parent process
Differential Revision: https://phabricator.services.mozilla.com/D23330
--HG--
extra : moz-landing-system : lando
Left shifts exhibit undefined behavior if the sign bit changes, which
would happen in this code for indices that are 31 mod 32. Fix this by
always making sure the shifted value is an unsigned integer.
Differential Revision: https://phabricator.services.mozilla.com/D25476
--HG--
extra : moz-landing-system : lando
This ensures the JS shell and browser behave the same way and it's nice for fuzzing.
Differential Revision: https://phabricator.services.mozilla.com/D25204
--HG--
extra : moz-landing-system : lando
Also adds a javascript.options.ion.full.threshold browser pref and similar shell
flags.
This doesn't rename the existing prefs yet.
Differential Revision: https://phabricator.services.mozilla.com/D24156
--HG--
extra : moz-landing-system : lando
In bug 256180, the size of the stack on 64-bit Windows was changed from 2MB to
8MB, and on 32-bit Windows, from 1MB to 1.5MB. This is so large that it takes
significantly longer for a runaway recursive function to throw "too much
recursion", which causes terrible performance in scripts obfuscated using
obfuscator.io.
This patch leaves the actual stack size as-is, but changes the
JS-engine-specific stack quota back to 2MB on 64-bit Windows (6MB if ASAN is
enabled). 32-bit Windows is unaffected by the new cap.
Differential Revision: https://phabricator.services.mozilla.com/D24597
--HG--
extra : moz-landing-system : lando
1. Adding a new attribute chromeContext in ConsoleEvent
2. Adding a new boolean attribute isFromChromeContext in nsIConsoleMessage
3. Sending IsFromChromeContext to the parent process
Differential Revision: https://phabricator.services.mozilla.com/D23330
--HG--
extra : moz-landing-system : lando
Replaced instances of callers in both C++ and JS files to query the state from the principal directly.
Differential Revision: https://phabricator.services.mozilla.com/D22532
--HG--
extra : moz-landing-system : lando
Noticed by Ian Moody [:Kwan] while watching test output. (!)
The test was introduced in bug 515475, 10 years ago. However, the syntax used
in this patch seems to have been removed from SpiderMonkey around that time,
maybe even before the patch landed; that is, this is not among the syntax
removed in bug 517580. As far as I can tell, this test has never functioned
properly.
Differential Revision: https://phabricator.services.mozilla.com/D24236
--HG--
extra : moz-landing-system : lando
This creates a shell command-line option, `--enable-experimental-fields`, and a
Gecko pref, `javascript.options.experimental.fields`.
Both are off by default everywhere, for now.
Differential Revision: https://phabricator.services.mozilla.com/D22045
--HG--
extra : moz-landing-system : lando
This creates a shell command-line option, `--enable-experimental-fields`, and a
Gecko pref, `javascript.options.experimental.fields`.
Both are off by default everywhere, for now.
Differential Revision: https://phabricator.services.mozilla.com/D22045
--HG--
extra : moz-landing-system : lando
`CompartmentPrivate::GetScope()` was added so callers don't have to do `scope.get()`
manually. The `scope` field is now private and was renamed to `mScope`.
Also replaces some `CompartmentPrivate::Get(obj)->scope` instances with
`ObjectScope(obj)`. It's equivalent but shorter.
Differential Revision: https://phabricator.services.mozilla.com/D22664
--HG--
extra : moz-landing-system : lando
XPCWrappedNativeScope is now allocated and destroyed with the CompartmentPrivate
that owns it. In follow-up bugs we could merge the two classes (see bug 1032928).
This also removes the dying-scopes list. XPCJSRuntime now stores the list of all
scopes as mozilla::LinkedList.
Differential Revision: https://phabricator.services.mozilla.com/D22492
--HG--
extra : moz-landing-system : lando
Renamed the test to reflect that it is really just a test of the script preloader
as well. I just moved it to get it close to the ScriptPreloader and near existing
tests.
Differential Revision: https://phabricator.services.mozilla.com/D22330
--HG--
rename : testing/marionette/harness/marionette_harness/tests/unit/test_startup_caches.py => js/xpconnect/tests/marionette/test_preloader_telemetry.py
extra : moz-landing-system : lando
Hiding document.createEvent("TouchEvent"), document.createTouch, document.createTouchList and ontouch* event
handlers on desktop to follow what Chrome has done.
This patch explicitly does not remove createTouch or createTouchList everywhere, although those seem to have been
removing already on some other browsers.
Devtools use TOUCHEVENTS_OVERRIDE_ENABLED for touch event testing, and this patch keeps the old behavior per discussion
with devtools devs.
Differential Revision: https://phabricator.services.mozilla.com/D22081
--HG--
extra : rebase_source : 562588a289632ba2f11db7f3ac8782c26c3b05f8
In bug 1264235 we have some indication that observed bugs with the
startup cache might have been resolved, but we don't really know
until we collect data. Collecting these stats will give us the
ability to have more certainty that the startup cache is functioning
correctly in the wild.
Differential Revision: https://phabricator.services.mozilla.com/D19573
--HG--
extra : moz-landing-system : lando
Add an RAII guarded initialization for the IO interposer to the
initialization process for xpc shell tests. This ensures that whenever
an xpcshell tests uses the IOInterposer, that it will correctly
catch all registered threads, and will not miss any.
Differential Revision: https://phabricator.services.mozilla.com/D20736
--HG--
extra : moz-landing-system : lando
The main behavior changes are:
1) We no longer create a new Window when doing document.open(). We use the
same Window but remove all the event listeners on it and on the existing DOM
tree before removing the document's existing kids.
2) We no longer create a new session history entry. The existing one always
gets replaced instead.
3) We now support document.open on documents that are not in a Window.
The reasons for the various test changes are as follows:
The change to browser_modifiedclick_inherit_principal.js is because we no
longer set the docshell to a wyciwyg URL when document.open() happens and the
test was depending on that to terminate.
browser_wyciwyg_urlbarCopying.js is being removed because it's trying to test
wyciwyg URIs, which no longer exist.
The changes in docshell/test/navigation are because document.open() no longer
affects session history. One of the tests was testing the interactions there
and is being removed; another is being repurposed to just test that
document.open() does not affect history.length.
The change to test_x-frame-options.html is because document.open() now removes
event listeners on the window, which it didn't use to do (and in the specific
case in this test reused the existing inner too, so the listener was still
around in practice). The new behavior matches other browsers.
The removal of test_bug172261.html is because document.open() no longer affects
session history, so you can't go back across it or forward to the "opened"
state, so the situation that test is trying to test no longer exists.
The changes to test_bug255820.html are because reloading a document after
document.open() will now just load the URL of the document that was the entry
document for the open() call, not reload the written content. So there's not
much point testing reload behavior, and in this test it was just reloading the
toplevel test file inside the frames.
The change to test_bug346659.html is because now we no longer create a new
Window on document.open().
The change to test_bug1232829.html is because document.open() (implicit in this
test) no longer adds history entries, so the back() was just leaving the test
page instead of going back across the document.open(). The test is a
crashtest in practice, so might still be testing something useful about how
document.open() interacts with animations.
The change to test_bug715739.html is because the URL of the document after
document.open() is now the URL of the entry document, not a wyciwyg URL, so
reload() has different behavior than it used to.
The change to test_bug329869.html is because now when we go back we're
reloading the original document we had, not doing a wyciwyg load, and the
security info now doesn't include the untrusted script.
The changes to the wpt expectations are removing a bunch of expected failures
now that we pass those tests and disabling some tests that are fundamentally
racy and hence fail randomly. The latter all have github issues filed for the
test problem.
The change to testing/web-platform/tests/common/object-association.js is fixing
tests that were not matching the spec (and were failing in other browsers).
The change to parser-uses-registry-of-owner-document.html is fixing tests that
were not matching the spec (and were failing in other browsers).
The change to document-write.tentative.html is because the test was buggy: it
was using the same iframe element for all its tests and racing loads from some
tests against API calls from other tests, etc. It's a wonder it ever managed
to pass, independent of these patches (and in fact it doesn't pass according to
wpt.fyi data, even in Firefox).
The changes in html/browsers/history/the-history-interface are because
document.open() no longer adds history entries. The test was failing in all
other browsers for the same reason.
The changes in html/browsers/history/the-location-interface are because
reloading a document.open()-created thing now loads the URL of the page that
was the entry document for the open() call. The test was failing in all other
browsers.
The change to reload_document_open_write.html is because we now reload the url
of the document that entered the script that called open() when we reload, not
the written content. Other browsers were failing this test too; Gecko with
the old document.open implementation was the only one that passed.
The change to http-refresh.py is to fix a test bug: it was not returning a
Content-Type header, so we were putting up helper app dialogs, etc.
The change to test_ext_contentscript.js is because we no create a new global
for document.open() calls. Kris Maglione OKed this part.
Differential Revision: https://phabricator.services.mozilla.com/D17323
--HG--
extra : moz-landing-system : lando
The watchdog thread's stack is intentionally kept as small as possible, since
the work that it does is trivial and should not require much stack space.
One result of this is that the stack segment is generally too small to
instantiate a new service (and we really don't want to be instantiating
services on it anyway).
The debug service is generally instantiated before the watchdog thread tries
to touch it, but this isn't guaranteed (and, in particular, it often isn't for
xpcshell runs on some noopt debug builds). Instantiating it before starting
the watchdog thread solves this problem.
Differential Revision: https://phabricator.services.mozilla.com/D20869
--HG--
extra : rebase_source : 2f9c3e412a7ad9a0b6f84c61907e4a7508f3b18c
BasicDllServices is used to gain access to the authenticode APIs in non-Gecko
contexts. One feature that WinDllServices provides is the ability to register
a callback interface to be notified when a DLL has been loaded.
This is not particularly useful in the BasicDllServices use case, and in the
"handle a launcher process failure on a background thread" use case, would
actually be harmful.
This patch modifies the DLLServices backend to offer a "basic" option that
omits the callback stuff.
Differential Revision: https://phabricator.services.mozilla.com/D19696
--HG--
extra : moz-landing-system : lando
Consequently, this removes:
- MOZ_LIBPRIO, which is now always enabled.
- non_msvc_compiler, which is now always true.
- The cl.py wrapper, since it's not used anymore.
- CL_INCLUDES_PREFIX, which was only used for the cl.py wrapper.
- NONASCII, which was only there to ensure CL_INCLUDES_PREFIX still
worked in non-ASCII cases.
This however keeps a large part of detecting and configuring for MSVC,
because we still do need it for at least headers, libraries, and midl.
Depends on D19614
Differential Revision: https://phabricator.services.mozilla.com/D19615
--HG--
extra : moz-landing-system : lando
Bug 1526508 - add profiler markers for importing a JS module, loading a JS XPCOM component or a subscript
Differential Revision: https://phabricator.services.mozilla.com/D19228
--HG--
extra : moz-landing-system : lando
This is a follow-up to the previous part, which actually changes one of
these callers to use Array<nsIIDRef> instead of [array] nsIIDPtr.
From doing this patch, it seems like we should consider changing
the type `nsIIDRef` to instead simply be `nsIID`, and treat it more like
the `AString` types from the POV of XPIDL. `nsIIDPtr` would then
continue to exist for backwards compatibility, but we can probably
remove almost all current consumers over time.
Depends on D19175
Differential Revision: https://phabricator.services.mozilla.com/D19176
--HG--
extra : moz-landing-system : lando
Currently the [ref] and [ptr] types share the same underlying
implementation. This is unfortunate, and can cause correctness problems
with outparam refs (as an example).
By using the same tools used to allow other larger objects (such as
jsid, nsTArray, and nsString) to be stored directly in the nsXPTCVariant
object, this patch directly stores the nsID in the nsXPTCVariant object
when calling from JS into C++.
Using this new strategy avoids an nsID* allocation every time we pass
one over XPConnect, and should also allow us to simplify callers.
In addition, some special casing is added to xpidl to make it possible
to use the nsid reference type objects directly inside of Array<T>,
which will allow us to remove `[array] nsIIDPtr` callers.
Differential Revision: https://phabricator.services.mozilla.com/D19175
--HG--
extra : moz-landing-system : lando
