Some classes of a11y are defined from both IUnknown and nsISupports. Both of
them have QueryInterface(). So, one of them is overridden by a sub class
hides the other and this is warned by clang. However, AccessibleWrap.h
prevents this warning since it's by design when the compiler is GCC.
So, it should do same thing when the compiler is clang.
Differential Revision: https://phabricator.services.mozilla.com/D4310
--HG--
extra : moz-landing-system : lando
This introduces the machinery needed to generate crash annotations from a YAML
file. The relevant C++ functions are updated to take a typed enum. JavaScript
calls are unaffected but they will throw if the string argument does not
correspond to one of the known entries in the C++ enum. The existing whitelists
and blacklists of annotations are also generated from the YAML file and all
duplicate code related to them has been consolidated. Once written out to the
.extra file the annotations are converted in string form and are no different
than the existing ones.
All existing annotations have been included in the list (and some obsolete ones
have been removed) and all call sites have been updated including tests where
appropriate.
--HG--
extra : source : 4f6c43f2830701ec5552e08e3f1b06fe6d045860
This again exposes the accelerator (e.g. Ctrl+T for New Tab) for menu items via IAccessible::accName on windows.
Bug 1438193 made Accessible::Name const.
On Windows, XULMenuitemAccessibleWrap::Name is overridden to append the accelerator to the name.
Unfortunately, this override was missed when converting Name to const (and wasn't a compile error because it didn't use the override keyword).
As such, this override stopped being called; we always called the const version, since that's the one inherited into AccessibleWrap.
Making this const causes this override to be called.
MozReview-Commit-ID: 7Gb199mzhHa
--HG--
extra : rebase_source : 41d15042a09e4743d127fb12378c775099aa09a0
We don't fire a state change for the focused state, since that would be redundant given that we fire a focus event.
However, that also means the handler cache previously didn't get invalidated when there was a focus change without any other update.
That resulted in stale focused states, which could mean an object was exposed as focused when it wasn't.
Invalidating the cache for focus events ensures the focused state is always up to date.
MozReview-Commit-ID: DZTJMbiTwgt
--HG--
extra : rebase_source : 9977699999218432b07c4af2149cba6e0de466af
For example, this can happen when choosing File menu -> new Tab.
Focus briefly returns to the document in the original tab, so we ask that document to restore focus.
The remote document then sends a focus event to the parent.
However, before the parent can process that event, focus has already moved to the address bar for the new tab.
With this check, we discover that focus is now in the chrome and thus avoid firing the event for the remote accessible.
MozReview-Commit-ID: 7k58dzREqZD
--HG--
extra : rebase_source : 51303f54293957f562e8540c8bf98f821694be54
For example, this can happen when choosing File menu -> new Tab.
Focus briefly returns to the document in the original tab, so we ask that document to restore focus.
The remote document then sends a focus event to the parent.
However, before the parent can process that event, focus has already moved to the address bar for the new tab.
With this check, we discover that focus is now in the chrome and thus avoid firing the event for the remote accessible.
MozReview-Commit-ID: 7k58dzREqZD
--HG--
extra : rebase_source : 070d3a6b5032bd6d4cd36fb054be04509bb0faae
There is one actual behavior change here, in the webidl version of
TreeBoxObject::GetCellAt. I believe this change fixes a leak of the
nsTreeColumn, but could use careful review.
I tried to avoid changes not needed to get this compiling. There will be a lot
more cleanup in the next few changesets.
This was done automatically replacing:
s/mozilla::Move/std::move/
s/ Move(/ std::move(/
s/(Move(/(std::move(/
Removing the 'using mozilla::Move;' lines.
And then with a few manual fixups, see the bug for the split series..
MozReview-Commit-ID: Jxze3adipUh
Bug 1421144 fixed IAccessible::accFocus to work when focus is within a remote document.
Unfortunately, this causes mysterious intermittent crashes when called from a UIA client in Windows 7.
Ideally, we'd deal with the actual cause of the crashes, but they seem to be deep in Windows RPC code and all attempts at tracking this down have failed.
Clients don't seem to need this too often anyway (and it's a minor annoyance if it doesn't work when they do).
MozReview-Commit-ID: IxvbBGJ2wxA
--HG--
extra : rebase_source : 504909703718b75fffcabf1fbe56c0bbff089b58
Bug 1421018 intended to block the Ask.Com Toolbar (tbnotifier.exe).
This is basically malware and is responsible for a huge number of unnecessary accessibility instantiations.
However, there seems to have been some confusion and we ended up blocking tbnnotifier.exe instead.
This changes that block to tbnotifier.exe.
MozReview-Commit-ID: 2gZF8sYeGtb
--HG--
extra : rebase_source : 3d14a24c12748edfc31ddf7dac51bca491abd744
These events were a hack implemented in bug 703198.
At that time, JAWS required focus events for selection changes in a collapsed combo box.
However, these events also fire for expanded combo boxes.
This is problematic with e10s because now, for an expanded combo box, the real focus events come from the XUL dropdown implemented in the parent process, which is not associated with the document a11y tree in any way.
JAWS seems to cope just fine with value changes for Firefox combo boxes now and VFO have agreed that this is the correct path forward.
MozReview-Commit-ID: Iefop25bFe0
--HG--
extra : rebase_source : a86e5d73d560853bb50e1d8e3bbd11431aba8eb0
I moved the IgnoreErrors decl so it would come after the OOMReporter decl and I
could add the new conversion operator.
MozReview-Commit-ID: B1S6DXmZfvE
To improve performance for cross-process COM, we disable COM garbage collection.
However, this means we never receive Release calls from clients, so defunct accessibles can never be deleted.
Since we know when an accessible is shutting down, we can work around this by forcing COM to disconnect this object from all of its remote clients, which will cause associated references to be released.
MozReview-Commit-ID: 5sIuxnaRJWj
--HG--
extra : rebase_source : e5c54e4f107b2db39d77277020aa939a6a283bd8
extra : histedit_source : 0a964ba20363cc1b8880d78e5d214dae0f758a01
To improve performance for cross-process COM, we disable COM garbage collection.
However, this means we never receive Release calls from clients, so defunct accessibles can never be deleted.
Since we know when an accessible is shutting down, we can work around this by forcing COM to disconnect this object from all of its remote clients, which will cause associated references to be released.
MozReview-Commit-ID: 5sIuxnaRJWj
--HG--
extra : rebase_source : 9059dbed02bf78b4be5b544c8b3bc14f0ca088c8
RealPlayer's use of accessibility was blocked in bug 1418535 due to severe performance problems caused by its use of accessibility.
This is fixed in newer builds, so we want to allow those while still blocking older builds.
MozReview-Commit-ID: 5XlY4IM5qHf
--HG--
extra : rebase_source : b14251b697c5d2a08c613b4a402214f6bd4f475f
1. Move IsModuleVersionLessThan into the Compatibility class and export it in the header file.
2. The function previously referred to the third component of the version as the minor version; i.e. it was testing major.bbbb.minor.dddd.
This is incorrect and might confuse people using this in future code.
The minor version is the second component; i.e. major.minor.cccc.dddd.
cccc and dddd are often named build and revision, but the naming here is less consistent.
3. Rather than accepting separate version components, the function now accepts a single 64 bit value.
This makes comparison easier and also allows for comparison against magic values in other code; e.g. a value meaning "all versions".
This value can be created from separate components using the MAKE_FILE_VERSION macro.
4. Previously, it was assumed that a dll path could not be longer than MAX_PATH, but it can actually be longer.
The function now handles this.
5. The function previously didn't do any error checking, which could have led to null pointer dereferences and possibly other pain.
This was fine when it was only being used for JAWS, which we know always has version info, but this could be problematic for other callers.
We return true if there is a failure, assuming that no version info implies an earlier version.
6. The code now uses smart pointers instead of raw pointers, making memory management simpler.
7. Updated the JAWS version check accordingly.
MozReview-Commit-ID: 9Y6gUQSX0P5
--HG--
extra : rebase_source : 595408140d8611d38fef1211ef41c53e4b65e90c
The change to RootAccessible.cpp fixes an obvious bug introduced in bug 741707.
The visibility changes in gfx/thebes are because NS_DECL_ISUPPORTS has a
trailing "public:" that those classes were relying on to have public
constructors.
MozReview-Commit-ID: IeB8KIJCGhU
We weren't checking for null previously, thus causing a crash when there was no caret.
MozReview-Commit-ID: 6NNRafcVPhb
--HG--
extra : rebase_source : 7d6bfc8c776c1be1429da8c2c94f4bda232c6314
Bug 1422201 changed GetIAccessibleFor so it only handles remote ids when called on the root accessible.
However, this breaks webextension popup documents.
These popups have their own HWND, so the root accessible of that HWND needs to handle accChild for ids in remote documents within that HWND.
Therefore, expand the restriction to cover the root accessible of any HWND, not just the main HWND.
MozReview-Commit-ID: 69v4XSeQLcS
--HG--
extra : rebase_source : bf5c5e0a475424b2276800d1a10ab1a46c136bd6
Previously, in the parent process, we were treating positive child ids as remote unique ids.
This of course failed when searching remote documents and returned early.
Make sure we only treat ids as remote if they are less than 0.
Ids above 0 are child indices and are handled later in the code for both local and remote children.
MozReview-Commit-ID: 2KmFj6rTXTV
--HG--
extra : rebase_source : 273496a3f6420d184f71795095937638e1e3e2ca
Previously, we could return remote accessibles which weren't actually descendants of the accessible on which accChild was called.
For example, calling accChild on a local document with the id of a remote document would happily return the remote accessible.
This confused clients such as NVDA which use accChild to check whether something is a descendant of a document.
MozReview-Commit-ID: 8mJ4m6RC3r2
--HG--
extra : rebase_source : c56040d84e09c2b9ede94985cdd94606c27160a3
For elements such as divs which have no enumerated MSAA role, we return a string.
First, we try the ARIA role, and failing that, we use the tag name.
However, if the author specifies role="", we previously failed.
Instead, we now fall back to the tag name for an empty role string.
That is, we treat the non-existence of the attribute or an empty string value the same way.
Although this is invalid markup, it occurs in the wild, and accRole failing breaks handler caching.
Note that this patch also removes a check for msaaRole != ROLE_SYSTEM_CLIENT when getting a string role.
This check is now pointless because we've already returned earlier if msaaRole != USE_ROLE_STRING.
That is, msaaRole can only be USE_ROLE_STRING at this point.
MozReview-Commit-ID: 7PVvU5V2uO4
--HG--
extra : rebase_source : 7aa1baee31393291ed15e8d6687e6a2d576f858c
The base implementation of accFocus can't handle the case when a remote document has focus and just returns no focus (VT_EMPTY).
Override accFocus on the root accessible to try the accessible for the remote document in the active tab in this case.
This fixes focus loss with NVDA when dismissing the System menu.
MozReview-Commit-ID: 1jhAv08rDFU
--HG--
extra : rebase_source : 7381b397724f21ba894dc94a051996e5d96c642d
As per the spec, if there are no hyperlinks, the hyperlinks array should be set to null and S_FALSE should be returned.
This saves pointless memory management when there are no hyperlinks.
MozReview-Commit-ID: 9wsiXBely6G
--HG--
extra : rebase_source : bc1f6b8a04205939b322393674414365fd89f39a
The Compatibility code previously only tested for OsmHooks.dll to detect Baum Cobra.
However, the 64 bit dll is called OsmHks64.dll.
Update the code to test for both so that detection works in 64 bit processes.
Aside from telemetry, this is also needed to ensure that the InSendMessageEx hook is used to work around RPC_E_CANTCALLOUT_ININPUTSYNCCALL.
MozReview-Commit-ID: FjliZybHrZH
--HG--
extra : rebase_source : a4b53abf86206dd8ecb4965a6ce1ba31545f0197
This allows for fetching of all hyperlinks instead of one at a time, which improves performance for a cross-process client.
MozReview-Commit-ID: 8wso3EqBqwP
--HG--
extra : rebase_source : f972076e5b65a8c882f9a07a82b67eeefc85d8a4
Bug 1407475 added support for accNavigate(NAVRELATION_EMBEDS) for e10s.
However, it's possible for a client to call this on the root accessible for a window which was since closed.
Therefore, we must check whether the root accessible is defunct before trying to retrieve the tab document.
MozReview-Commit-ID: 9iR6Kvzu5Mb
--HG--
extra : rebase_source : 82afdecf915cd09cc3eaed948587b87d8ac4173b
When we only have a single process, this can be (and was previously) handled the same way as any other relation.
However, for multi process, the normal relation mechanism doesn't work because it can't handle remote objects.
This patch overrides accNavigate for the root accessible to handle this, since this is only ever used on the root.
MozReview-Commit-ID: JLm5zITfG6Y
--HG--
extra : rebase_source : 3666ffe699d861c06b763200e7d59fbd75a581ee
(Path is actually r=froydnj.)
Bug 1400459 devirtualized nsIAtom so that it is no longer a subclass of
nsISupports. This means that nsAtom is now a better name for it than nsIAtom.
MozReview-Commit-ID: 91U22X2NydP
--HG--
rename : xpcom/ds/nsIAtom.h => xpcom/ds/nsAtom.h
extra : rebase_source : ac3e904a21b8b48e74534fff964f1623ee937c67
The first accessibility client to ask for the root accessible will get a surrogate LazyInstantiator object. That instantiates accessibility if appropriate and then passes calls through.
However, it previously didn't implement accNavigate, which broke things for clients that used this on the root.
In particular, this was bad for NAVRELATION_EMBEDS, which is used by some clients to retrieve the tab document.
This patch simply passes accNavigate through to the real accessible, just as is already done for other methods.
MozReview-Commit-ID: KIoA9BwMDhc
--HG--
extra : rebase_source : 7514a0a6fb3750df128ad1e790e1fd1978c031e8
