What happens here if aListHead is an empty selector list:
(1) next selector group is parsed and continues to the next iteration if it's empty or invalid
(2) if we're a forgiving selector list and aListHead is empty, replace it with the selector group that we've just parsed
(3) step 1 ignores invalid/empty, so we assert that step 2 should never have an empty selector list
This should allow pseudo-classes that accept selector lists to properly handle complex selectors. `:visited` is also matched inside selector lists.
A previous iteration of this commit had incorrect matching behavior when called from `RestrictedSelectorListMatches`.
This makes it clear that we're accessing a selector list (nsCSSSelectorList), which is different from the selectors (nsCSSSelector) contained inside a selector list. Previously, both were confusingly referred to as mSelectors.
This modifies selector list parsing to accommodate being "forgiving". Aliases for the :is selector's former names were also included. Note that the older and prefixed variant `-moz-any` remains unforgiving.
This is for those cases where the hostnme of the email address is not
the actual hostname of the e-mail service (e.g. GMail with a custom
domain).
This should resolve issue athenian200/epyrus#53.
Signed-off-by: Mavridis Philippe <mavridisf@gmail.com>
Editor changes caret visibility during drag and drop. But when destroying
editor, we don't restore caret state. So we should restore it when destroying
editor.
Co-authored-by: Makoto Kato <m_kato@ga2.so-net.ne.jp>
This implements the "clip" attribute value for the "crop" attribute found in XUL label elements. It functions similarly to having the text-overflow property set to an empty string or "". This shouldn't break previously established behavior as it only adds a new value to be checked.
The behavior of the "none" attribute value in the documentation (https://udn.realityripple.com/docs/Archive/Mozilla/XUL/Attribute/crop) is incorrect. The "none" attribute value literally means "don't crop anything" if you'd check the code and is also equivalent to not setting the "crop" attribute at all. This has always been the case since Firefox 2 (earliest version I checked) and apparently, this section of the documentation was copied verbatim from XULPlanet without even checking if it's correct.
because crash still happens. crash log:
RegExpShared::finalize(FreeOp* fop)
{
for (auto& comp : compilationArray)
js_free(comp.byteCode);
for (size_t i = 0; i < tables.length(); i++)
js_free(tables[i]);
tables.~JitCodeTables(); // <----
}
Stack trace:
mozglue.dll!arena_dalloc(void * ptr, unsigned int offset) Line 4209 C
mozglue.dll!je_free(void * ptr) Line 6099 C
> xul.dll!js::RegExpShared::finalize(js::FreeOp * fop) Line 982 C++
xul.dll!js::gc::Arena::finalize<js::RegExpShared>(js::FreeOp * fop, js::gc::AllocKind thingKind, unsigned int thingSize) Line 460 C++
xul.dll!FinalizeTypedArenas<js::RegExpShared>(js::FreeOp * fop, js::gc::Arena * * src, js::gc::SortedArenaList & dest, js::gc::AllocKind thingKind, js::SliceBudget & budget, js::gc::ArenaLists::KeepArenasEnum keepArenas) Line 521 C++
xul.dll!js::gc::ArenaLists::backgroundFinalize(js::FreeOp * fop, js::gc::Arena * listHead, js::gc::Arena * * empty) Line 2597 C++
xul.dll!js::gc::GCRuntime::sweepBackgroundThings(js::gc::ZoneList & zones, js::LifoAlloc & freeBlocks) Line 2982 C++
xul.dll!js::GCHelperState::doSweep(js::AutoLockGC & lock) Line 3183 C++
xul.dll!js::GCHelperState::work() Line 3103 C++
xul.dll!js::HelperThread::handleGCHelperWorkload(js::AutoLockHelperThreadState & locked) Line 1823 C++
xul.dll!js::HelperThread::threadLoop() Line 1888 C++
xul.dll!js::detail::ThreadTrampoline<void (__cdecl&)(void *),js::HelperThread *>::Start(void * aPack) Line 227 C++
[External Code]
[Frames below may be incorrect and/or missing, no symbols loaded for ucrtbase.dll]
