PLDHashTable::Search() does not modify any members. So, this method and
methods called by it should be marked as const.
MozReview-Commit-ID: 6g4jrYK1j9E
--HG--
extra : rebase_source : eda6c50c538fec0e8c09cb2ba629735eea6ec711
This also replaces the custom logic in ObserverList with an nsTObserverArray
which has all the necessary logic for stable iteration over a potentially
changing list of items. Unused dependencies were also removed.
--HG--
extra : source : 303478f7f248470a1c747f42dad9cb85c3129f0a
This also replaces the custom logic in ObserverList with an nsTObserverArray
which has all the necessary logic for stable iteration over a potentially
changing list of items. Unused dependencies were also removed.
--HG--
extra : source : 1a83516d2ee5939052c5fb226b81563a0d114ff9
This patch is an automatic replacement of s/NS_NOTREACHED/MOZ_ASSERT_UNREACHABLE/. Reindenting long lines and whitespace fixups follow in patch 6b.
MozReview-Commit-ID: 5UQVHElSpCr
--HG--
extra : rebase_source : 4c1b2fc32b269342f07639266b64941e2270e9c4
extra : source : 907543f6eae716f23a6de52b1ffb1c82908d158a
This reduces memory usage because we only need one allocation instead of two
for the dynamic atom and its chars, and because we don't need to store a
refcount and a size. It precludes sharing of chars between dynamic atoms, but
we weren't benefiting much from that anyway.
This reduces per-process memory usage by up to several hundred KiB on my
Linux64 box.
One consequence of this change is that we need to allocate + copy in
DOMString::SetKnownLiveAtom(), which could make some things slower.
--HG--
extra : rebase_source : ba4065ea31e509dd985c003614199f73def0596c
Cache AccessibleNode and make it able to operate the same instance by nsINode::GetAccessibleNode
--HG--
extra : rebase_source : 063eec8658af020f5408260d7d581ee76a04bd37
* Create BLOCKQUOTE internal role (also needed for correct exposure of
blockquote element on ATK)
* Add new ARIA roles to internal ARIA map
* Add new roles to existing ARIA roles mochitest and update blockquote
element test to reflect mapping to the new internal role
* Create BLOCKQUOTE internal role (also needed for correct exposure of
blockquote element on ATK)
* Add new ARIA roles to internal ARIA map
* Add new roles to existing ARIA roles mochitest and update blockquote
element test to reflect mapping to the new internal role
Same approach as the other bug, mostly replacing automatically by removing
'using mozilla::Forward;' and then:
s/mozilla::Forward/std::forward/
s/Forward</std::forward</
The only file that required manual fixup was TestTreeTraversal.cpp, which had
a class called TestNodeForward with template parameters :)
MozReview-Commit-ID: A88qFG5AccP
The DoMouseClick helper is also removed because no other caller can now pass a null aEvent. Other MouseClicked implementations are also updated since aEvent cannot be null, which was already the case.
MozReview-Commit-ID: 3bTJ6cZW9ZA
--HG--
extra : rebase_source : ae1bafe7144f6f428e2ef4e7047d5c64e0a19e8c
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
The schema handling for this is currently a bit ugly, for the sake of
simplifying uplift. In the figure, we should find a way to change the schema
pattern matching based on whether or not the extension is privileged.
MozReview-Commit-ID: CU9WR2Ika6k
--HG--
extra : rebase_source : 28e1c43cf0f0efc01b40757e4e65c4ac5d882258
These were found using some ugly text searches, so it's possible some unused
atoms remain. In the future, we should enforce removing unused atoms using
static analysis. Or just generate the static atoms table based on string atom
names in our code.
This patch leaves unused RDF atoms in place, since those are being dealt with
in another bug.
MozReview-Commit-ID: 1KpH9KsHzQy
--HG--
extra : rebase_source : 8138faa2b16e847da31861abae2bbc1c7bac4e02
Move tracking of persistent window state into nsXULWindow. Also, move
special handling of the width/height of the window into nsXULWindow.
MozReview-Commit-ID: LOmHGyYeNSU
--HG--
extra : rebase_source : bcea16eb6209ff789948644a64968a7325cea4ef
This also removes any redundant Ci.nsISupports elements in the interface
lists.
This was done using the following script:
acecb401b7/processors/chromeutils-generateQI.jsm
MozReview-Commit-ID: AIx10P8GpZY
--HG--
extra : rebase_source : a29c07530586dc18ba040f19215475ac20fcfb3b
For reasons unknown, if you give MSVC:
// Foo.h
struct Foo
{
...
};
extern const Foo gFoo;
// Foo.cpp, which necessarily includes Foo.h.
extern constexpr Foo gFoo = {
};
MSVC will create a static initializer for gFoo and place it in the
read/write data section, rather than the read-only data section.
Removing the `extern const` declaration seems to be enough to make this
problem go away. We need to adjust the declaration of other variables
to compensate for the non-visibility of gFoo in the header file.
Due to the decision to keep the old API on nsXPTInterfaceInfo in part 4, this is
a fairly straightforward patch.
1. I had to change a couple of consumers of `IsRetval()` due to the movement of
that flag.
2. I changed all code which held a nsIInterfaceInfo to hold an `const
nsXPTInterfaceInfo*` instead.
3. I changed code which used the nsIInterfaceInfoManager to instead call the
static methods on nsXPTInterfaceInfo.
This removes properties of XULElement that can easily seen to be unused, even if the attributes they control are still in use. There are other properties that may still be used once or twice, and they are not removed here.
MozReview-Commit-ID: IL6mCvtGQAG
--HG--
extra : rebase_source : 4b22b330d311ef22e3466f517c04d5a19512ab71
Each nsStaticAtomSetup contains a pointer to a static atom, and also a pointer
to the canonical pointer to that static atom. Which is pretty weird! The
notable thing thing about it is that these structs are in an array, and that
gives us the only way to iterate over all static atoms in a single class, for
registration and lookups.
But thanks to various other recent changes to the implementation of static
atoms, we can now put the static atoms themselves into an array, which can be
iterated over. So this patch does that. With that done, nsStaticAtomSetup is no
longer necessary.
According to the `size` utility, on Linux64 this reduces the size of libxul.so
by the following amounts:
> text: 62008 bytes
> data: 20992 bytes
> bss: 21040 bytes
> total: 104040 bytes
- The bss reduction is one word per atom, because the canonical static atom
pointers (e.g. nsGkAtoms::foo) have moved from .bss to .data, because they're
now initialized at compile time instead of runtime.
- The data reduction is one word per atom, because we remove two words per atom
for the nsStaticAtomSetup removal, but gain one word per atom from the
previous bullet point.
- I'm not sure about the text reduction. It's three words per atom. Maybe
because there is one less relocation per atom?
Other notable things in the patch:
- nsICSSAnonBoxPseudo and nsICSSPseudoElement now inherit from nsStaticAtom,
not nsAtom, because that's more precise.
- Each static atoms array now has an enum associated with it, which is used in
various ways.
- In the big comment about the macros at the top of nsStaticAtom.h, the pre-
and post-expansion forms are now shown interleaved. The interleaving reduces
duplication and makes the comment much easier to read and maintain. The
comment also has an introduction that explains the constraints and goals of
the implementation.
- The SUBCLASS macro variations are gone. There are few enough users of these
macros now that always passing the atom type has become simpler.
MozReview-Commit-ID: 1GmfKidLjaU
--HG--
extra : rebase_source : 2352590101fc6693ba388f885ca4714a42963943
This will allow us to work with direct pointers to static atoms that are
initialized at compile-time, rather than pointers to pointers to static atoms
that are initialized at runtime.
MozReview-Commit-ID: K04pEicuqu3
--HG--
extra : rebase_source : c5d895b34567c621f4591769390fc95a4c6ff405
In SourceSurfaceImage::GetTextureClient, we use LookupForAdd. This is
because we typically will create a new TextureClient if there isn't
already one created. This creation can fail because the size is too big,
or we don't have the memory available for it. Unfortunately LookupForAdd
is an infallible operation; it is expected we will always add something
to the hashtable if we don't find an entry. This patch adds an OrRemove
method to cover the corner case where we are unable to complete the
insertion.
The last piece of code that was using these attributes was removed in
https://hg.mozilla.org/mozilla-central/rev/35b7fa5ebd58
MozReview-Commit-ID: FyF7kzlpqMz
--HG--
extra : rebase_source : edaf64a4be92be1f72ff67f90c03eed09173cce3
Currently static atoms are stored on the heap, but their char buffers are
stored in read-only static memory.
This patch changes the representation of nsStaticAtom (thus making it a
non-trivial subclass of nsAtom). Instead of a pointer to the string, it now has
an mStringOffset field which is a 32-bit offset to the string. (This requires
placement of the string and the atom within the same object so that the offset
is known to be small. The docs and macros in nsStaticAtom.h handle that.)
Static and dynamic atoms now store their chars in different ways: nsStaticAtom
stores them inline, nsDynamicAtom has a pointer to separate storage. So
`mString` and GetStringBuffer() move from nsAtom to nsDynamicAtom.
The change to static atoms means they can be made constexpr and stored in
read-only memory instead of on the heap. On 64-bit this reduces the per-process
overhead by 16 bytes; on 32-bit the saving is 12 bytes. (Further reductions
will be possible in follow-up patches.)
The increased use of constexpr required multiple workarounds for MSVC.
- Multiple uses of MOZ_{PUSH,POP}_DISABLE_INTEGRAL_CONSTANT_OVERFLOW_WARNING to
disable warnings about (well-defined!) overflow of unsigned integer
arithmetic.
- The use of -Zc:externConstexpr on all files defining static atoms, to make
MSVC follow the C++ standard(!) and let constexpr variables have external
linkage.
- The use of -constexpr:steps300000 to increase the number of operations
allowed in a constexpr value, in order to handle gGkAtoms, which requires
hashing ~2,500 atom strings.
The patch also changes how HTML5 atoms are handled. They are now treated as
dynamic atoms, i.e. we have "dynamic normal" atoms and "dynamic HTML5 atoms",
and "dynamic atoms" covers both cases, and both are represented via
nsDynamicAtom. The main difference between the two kinds is that dynamic HTML5
atoms still aren't allowed to be used in various operations, most notably
AddRef()/Release(). All this also required moving nsDynamicAtom into the header
file.
There is a slight performance cost to all these changes: now that nsStaticAtom
and nsDynamicAtom store their chars in different ways, a conditional branch is
required in the following functions: Equals(), GetUTF16String(),
WeakAtom::as_slice().
Finally, in about:memory the "explicit/atoms/static/atom-objects" value is no
longer needed, because that memory is static instead of heap-allocated.
MozReview-Commit-ID: 4AxPv05ngZy
These functions no longer perform any refcounting, so the existing names are
misleading.
MozReview-Commit-ID: LX55e0bUP8N
--HG--
extra : rebase_source : 89a3da577325286c1d31723acfd4153754f49703
This function captures a common usage pattern, generalizing the existing
IsMember() function.
MozReview-Commit-ID: 5Pt7kqyGD6Y
--HG--
extra : rebase_source : b03e4cc8f5a4a25da9236420f4b64493664b70e0
It seems silly to have a tiny utils class with a single function in its own
module. This patch moves it into nsStaticAtom.h/nsAtomTable.cpp. It also
renames nsAtomListUtils as nsStaticAtomUtils. Finally, it uses templates to
remove the need for the `aCount` parameter at callsites.
MozReview-Commit-ID: DvJVoZFv89c
--HG--
extra : rebase_source : 1f1dd27d56e46c71c30c10102ac6132a721e23d1
Most of the noise is from the fact that clang-format on parser/html/*.{h,cpp}
reformatted all sorts of stuff. Not running it caused lots of format changes
from the generator... I guess we changed the format rules since the last time
this got run?
MozReview-Commit-ID: IA2G87zUIKN
In SourceSurfaceImage::GetTextureClient, we use LookupForAdd. This is
because we typically will create a new TextureClient if there isn't
already one created. This creation can fail because the size is too big,
or we don't have the memory available for it. Unfortunately LookupForAdd
is an infallible operation; it is expected we will always add something
to the hashtable if we don't find an entry. This patch adds an OrRemove
method to cover the corner case where we are unable to complete the
insertion.
Now that nsGkAtoms is in xpcom/, we can call nsGkAtoms::AddRefAtoms() from
NS_InitAtomTable(), which removes the need for DefaultAtoms, and also removes a
duplicate static atom.
MozReview-Commit-ID: CyfvnvZomzZ
--HG--
extra : rebase_source : 53ead62323a340038c1b4594b1a3eb225aa19626
This list of atoms isn't particularly DOM-ish, and having it in xpcom/ will
help with the next patch.
MozReview-Commit-ID: 1Y3Fhn9lNbh
--HG--
rename : dom/base/nsGkAtomList.h => xpcom/ds/nsGkAtomList.h
rename : dom/base/nsGkAtoms.cpp => xpcom/ds/nsGkAtoms.cpp
rename : dom/base/nsGkAtoms.h => xpcom/ds/nsGkAtoms.h
extra : rebase_source : 0a0f6ab4432e0d58ea4662299b750a8c52325ad5
Without this patch, it's impossible for clients to move a reference into an
nsInterfaceHashtable. That causes at least one extra addref/release pair when
they otherwise could. With this patch, a client can do `hashTable.Put(key,
comptr.forget());` to avoid the additional refcounting.
MozReview-Commit-ID: Ghm7n41ziZp
--HG--
extra : rebase_source : a3e842bf9dfe202134c58e447ecf4fa79851c076
Because (a) that name better indicates that it's a pointer to a pointer, and
(b) because nsStaticAtom::mString is going to be renamed as mAtom in bug
1411469.
MozReview-Commit-ID: D5tuNOstMgr
--HG--
extra : rebase_source : 9344eeea0288c8c52c069ce21e8bc55f6e0f3f6f
By removing the "Atom" suffix, which is redundant.
MozReview-Commit-ID: 4MCX9Icfjrw
--HG--
extra : rebase_source : c3c759a508a8938b59d36dbb20448d2964b98c91
The patch also uses GetStringBuffer() in a couple of appropriate places.
MozReview-Commit-ID: JufCUgmO8JL
--HG--
extra : rebase_source : ecd3f17b5560b19622c86759d605fa122d70e48a
The refcount is only used for dynamic atoms.
On 64-bit, this reduces sizeof(nsStaticAtom) from 24 bytes to 16 bytes, and the
on-heap size from 32 bytes to 16 bytes. This saves 42 KiB per process.
On 32-bit, this reduces sizeof(nsStaticAtom) from 16 bytes to 12 bytes, but the
on-heap size stays at 16 bytes, so memory usage is unchanged.
MozReview-Commit-ID: 7d9H7MRHN9a
--HG--
extra : rebase_source : d3eafb3aaf44a592afd6c89cd0519d043056e65a
Various atom-related things have improved recently.
- The main atom table is now threadsafe (bug 1275755) and so can be accessed on
any thread. It has also been split into pieces (bug 1440824), which greatly
reduces lock contention.
- A cache has been added to the HTML5 parser (bug 1352874) that removes the
need for most of the full table lookups.
As a result, there is no point having a separate static atom table. This patch
removes it.
MozReview-Commit-ID: 8ou1BrnPAwd
--HG--
extra : rebase_source : 0c6ab073b1a20b703705582d28731a68456741e1
The old output had a single value: "atoms-table". The new output looks like
this:
> 649,904 B (00.39%) -- atoms
> ├──350,256 B (00.21%) -- dynamic
> │ ├──235,056 B (00.14%) ── unshared-buffers
> │ └──115,200 B (00.07%) ── atom-objects
> ├──212,992 B (00.13%) ── table
> └───86,656 B (00.05%) ── static/atom-objects
MozReview-Commit-ID: 924vUmxHAlh
--HG--
extra : rebase_source : 6c977546a69eeee62ebc87e335982e8278217484
This converts from the odd `tmp` for `rv` pattern and just returns immediately
on failure instead.
--HG--
extra : rebase_source : 1ad5882c1411e9e10f99201b6233ed87c71a20cc
This patch was autogenerated by my decomponents.py
It covers almost every file with the extension js, jsm, html, py,
xhtml, or xul.
It removes blank lines after removed lines, when the removed lines are
preceded by either blank lines or the start of a new block. The "start
of a new block" is defined fairly hackily: either the line starts with
//, ends with */, ends with {, <![CDATA[, """ or '''. The first two
cover comments, the third one covers JS, the fourth covers JS embedded
in XUL, and the final two cover JS embedded in Python. This also
applies if the removed line was the first line of the file.
It covers the pattern matching cases like "var {classes: Cc,
interfaces: Ci, utils: Cu, results: Cr} = Components;". It'll remove
the entire thing if they are all either Ci, Cr, Cc or Cu, or it will
remove the appropriate ones and leave the residue behind. If there's
only one behind, then it will turn it into a normal, non-pattern
matching variable definition. (For instance, "const { classes: Cc,
Constructor: CC, interfaces: Ci, utils: Cu } = Components" becomes
"const CC = Components.Constructor".)
MozReview-Commit-ID: DeSHcClQ7cG
--HG--
extra : rebase_source : d9c41878036c1ef7766ef5e91a7005025bc1d72b
This was done using the following script:
37e3803c7a/processors/chromeutils-import.jsm
MozReview-Commit-ID: 1Nc3XDu0wGl
--HG--
extra : source : 12fc4dee861c812fd2bd032c63ef17af61800c70
extra : intermediate-source : 34c999fa006bffe8705cf50c54708aa21a962e62
extra : histedit_source : b2be2c5e5d226e6c347312456a6ae339c1e634b0
This removes an unnecessary level of indirection by replacing all
nsStringGlue.h instances with just nsString.h.
--HG--
extra : rebase_source : 340989240af4018f3ebfd92826ae11b0cb46d019
When this was added, the xpcom glue was still a thing, and there was a
distinction between things that would build with mozalloc available and
others. There is no such distinction anymore. Anything that has access
to xpcom has access to infallible allocator functions.
--HG--
extra : rebase_source : 04bce114e940c53709275d0354ea7240df4a051e
Now that nsArray uses nsCOMArray under the hood, we don't have to do weird
ForwardEnumeration hacks to start IndexOf at a non-zero index.
MozReview-Commit-ID: 3ReDV0BT0hn
--HG--
extra : rebase_source : 399a988d00fabaa314eb7592b2aacf277a2ca477
Now that xpcom strings use templates we can combine their `ArenaStrdup`
implementations. `nsTStringRepr` is used so as to allow handling of both
`nsTLiteralString` and `nsTSubstring` types.
--HG--
extra : rebase_source : e77891da589f320507b45f524a3203b3dc9f38e6
extra : histedit_source : 79d3b9a1add191563e0985f0c0e416bd29f24351
This adds support for duplicating both raw `char*` strings and raw `char16_t*`
strings by making the character type generic.
--HG--
extra : rebase_source : bb5245ed71161c8c785684e5a56ac894f03874ea
extra : histedit_source : e94eb738a3982f0cb63a894a0cdfbdf0be2b9cad
It's a sub-class of nsAtom, useful for cases where you know you are dealing
exclusively with static atoms. The nice thing about it is that you can use
raw nsStaticAtom pointers instead of RefPtr<>. (In fact, the AddRef/Release
implementations ensure that we'll crash if we use RefPtr<nsStaticAtom>.)
MozReview-Commit-ID: 4Q6QHX5h44V
--HG--
extra : rebase_source : e4237f85b4821b684db0ef84d1f9c5e17cdee428
It's easy to mess up the scoping so that (a) the label is pushed and then
immediately popped, and/or (b) the string doesn't live long enough. It's also
easy to do a utf16-to-utf8 conversion unnecessarily when the profiler is
inactive. This patch splits that macro into three new ones that are harder to
mess up.
- AUTO_PROFILER_LABEL_DYNAMIC_CSTR: same as current.
- AUTO_PROFILER_LABEL_DYNAMIC_NSCSTRING: for nsCStrings.
- AUTO_PROFILER_LABEL_DYNAMIC_LOSSY_NSSTRING: for nsStrings.
--HG--
extra : rebase_source : 3e2bbec4737b696e1c86579ae54be4cb3186c100
There are four things that must be provided for every static atom, two of which
have a macro:
- the atom pointer declaration (no macro);
- the atom pointer definition (no macro);
- the atom char buffer (NS_STATIC_ATOM_BUFFER);
- the StaticAtomSetup struct (NS_STATIC_ATOM_SETUP).
This patch introduces new macros for the first two things: NS_STATIC_ATOM_DECL
and NS_STATIC_ATOM_DEFN, and changes the arguments of the existing two macros
to make them easier to use (e.g. all the '##' concatenation now happens within
the macros).
One consequence of the change is that all static atoms must be within a class,
so the patch adds a couple of classes where necessary (DefaultAtoms, TSAtoms).
The patch also adds a big comment explaining how the macros are used, and what
their expansion looks like. This makes it a lot easier to understand how static
atoms work. Correspondingly, the patch removes some small comments scattered
around the macro use points.
MozReview-Commit-ID: wpRyrEOTHE
--HG--
extra : rebase_source : 9f85d477b4d06c9a9e710c757de1f1476edb6efe
Because it's the type we use to set up static atoms at startup, not the static
atom itself.
The patch accordingly renames some parameters, variables, and NS_STATIC_ATOM,
for consistency.
MozReview-Commit-ID: 1a0KvhYNNw2
--HG--
extra : rebase_source : 5c66e5b2dfe053a368bf3584d957198aec4cce91
These were generated with |./mach eslint --fix xpcom| with the
.eslintignore and xpcom/tests/unit/.eslintrc.js changes from the next
patch.
MozReview-Commit-ID: 8pKkICSK3JQ
--HG--
extra : rebase_source : bbc98050928f27160d8ca63d38aa0c383be95878
This change requires introducing nsID::Clone(). Because it's infallible, the
patch also removes some redundant failure-handling code. (nsMemory::Clone() is
also infallible, so this code was redundant even before this change.)
--HG--
extra : rebase_source : ef22757d3fa814320490bf7e19e822b8f0c4bdc3
The new code is slightly less efficient because it requires measuring the
string length, but these strings are all short so it shouldn't matter.
Note that the case in DataToString() is a little different. The std::min() that
was there appears to be excessive caution -- this code is always printf'ing
some kind of number, so 32 chars should never be reached -- but it was bogus
anyway, because if 32 was exceeded then (a) we would have overflowed `buf`, and
(b) we'd be returning a non-null-terminated string.
--HG--
extra : rebase_source : b666ad72c09d8c32b98bb9abc9dce1bd0c912c9b
They are equivalent -- both infallible, both requiring measuring the length of
the string -- but moz_xstrdup is much more readable. (One place deals with
16-bit strings and so uses NS_strdup instead, which is also infallible.)
The patch also removes some failure-path code that will never execute due to
the infallibility.
--HG--
extra : rebase_source : 115574cf55db90b60e6bee59e5dc6ee409374159
Currently nsAtom::mString points to the interior of an nsStringBuffer. For
static atoms this requires the use of nsFakeStringBuffer, which is pretty
gross.
This patch changes things so that nsAtom::mString points to a static char
buffer for static atoms. This simplifies a number of things:
- nsFakeStringBuffer and CheckStaticAtomSizes are no longer needed.
- FakeBufferRefCountHelper is no longer needed.
- nsAtom's constructor for static atoms is simpler.
- RegisterStaticAtoms() is simpler.
On the flip-side, a couple of things get more complicated.
- nsAtom::ToString() treats static and dynamic atoms differently.
- nsAtom::GetStringBuffer() is now only valid for dynamic atoms. This
function is only used in two places, both involving DOMString, so those
locations are updated appropriately. This also requires updating some other
code assigning nsStrings to DOMStrings, because we can't assume that
nsStrings are shared.
On Linux64 this change reduces the size of the binary by 8752 B, and moves
81968 B from the .data to the .rodata section, where it can be shared between
processes.
--HG--
extra : rebase_source : 0f6fcdec1c525aa66222e208b66a9f9026f69bcb
Bug 1134923 removed the use of those functions in gecko, and left some
for the XPCOM standalone glue. The XPCOM standalone glue was severely
stripped down in bug 1306327, with the effect of removing the
implementation for those functions.
The remains in nsXPCOM.h are:
XPCOM_API(void*) NS_Alloc(size_t aSize);
XPCOM_API(void*) NS_Realloc(void* aPtr, size_t aSize);
XPCOM_API(void) NS_Free(void* aPtr);
With no implementation left, the first arm is never actually used, and
the second arm means every remaining use of those functions in the tree
is a macro expansion to one of moz_xmalloc, moz_xrealloc or free.
--HG--
extra : rebase_source : fd1669abc5a25d8edbd5c3a8522e22a5c3f558e2
It's easy to mess up the scoping so that (a) the label is pushed and then
immediately popped, and/or (b) the string doesn't live long enough. It's also
easy to do a utf16-to-utf8 conversion unnecessarily when the profiler is
inactive.
This patch splits that macro into three new ones that are harder to mess up.
- AUTO_PROFILER_LABEL_DYNAMIC_CSTR: same as current.
- AUTO_PROFILER_LABEL_DYNAMIC_NSCSTRING: for nsCStrings.
- AUTO_PROFILER_LABEL_DYNAMIC_LOSSY_NSSTRING: for nsStrings.
--HG--
extra : rebase_source : 53c8b43b6a1be06d00618a133e28bf95c46a3ba3
It's easy to mess up the scoping so that (a) the label is pushed and then
immediately popped, and/or (b) the string doesn't live long enough. It's also
easy to do a utf16-to-utf8 conversion unnecessarily when the profiler is
inactive.
This patch splits that macro into three new ones that are harder to mess up.
- AUTO_PROFILER_LABEL_DYNAMIC_CSTR: same as current.
- AUTO_PROFILER_LABEL_DYNAMIC_NSCSTRING: for nsCStrings.
- AUTO_PROFILER_LABEL_DYNAMIC_LOSSY_NSSTRING: for nsStrings.
--HG--
extra : rebase_source : 59f77df0124249bfd11fee3585420a17b4201d37
(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
This fixes the build in non-unified mode, as nsCOMArray::Sort() method uses
NS_QuickSort().
MozReview-Commit-ID: CDBLIX8D3mL
--HG--
extra : rebase_source : 67902cf10b9311adbd97fecf7906762bb86c52e4
This patch merges nsAtom into nsIAtom. For the moment, both names can be used
interchangeably due to a typedef. The patch also devirtualizes nsIAtom, by
making it not inherit from nsISupports, removing NS_DECL_NSIATOM, and dropping
the use of NS_IMETHOD_. It also removes nsIAtom's IIDs.
These changes trigger knock-on changes throughout the codebase, changing the
types of lots of things as follows.
- nsCOMPtr<nsIAtom> --> RefPtr<nsIAtom>
- nsCOMArray<nsIAtom> --> nsTArray<RefPtr<nsIAtom>>
- Count() --> Length()
- ObjectAt() --> ElementAt()
- AppendObject() --> AppendElement()
- RemoveObjectAt() --> RemoveElementAt()
- ns*Hashtable<nsISupportsHashKey, ...> -->
ns*Hashtable<nsRefPtrHashKey<nsIAtom>, ...>
- nsInterfaceHashtable<T, nsIAtom> --> nsRefPtrHashtable<T, nsIAtom>
- This requires adding a Get() method to nsRefPtrHashtable that it lacks but
nsInterfaceHashtable has.
- nsCOMPtr<nsIMutableArray> --> nsTArray<RefPtr<nsIAtom>>
- nsArrayBase::Create() --> nsTArray()
- GetLength() --> Length()
- do_QueryElementAt() --> operator[]
The patch also has some changes to Rust code that manipulates nsIAtom.
MozReview-Commit-ID: DykOl8aEnUJ
--HG--
extra : rebase_source : 254404e318e94b4c93ec8d4081ff0f0fda8aa7d1
nsHtml5Atoms are very similar to dynamic nsAtoms. This patch removes the former
in favour of the latter, which leaves nsAtom as the only subclass of nsIAtom.
nsAtom::mKind is still used to distinguish dynamic atoms from HTML5 atoms, and
the HTML5 parser still uses manual memory management to handle its HTML5 atoms.
nsHtml5AtomEntry::mAtom had to be changed from an nsAutoPtr to a raw pointer
because nsAtom's destructor is private.
MozReview-Commit-ID: 1pBzwkog3ut
--HG--
extra : rebase_source : fbb819e527cb30606348da9ce3eede62e00fb936
This patch moves nsAtom's declaration to nsIAtom.h. In order to keep most of
nsAtom's members private the patch also does the following.
- It introduces a new class, nsAtomFriend, which encapsulates the functions
that will need access to nsAtom's private members.
- It moves GCKind, GCAtomTable(), and GCAtomTableLocked() out of nsAtom.
- It removes the factory methods, and replaces their uses with direct
constructor calls.
MozReview-Commit-ID: L8vfrHsR2cS
--HG--
extra : rebase_source : d38a65515ae82c355966ca80ab235bea11b638af
Because it's going to be exposed via nsIAtom.h in the next few patches.
MozReview-Commit-ID: A81s1nWrvmB
--HG--
extra : rebase_source : 43990feb460fc7343c994f72fb6c6679a3aeb1cc
This patch reduces sizeof(PLDHashTable) as follows.
- 64-bit: from 40 bytes to 32
- 32-bit: from 28 bytes to 20
It does this by doing the following.
- It moves mGeneration from EntryStore to PLDHashTable, to avoid unnecessary
padding on 64-bit. This requires tweaking EntryStore::Set() as explained in a
comment.
- It also shrinks mGeneration from uint32_t to uint16_t, saving 2 bytes of
data.
- It shrinks mEntrySize from uint32_t to uint8_t, to cut 3 bytes of data.
- It shrinks mHashShift from int16_t to uint8_t, trimming another byte of data,
and moves it, saving another 2 bytes of padding.
And it reorders the fields so the word-sized ones are at the start, which makes
it easier to imagine the memory layout.
The patch also adds a test, and fixes some misordered function arguments in
existing tests.
--HG--
extra : rebase_source : 6ed6f7be68477fd4a82f07dd2f51c1f1d9b92dcc
The current code replaces one PLDHashTable's mGeneration with another. If the
two generation values happen to be equal, it will look as though the storage
hasn't changed when really it has.
The fix is to use EntryStore::Set() to update mEntryStore, which increments
mGeneration appropriately.
--HG--
extra : rebase_source : d1779a143746c8d1a3e67bc3685e703496206b0f
Its return value is never used, and most implementations return nullptr anyway.
MozReview-Commit-ID: 8rxC053mmE8
--HG--
extra : rebase_source : 61a0b8b1373396182efd27d3c01b96e5e5541364
Now that nsIAtom is non-scriptable, a .idl file isn't needed.
I made the new nsIAtom.h file by starting with a generated nsIAtom.h file, and
then cleaning it up and removing some stuff that wasn't necessary.
--HG--
extra : rebase_source : 9655fd38984512bd96cf5555048f7774414f6d92
There's no reason for them to be separate, and we can use the |kind| field to
distinguish the two kinds when necessary.
This lets us remove the duplication of ScriptableToString(), ToUTF8String(),
and ScriptableEquals().
It also lets us use |Atom*| pointers instead of |nsIAtom*| pointers in various
places within nsAtomTable.cpp, which de-virtualizes various calls and removes
the need for some static_casts.
--HG--
extra : rebase_source : 2f9183323446e353f8cc5dcedf57d9dc9a38f0a7
This imports Chromium's `make_dafsa.py` script [1]. It takes in a gperf
formatted file (note: gperf is *not* required) and converts that to a compact
binary representation of the string data in the form of a deterministic
acyclic finite state automaton (DAFSA) [2].
The only change made to the script was to make it handle the arguments our
file generation script passes in to the `main` function.
It also imports the logic for traversing the DAFSA [3] almost verbatim in
`Dafsa.cpp`. A thin wrapper was added so that we can reuse the DAFSA structure
for multiple tables.
The only change made to the original logic was to swap in mozilla style
assertions and rename the not found constant from `kNotFound` to
`Dafsa::kKeyNotFound` in order to avoid a collision with `kNotFound` defined in
our nsString code.
[1] 6ba04a9056/tools/dafsa/make_dafsa.py
[2] https://en.wikipedia.org/wiki/Deterministic_acyclic_finite_state_automaton
[3] a2a90a35aa/net/base/registry_controlled_domains/registry_controlled_domain.cc (72)
MozReview-Commit-ID: Eion9POHZm5
This imports Chromium's `make_dafsa.py` script [1]. It takes in a gperf
formatted file (note: gperf is *not* required) and converts that to a compact
binary representation of the string data in the form of a deterministic
acyclic finite state automaton (DAFSA) [2].
The only change made to the script was to make it handle the arguments our
file generation script passes in to the `main` function.
It also imports the logic for traversing the DAFSA [3] almost verbatim in
`Dafsa.cpp`. A thin wrapper was added so that we can reuse the DAFSA structure
for multiple tables.
The only change made to the original logic was to swap in mozilla style
assertions and rename the not found constant from `kNotFound` to
`Dafsa::kKeyNotFound` in order to avoid a collision with `kNotFound` defined in
our nsString code.
[1] 6ba04a9056/tools/dafsa/make_dafsa.py
[2] https://en.wikipedia.org/wiki/Deterministic_acyclic_finite_state_automaton
[3] a2a90a35aa/net/base/registry_controlled_domains/registry_controlled_domain.cc (72)
MozReview-Commit-ID: Eion9POHZm5
This imports Chromium's `make_dafsa.py` script [1]. It takes in a gperf
formatted file (note: gperf is *not* required) and converts that to a compact
binary representation of the string data in the form of a deterministic
acyclic finite state automaton (DAFSA) [2].
The only change made to the script was to make it handle the arguments our
file generation script passes in to the `main` function.
It also imports the logic for traversing the DAFSA [3] almost verbatim in
`Dafsa.cpp`. A thin wrapper was added so that we can reuse the DAFSA structure
for multiple tables.
The only change made to the original logic was to swap in mozilla style
assertions and rename the not found constant from `kNotFound` to
`Dafsa::kKeyNotFound` in order to avoid a collision with `kNotFound` defined in
our nsString code.
[1] 6ba04a9056/tools/dafsa/make_dafsa.py
[2] https://en.wikipedia.org/wiki/Deterministic_acyclic_finite_state_automaton
[3] a2a90a35aa/net/base/registry_controlled_domains/registry_controlled_domain.cc (72)
MozReview-Commit-ID: Eion9POHZm5
nsTArray::AppendElement{,s} uses an IncrementLength method to adjust the
length of the array after appending elements. There are checks in
IncrementLength to ensure that we're not incrementing the length stored
in the static empty header object; these checks are necessary in cases
such as appending a zero-length array to another zero-length array.
But we do not need this check when we're calling AppendElement: we know
the header is obviously not the empty header, because we increased the
length of the array by one for the newly appended element. Incrementing
the length can therefore be inlined, saving ~90K of codesize on x86-64
Linux.
All the SizeOf{In,Ex}cludingThis() functions take a MallocSizeOf function
which measures memory blocks. This patch introduces a new type, SizeOfState,
which includes a MallocSizeOf function *and* a table of already-measured
pointers, called SeenPtrs. This gives us a general mechanism to measure
graph-like data structures, by recording which nodes have already been
measured. (This approach is used in a number of existing reporters, but not in
a uniform fashion.)
The patch also converts the window memory reporting to use SizeOfState in a lot
of places, all the way through to the measurement of Elements. This is a
precursor for bug 1383977 which will measure Stylo elements, which involve
Arcs.
The patch also converts the existing mAlreadyMeasuredOrphanTrees table in the
OrphanReporter to use the new mechanism.
--HG--
extra : rebase_source : 2c23285f8b6c3b667560a9d14014efc4633aed51
The simplistic shift-based hashing function creates a lot of collisions
for pointers pointing to arrays as it doesn't do a great job at distributing
the data randomly based on the input bytes.
Running with XPCOM_MEM_LOG_CLASSES=nsStringBuffer triggers an
assertion in refcount logging for nsFakeStringBuffers. These are given
an initial refcount of 1, without calling NS_LOG_ADDREF. Then,
AddRef() is called on these objects in StaticAtom::StaticAtom(), and
we tell the refcount logging system about the fake buffer, and that it
has a refcount of 0, triggering the assertion.
The first part of the fix is to call NS_LOG_ADDREF for this initial
refcount, in StaticAtom().
This first fix causes refcount logging to start reporting that the
fake string buffers leak, when XPCOM_MEM_LOG_CLASSES is not set. This
is because refcount logging is now getting told about these objects
being AddRefed at 1, which it takes to mean that an object is created.
To work around this issue, I add an array gFakeBuffers that contains
every fake string buffer we create, and tell the refcount logging
system that these objects are all being destroyed, when the atom table
is being shut down. This could result in some bogosity if the fake
buffers are "leaked" but hopefully this is still an improvement over
the current state.
MozReview-Commit-ID: 5AxoBYAlYRU
--HG--
extra : rebase_source : ba0763cb494894918141774025db525cea9f9c75
Running with XPCOM_MEM_LOG_CLASSES=nsStringBuffer triggers an
assertion in refcount logging for nsFakeStringBuffers. These are given
an initial refcount of 1, without calling NS_LOG_ADDREF. Then,
AddRef() is called on these objects in StaticAtom::StaticAtom(), and
we tell the refcount logging system about the fake buffer, and that it
has a refcount of 0, triggering the assertion.
The first part of the fix is to call NS_LOG_ADDREF for this initial
refcount, in StaticAtom().
This first fix causes refcount logging to start reporting that the
fake string buffers leak, when XPCOM_MEM_LOG_CLASSES is not set. This
is because refcount logging is now getting told about these objects
being AddRefed at 1, which it takes to mean that an object is created.
To work around this issue, I add an array gFakeBuffers that contains
every fake string buffer we create, and tell the refcount logging
system that these objects are all being destroyed, when the atom table
is being shut down. This could result in some bogosity if the fake
buffers are "leaked" but hopefully this is still an improvement over
the current state.
MozReview-Commit-ID: 5AxoBYAlYRU
--HG--
extra : rebase_source : 967520e825d26c369a11acc478d223d190137a43
This patch makes the following changes to the macros.
- Removes PROFILER_LABEL_FUNC. It's only suitable for use in functions outside
classes, due to PROFILER_FUNCTION_NAME not getting class names, and it was
mostly misused.
- Removes PROFILER_FUNCTION_NAME. It's no longer used, and __func__ is
universally available now anyway.
- Combines the first two string literal arguments of PROFILER_LABEL and
PROFILER_LABEL_DYNAMIC into a single argument. There was no good reason for
them to be separate, and it forced a '::' in the label, which isn't always
appropriate. Also, the meaning of the "name_space" argument was interpreted
in an interesting variety of ways.
- Adds an "AUTO_" prefix to PROFILER_LABEL and PROFILER_LABEL_DYNAMIC, to make
it clearer they construct RAII objects rather than just being function calls.
(I myself have screwed up the scoping because of this in the past.)
- Fills in the 'js::ProfileEntry::Category::' qualifier within the macro, so
the caller doesn't need to. This makes a *lot* more of the uses fit onto a
single line.
The patch also makes the following changes to the macro uses (beyond those
required by the changes described above).
- Fixes a bunch of labels that had gotten out of sync with the name of the
class and/or function that encloses them.
- Removes a useless PROFILER_LABEL use within a trivial scope in
EventStateManager::DispatchMouseOrPointerEvent(). It clearly wasn't serving
any useful purpose. It also serves as extra evidence that the AUTO_ prefix is
a good idea.
- Tweaks DecodePool::SyncRunIf{Preferred,Possible} so that the labelling is
done within them, instead of at their callsites, because that's a more
standard way of doing things.
--HG--
extra : rebase_source : 318d1bc6fc1425a94aacbf489dd46e4f83211de4
All the instances are converted as follows.
- nsSubstring --> nsAString
- nsCSubstring --> nsACString
--HG--
extra : rebase_source : cfd2238c52e3cb4d13e3bd5ddb80ba6584ab6d91
All the instances are converted as follows.
- nsAFlatString --> nsString
- nsAFlatCString --> nsCString
--HG--
extra : rebase_source : b37350642c58a85a08363df2e7c610873faa6e41
Also change nsClassHashtable::LookupOrAdd to not regard existing entries with
a nullptr value as non-existent. This is to make it consistent with
nsBaseHashtable::LookupForAdd() and other methods.
MozReview-Commit-ID: 1wYqK8XQbyW
We already have a handle to the entry in the hashtable, we can remove
that directly, rather than implicitly looking up the entry to remove in
Remove().
PLDHashTable takes the result of the hash function and multiplies it by
kGoldenRatio to ensure that it has a good distribution of bits across
the 32-bit hash value, and then zeroes out the low bit so that it can be
used for the collision flag. This result is called hash0. From hash0
it computes two different numbers used to find entries in the table
storage: hash1 is used to find an initial position in the table to
begin searching for an entry; hash2 is then used to repeatedly offset
that position (mod the size of the table) to build a chain of positions
to search.
In a table with capacity 2^c entries, hash1 is simply the upper c bits
of hash0. This patch does not change this.
Prior to this patch, hash2 was the c bits below hash1, padded at the low
end with zeroes when c > 16. (Note that bug 927705, changeset
1a02bec165e16f370cace3da21bb2b377a0a7242, increased the maximum capacity
from 2^23 to 2^26 since 2^23 was sometimes insufficient!) This manner
of computing hash2 is problematic because it increases the risk of long
chains for very large tables, since there is less variation in the hash2
result due to the zero padding.
So this patch changes the hash2 computation by using the low bits of
hash0 instead of shifting it around, thus avoiding 0 bits in parts of
the hash2 value that are significant.
Note that this changes what hash2 is in all cases except when the table
capacity is exactly 2^16, so it does change our hashing characteristics.
For tables with capacity less than 2^16, it should be using a different
second hash, but with the same amount of random-ish data. For tables
with capacity greater than 2^16, it should be using more random-ish
data.
Note that this patch depends on the patch for bug 1353458 in order to
avoid causing test failures.
MozReview-Commit-ID: JvnxAMBY711
--HG--
extra : transplant_source : 2%D2%C2%CE%E1%92%C8%F8H%D7%15%A4%86%5B%3Ac%0B%08%3DA
PLDHashTable's entry store has two types of unoccupied entries: free
entries and removed entries. The search of a chain of entries
(determined by the hash value) in the entry store to search for an entry
can stop at free entries, but it continues across removed entries,
because removed entries are entries that may have been skipped over when
we were adding the value we're searching for to the hash, but have since
been removed. For live entries, we also maintain this distinction by
using one bit of storage for a collision flag, which notes that if the
hashtable entry is removed, its place in the entry store must become a
removed entry rather than a free entry.
When we add a new entry to the table, Add's semantics require that we
return an existing entry if there is one, and only create a new entry if
no existing entry exists. (Bug 1352198 suggests the possibility of a
faster alternative Add API where the caller guarantees that the key is
not already in the hashtable.) When we search for the existing entry,
we must thus continue the search across removed entries, even though we
record the first removed entry found to return if the search for an
existing entry fails.
The existing code adds the collision flag through the entire table
search during an Add. This patch changes that behavior so that we only
add the collision flag prior to finding the first removed entry. Adding
it after we find the first removed entry is unnecessary, since we are
not making that entry part of a path to a new entry. If it is part of a
path to an existing entry, it will already have the collision flag set.
This patch effectively puts an if (!firstRemoved) around the else branch
of the if (MOZ_UNLIKELY(EntryIsRemoved(entry))), and then refactors that
condition outwards since it is now around the contents of both the if
and else branches.
MozReview-Commit-ID: CsXnMYttHVy
--HG--
extra : transplant_source : %80%9E%83%EC%CCY%B4%B0%86%86%18%99%B6U%21o%5D%29%AD%04
The OrInsert() method adds the new entry to the hashtable if needed, and
returns the newly added entry or the pre-existing one. It allows for a
more concise syntax at the call site.
PLDHashTable takes the result of the hash function and multiplies it by
kGoldenRatio to ensure that it has a good distribution of bits across
the 32-bit hash value, and then zeroes out the low bit so that it can be
used for the collision flag. This result is called hash0. From hash0
it computes two different numbers used to find entries in the table
storage: hash1 is used to find an initial position in the table to
begin searching for an entry; hash2 is then used to repeatedly offset
that position (mod the size of the table) to build a chain of positions
to search.
In a table with capacity 2^c entries, hash1 is simply the upper c bits
of hash0. This patch does not change this.
Prior to this patch, hash2 was the c bits below hash1, padded at the low
end with zeroes when c > 16. (Note that bug 927705, changeset
1a02bec165e16f370cace3da21bb2b377a0a7242, increased the maximum capacity
from 2^23 to 2^26 since 2^23 was sometimes insufficient!) This manner
of computing hash2 is problematic because it increases the risk of long
chains for very large tables, since there is less variation in the hash2
result due to the zero padding.
So this patch changes the hash2 computation by using the low bits of
hash0 instead of shifting it around, thus avoiding 0 bits in parts of
the hash2 value that are significant.
Note that this changes what hash2 is in all cases except when the table
capacity is exactly 2^16, so it does change our hashing characteristics.
For tables with capacity less than 2^16, it should be using a different
second hash, but with the same amount of random-ish data. For tables
with capacity greater than 2^16, it should be using more random-ish
data.
Note that this patch depends on the patch for bug 1353458 in order to
avoid causing test failures.
MozReview-Commit-ID: JvnxAMBY711
--HG--
extra : rebase_source : dbde93b9312dd10fb3a549ee4098b57e1df5cadf
This removes the last uses of PR_smprintf from the tree (excluding the
security and nsprpub directories). It also fixes a related latent bug
in nsAppRunner.cpp (which was incorrectly freeing the pointer passed to
PR_SetEnv).
MozReview-Commit-ID: GynP2PhuWWO
--HG--
extra : rebase_source : c3b83c7bd08b1c222e137a00323caf5481352845
PLDHashTable's entry store has two types of unoccupied entries: free
entries and removed entries. The search of a chain of entries
(determined by the hash value) in the entry store to search for an entry
can stop at free entries, but it continues across removed entries,
because removed entries are entries that may have been skipped over when
we were adding the value we're searching for to the hash, but have since
been removed. For live entries, we also maintain this distinction by
using one bit of storage for a collision flag, which notes that if the
hashtable entry is removed, its place in the entry store must become a
removed entry rather than a free entry.
When we add a new entry to the table, Add's semantics require that we
return an existing entry if there is one, and only create a new entry if
no existing entry exists. (Bug 1352198 suggests the possibility of a
faster alternative Add API where the caller guarantees that the key is
not already in the hashtable.) When we search for the existing entry,
we must thus continue the search across removed entries, even though we
record the first removed entry found to return if the search for an
existing entry fails.
The existing code adds the collision flag through the entire table
search during an Add. This patch changes that behavior so that we only
add the collision flag prior to finding the first removed entry. Adding
it after we find the first removed entry is unnecessary, since we are
not making that entry part of a path to a new entry. If it is part of a
path to an existing entry, it will already have the collision flag set.
This patch effectively puts an if (!firstRemoved) around the else branch
of the if (MOZ_UNLIKELY(EntryIsRemoved(entry))), and then refactors that
condition outwards since it is now around the contents of both the if
and else branches.
MozReview-Commit-ID: CsXnMYttHVy
--HG--
extra : transplant_source : W4%B8%BA%D5p%102%1B%8D%83%23%E0s%B3%B0f%0D%05%AE
We need this because the stored values in the hash table may themselves
be main-thread only objects, and destroying them off the main thread
will cause crashes.
PLDHashTable takes the result of the hash function and multiplies it by
kGoldenRatio to ensure that it has a good distribution of bits across
the 32-bit hash value, and then zeroes out the low bit so that it can be
used for the collision flag. This result is called hash0. From hash0
it computes two different numbers used to find entries in the table
storage: hash1 is used to find an initial position in the table to
begin searching for an entry; hash2 is then used to repeatedly offset
that position (mod the size of the table) to build a chain of positions
to search.
In a table with capacity 2^c entries, hash1 is simply the upper c bits
of hash0. This patch does not change this.
Prior to this patch, hash2 was the c bits below hash1, padded at the low
end with zeroes when c > 16. (Note that bug 927705, changeset
1a02bec165e16f370cace3da21bb2b377a0a7242, increased the maximum capacity
from 2^23 to 2^26 since 2^23 was sometimes insufficient!) This manner
of computing hash2 is problematic because it increases the risk of long
chains for very large tables, since there is less variation in the hash2
result due to the zero padding.
So this patch changes the hash2 computation by using the low bits of
hash0 instead of shifting it around, thus avoiding 0 bits in parts of
the hash2 value that are significant.
Note that this changes what hash2 is in all cases except when the table
capacity is exactly 2^16, so it does change our hashing characteristics.
For tables with capacity less than 2^16, it should be using a different
second hash, but with the same amount of random-ish data. For tables
with capacity greater than 2^16, it should be using more random-ish
data.
MozReview-Commit-ID: JvnxAMBY711
--HG--
extra : transplant_source : %8A%25%FB%E3H%B8_%F1G%F6%3E%0B%29%DF%20%FF%D8%E1%AEw
PLDHashTable's entry store has two types of unoccupied entries: free
entries and removed entries. The search of a chain of entries
(determined by the hash value) in the entry store to search for an entry
can stop at free entries, but it continues across removed entries,
because removed entries are entries that may have been skipped over when
we were adding the value we're searching for to the hash, but have since
been removed. For live entries, we also maintain this distinction by
using one bit of storage for a collision flag, which notes that if the
hashtable entry is removed, its place in the entry store must become a
removed entry rather than a free entry.
When we add a new entry to the table, Add's semantics require that we
return an existing entry if there is one, and only create a new entry if
no existing entry exists. (Bug 1352198 suggests the possibility of a
faster alternative Add API where the caller guarantees that the key is
not already in the hashtable.) When we search for the existing entry,
we must thus continue the search across removed entries, even though we
record the first removed entry found to return if the search for an
existing entry fails.
The existing code adds the collision flag through the entire table
search during an Add. This patch changes that behavior so that we only
add the collision flag prior to finding the first removed entry. Adding
it after we find the first removed entry is unnecessary, since we are
not making that entry part of a path to a new entry. If it is part of a
path to an existing entry, it will already have the collision flag set.
This patch effectively puts an if (!firstRemoved) around the else branch
of the if (MOZ_UNLIKELY(EntryIsRemoved(entry))), and then refactors that
condition outwards since it is now around the contents of both the if
and else branches.
MozReview-Commit-ID: CsXnMYttHVy
--HG--
extra : transplant_source : 0T%B0%FA%C0%85v%8B%16%E7%81%03p%F5K%97%B1%9E%92%27
This adds an arena allocator that can be used as a drop-in replacement for
NSPR's PLArena. Example usage for defining an 8-byte aligned allocator that
uses a 4K arena size:
mozilla::ArenaAllocator<4096,8> a;
void* memory = a.Allocate(200);
This adds release bounds checking to ReplaceElementsAt, InsertElementAt, and
InsertElementsAt to make sure the insertion point is within the current array
bounds.
MozReview-Commit-ID: 1pFr8LuOROI
Handling potential nsDeque size changes means a bit of extra work.
But if the nsDeque is const, we can assume that it shouldn't get modified, so
we can provide a more optimized iterator that doesn't need to handle size
changes.
Optimizing a range-for loop in which the deque is not modified, can be done
by writing: `for (void* item : const_cast<const nsDeque&>(deque)) {...}`
MozReview-Commit-ID: AFupjoTsoH3
--HG--
extra : rebase_source : a71b09c9cb73787ce686c7c762f92ef0c208e76a
Note that iterators stay at the same index if the deque size changes
(including end-iterators staying at the end).
This means that after front operations, iterators will effectively point at
different elements! (Possibly skipping or re-visiting some.)
But this is consistent with ForEach and hand-crafted index-based for loops.
MozReview-Commit-ID: 5IvazJR68dG
--HG--
extra : rebase_source : c574fd2d2642d784482698c0fc861269200d1059
It's now possible to write:
for (void* item : deque) { ... }
MozReview-Commit-ID: FLoczCZd77y
--HG--
extra : rebase_source : 237293e94b478beb2bf352c1179d42c289dda145
This could have been done more simply, but the small amount of
refactoring that takes place in this comment enables better error
messages in the case where something does go wrong.