In QtCreator (and under weird configurations with gcc), this used to
fail with the error messages like: "error: member of anonymous union
redeclares '___'".
This patch gives each member a unique name.
VM::resolve_binding() can now return a Reference that knows the exact
binding index if it's pointing into a DeclarativeEnvironment.
Reading/writing through the Reference will now use direct environment
access when possible.
This patch adds two DeclarativeEnvironment APIs:
- get_binding_value_direct()
- set_mutable_binding_direct()
These work identically to their non-direct-suffixed counterparts, but
take an index instead of a bound name. This will allow someone who has
a binding index to get/set that binding directly without any additional
hash lookups.
The previous storage for DeclarativeEnvironment looked like this:
HashMap<FlyString, Binding> m_bindings;
This patch changes that to:
HashMap<FlyString, size_t> m_names;
Vector<Binding> m_bindings;
The main goal here is to give each binding an index that can ultimately
be cached and used for optimized environment accesses.
Per the spec, only a BlockContainer" can have line boxes, so let's not
clutter up every Layout::Box with line boxes.
This also allows us to establish an invariant that BFC and IFC always
operate on a Layout::BlockContainer.
Note that if BlockContainer has all block-level children, its line boxes
are not used for anything. They are only used in the all inline-level
children scenario.
List item markers will never have children, so let's mark them as such,
which now allows our layout system to skip over their "insides" and
going straight to positioning instead.
Some boxes cannot have children (most commonly replaced elements),
and so there is nothing meaningful inside them to layout.
We now use the can_have_children() flag to quickly skip over such boxes
instead of creating a formatting context and other pointless busywork.
There's a subtle difference here. A "block box" in the spec is a
block-level box, while a "block container" is a box whose children are
either all inline-level boxes in an IFC, or all block-level boxes
participating in a BFC.
Notably, an "inline-block" box is a "block container" but not a "block
box" since it is itself inline-level.
The CSS spec uses the name "block formatting context root" when talking
about a box that establishes a BFC. So let's call it BFC::root() in our
code as well, instead of the less specific BFC::context_box().
Until now, we've internally thought of the CSS "display" property as a
single-value property. In practice, "display" is a much more complex
property that comes in a number of configurations.
The most interesting one is the two-part format that describes the
outside and inside behavior of a box. Switching our own internal
representation towards this model will allow for much cleaner
abstractions around layout and the various formatting contexts.
Note that we don't *parse* two-part "display" yet, this is only about
changing the internal representation of the property.
Spec: https://drafts.csswg.org/css-display
Instead of doing layout synchronously whenever something changes,
we now use a basic event loop timer to defer and coalesce relayouts.
If you did something that requires a relayout of the page, make sure
to call Document::set_needs_layout() and it will get coalesced with all
the other layout updates.
There's lots of room for improvement here, but this already makes many
web pages significantly snappier. :^)
Also, note that this exposes a number of layout bugs where we have been
relying on multiple relayouts to calculate the correct dimensions for
things. Now that we only do a single layout in many cases, these kind of
problems are much more noticeable. That should also make them easier to
figure out and fix. :^)
This patch adds contexts to line iterator for nesting list items and
blockquotes. It also incidentally makes the api for LineIterator
simpler, and will make it easier to add other containers in the future.
Consider the situation where two shared libraries libA and libB, both
depending (as in having a NEEDED dtag) on libC. libA is first
dlopen()-ed, which produces libC to be mapped and linked. When libB is
dlopen()-ed the DynamicLinker would re-map and re-link libC though,
causing any previous references to its old location to be invalid. And
if libA's PLT has been patched to point to libC's symbols, then any
further invocations to libA will cause the code to jump to a virtual
address that isn't mapped anymore, therefore causing a crash. This
situation was reported in #10014, although the setup was more convolved
in the ticket.
This commit fixes the issue by distinguishing between a main program
loading being performed by Loader.so, and a dlopen() call. The main
difference between these two cases is that in the former the
s_globals_objects maps is always empty, while in the latter it might
already contain dependencies for the library being dlopen()-ed. Hence,
when collecting dependencies to map and link, dlopen() should skip those
that are present in the global map to avoid the issue described above.
With this patch the original issue seen in #10014 is gone, with all
python3 modules (so far) loading correctly.
A unit test reproducing a simplified issue is also included in this
commit. The unit test includes the building of two dynamic libraries A
and B with both depending on libline.so (and B also depending on A); the
test then dlopen()s libA, invokes one its function, then does the same
with libB.
Similarly to regexp_initialize() this can be a member function instead
of taking a RegExpObject argument.
Having it available outside RegExpPrototype is also useful for other
things that need RegExp.prototype.source behavior - e.g. the REPL for
pretty-printing.
Instead of iterating *all* swept cells when pruning weak containers,
only iterate the cells actually *in* the container.
Also, instead of compiling a list of all swept cells, we can simply
check the Cell::state() flag to know if something should be pruned.
Currently, `evaluate()` recalculates whether the MediaQuery matches or
not, and stores it in `m_matches`, which users can query using
`matches()`. This allows us to know when the match-state changes, which
is required to fire MediaQueryList's change event.
In cases where we know the Length is absolute, we know we don't need to
pass in a Layout::Node or FontMetrics etc, and yet we were required to
before. Splitting it means jumping through less hoops that we don't have
to. :^)
This method provides the needed information to evaluate media queries.
Every feature in Media Queries Level 4 is present, either as code or as
a FIXME: https://www.w3.org/TR/mediaqueries-4/#media-descriptor-table
There's a draft Level 5 which I have ignored for now.
Some are unimplemented for now since we do not have access to the
requested information. Some require StyleValue types that we do not yet
support. Many are hard-coded for now since we do not (and may never)
support monochrome or text-only displays for Browser.
Instead of checking storage_has(), followed by storage_get(), we can do
storage_get() directly and avoid a redundant property lookup.
This exposed a bug in SimpleIndexedPropertyStorage::get() which would
previously succeed for array holes.
The idea here is simple: If the block statement doesn't contain any
lexical declarations, we don't need to allocate, initialize and
eventually garbage collect a new declarative environment.
This even makes lookups across nested blocks slightly faster as we don't
have to traverse a chain of empty environments anymore - instead, the
execution context just stores the outermost non-empty one.
This doesn't speed up test-js considerably, but has a noticeable effect
on test262 and real-world web content :^)
From the commonmark spec:
A list is loose if any of its constituent list items are separated by
blank lines, or if any of its constituent list items directly contain
two block-level elements with a blank line between them. Otherwise a
list is tight. (The difference in HTML output is that paragraphs in a
loose list are wrapped in <p> tags, while paragraphs in a tight list are
not.)
This gives FunctionNode a "might need arguments object" boolean flag and
sets it based on the simplest possible heuristic for this: if we
encounter an identifier called "arguments" or "eval" up to the next
(nested) function declaration or expression, we won't need an arguments
object. Otherwise, we *might* need one - the final decision is made in
the FunctionDeclarationInstantiation AO.
Now, this is obviously not perfect. Even if you avoid eval, something
like `foo.arguments` will still trigger a false positive - but it's a
start and already massively cuts down on needlessly allocated objects,
especially in real-world code that is often minified, and so a full
"arguments" identifier will be an actual arguments object more often
than not.
To illustrate the actual impact of this change, here's the number of
allocated arguments objects during a full test-js run:
Before:
- Unmapped arguments objects: 78765
- Mapped arguments objects: 2455
After:
- Unmapped arguments objects: 18
- Mapped arguments objects: 37
This results in a ~5% speedup of test-js on my Linux host machine, and
about 3.5% on i686 Serenity in QEMU (warm runs, average of 5).
The following microbenchmark (calling an empty function 1M times) runs
25% faster on Linux and 45% on Serenity:
function foo() {}
for (var i = 0; i < 1_000_000; ++i)
foo();
test262 reports no changes in either direction, apart from a speedup :^)
In ECMAScriptFunctionObject::function_declaration_instantiation() we
iterate over all lexically declared names of the function scope body to
determine whether any of them is named 'arguments', because we don't
need to create an arguments object in that case. We can also stop at
that point, because the decision won't change anymore.
Up to now the only ``SELECT`` statement that worked was ``SELECT *
FROM <table>``. This commit allows a column list consisting of
column names and expressions in addition to ``*``. ``WHERE``
still doesn't work though.
The database the sql client connected to was 'hardcoded' to the login
name of the calling user.
- Extended the IPC API to be more expressive when connecting, by
returning the name of the database the client connected to in the
'connected' callback.
- Gave the sql client a command line argument (-d/--database) allowing
an alternative database name to be specified
A subsequent commit will have a dot command allowing the user to
connect to different databases from the same sql session.
If the specified value for these properties is "none", we end up storing
it as an "undefined" CSS::Length in the computed values.
We need to convert it back into "none" for getComputedStyle().
