To achieve this it now uses recursive descend, to make the state
managements easier.
With this we now correctly handle try-catch-finally blocks correctly,
modeling all possible controll flows between these blocks, which allows
analysis and optimization passes to make more accurate descisions about
accessibility of the enclosed blocks
Instead of just iterating indexed properties and direct properties on
the object's shape, let's walk the prototype chain and grab every
non-symbol property by re-using the AO for for-in statements.
We'll probably want to move this to a custom algorithm just for pretty
printing that doesn't skip some property types, but this already gives
a major improvement when printing LegacyPlatformObjects from LibWeb to
the console.
Instead of just calling JS::Value::to_string_without_side_effects() when
printing values to the console, have all the console clients use
the same JS::Print that the REPL does to print values.
This method leaves some things to be desired as far as OOM hardening
goes, however. We should be able to create a String in a way that
doesn't OOM on failure so hard.
To add grouping to a number, we take a string such as "123456.123" and
break it into integer and fraction parts. Then we take the integer part
and break it into locale-specific sized groups to inject the locale's
group separator (e.g. a comma in en-US). We currently create new strings
for each of these groups. Instead, we can use the shared superstring
method to avoid all of that string copying.
The iterator used to find an intrinsic accessor is used after calling
`HashMap.remove()` on it, which works for our current implementation but
will fall apart when you consider that modifications to the hash map
might invalidate all existing iterators that came from it, as many
implementations do.
Since we're aiming to replace our `HashTable` implementation with
something new, let's fix this first :^)
This includes an Error::create overload to create an Error from a UTF-8
StringView. If creating a String from that view fails, the factory will
return an OOM InternalError instead. VM::throw_completion can also make
use of this overload via its perfect forwarding.
If we are out of memory, we can't try to allocate a string that could
fail as well. When Error is converted to String, this would result in an
endless OOM-throwing loop. Instead, pre-allocate the string on the VM,
and use it to construct the Error.
Note that as of this commit, the OOM string is still a DeprecatedString.
This is just preporatory for Error's conversion to String.
This makes it ever-so-slightly faster, but more importantly, it fixes
the bug where a `/\//` regex's `source` property would return `\\/`
("\\\\/") instead of `\/` due to the existing '/' -> '\/' replace()
call.
This will be needed by Value::to_string_without_side_effects, which can
be called in contexts without a VM (e.g. in AK::Format specializations).
So to_string_without_side_effects will need to be callable without a VM,
thus NumberToString must be as well.
For example the words "can't" and "32.3" should not have boundaries
detected on the "'" and "." code points, respectively.
The String test cases fixed here are because "b'ar" is now considered
one word.
Depending on how this is invoked, the preprocessor may get confused when
pasting the VM parameter into this expression. For example, it trips up
on the JS REPL in cases such as:
TRY_OR_THROW_OOM(*g_vm, ...);
Not an issue currently, but while developing, it's easy to miss cases
where an infallible AK::StringBuilder method is still used. By making
this inheritance private, and explicitly pulling in base methods we
can safely use, we get extra help from the compiler to indicate such
mistakes immediately.
