Instead of aborting the program when we hit an assertion, just print a
message and keep going.
This allows us to write tests that provoke assertions on purpose.
There was a bug in the "prepend_vector_object" test but it was masked
by us not printing failures. (The bug was that we were adding three
elements to the "objects" vector and then checking that another
vector called "more_objects" indeed had three elements. Oops!)
This is a complement to append() that works by constructing the new
element in-place via placement new and forwarded constructor arguments.
The STL calls this emplace_back() which looks ugly, so I'm inventing
a nice word for it instead. :^)
It doesn't seem sane to try to iterate over a HashTable while it's in
the middle of being cleared. Since this might cause strange problems,
this patch adds an assertion if an iterator is constructed during
clear() or rehash() of a HashTable.
An operation often has two pieces of underlying information:
* the data returned as a result from that operation
* an error that occurred while retrieving that data
Merely returning the data is not good enough. Result<> allows exposing
both the data, and the underlying error, and forces (via clang's
consumable attribute) you to check for the error before you try to
access the data.
Put simply, Error<> is a way of forcing error handling onto an API user.
Given a function like:
bool might_work();
The following code might have been written previously:
might_work(); // but what if it didn't?
The easy way to work around this is of course to [[nodiscard]] might_work.
But this doesn't work for more complex cases like, for instance, a
hypothetical read() function which might return one of _many_ errors
(typically signalled with an int, let's say).
int might_read();
In such a case, the result is often _read_, but not properly handled. Like:
return buffer.substr(0, might_read()); // but what if might_read returned an error?
This is where Error<> comes in:
typedef Error<int, 0> ReadError;
ReadError might_read();
auto res = might_read();
if (might_read.failed()) {
switch (res.value()) {
case EBADF:
...
}
}
Error<> uses clang's consumable attributes to force failed() to be
checked on an Error instance. If it's not checked, then you get smacked.
We had some kernel-specific gizmos in AK that should really just be in the
Kernel subdirectory instead. The only thing remaining after moving those
was mmx_memcpy() which I moved to the ARCH(i386)-specific section of
LibC/string.cpp.
So we already have ByteBuffer::wrap() which is like a StringView for random
data. This might not be the best abstraction actually, but this will be
immediately useful so let's add it.
Clang loses the typestate when passing NonnullRefPtr's via lambda captures.
This is unfortunate, but not much we can do about it. Allowing ptr() makes
it possible to use captured NonnullRefPtrs as you'd expect.
Add an "ElementType" typedef to NonnullOwnPtr and NonnullRefPtr to allow
clients to easily find the pointee type. Then use this to remove a template
argument from NonnullPtrVector. :^)
It's not possible to grow one of these vectors beyond what's already in them
since it's not possible to default-construct Nonnull{Own,Ref}Ptr.
Add Vector::shrink() which can be used when you want to shrink the Vector
and delete resize() from the specialized Vectors.
This works just like NonnullRefPtr, except for NonnullOwnPtr's instead.
NonnullOwnPtrVector<T> inherits from Vector<NonnullOwnPtr<T>>, and adds some
comforts on top, like making accessors return T& so we can chase dots (.)
instead of arrows (->) :^)
This is just like OwnPtr (also single-owner), except it cannot be null.
NonnullOwnPtr is perfect as the return type of functions that never need to
return nullptr.
It's also useful as an argument type to encode the fact that the argument
must not be nullptr.
The make<Foo>() helper is changed to return NonnullOwnPtr<Foo>.
Note: You can move() out of a NonnullOwnPtr, and after that the object is
in an invalid state. Internally it will be a nullptr at this point, so we'll
still catch misuse, but the only thing that should be done in this state
is running the destructor. I've used consumable annotations to generate some
warnings when using a NonnullOwnPtr after moving from it, but these only
work when compiling with clang, so be aware of that.
Restructure the makefile a little so it only builds objects once, and
then run them on make clean.
This is a little slower (since we're relinking tests each makeall), but
it also ensures that it will work.
And use it in the scheduler.
IntrusiveList is similar to InlineLinkedList, except that rather than
making assertions about the type (and requiring inheritance), it
provides an IntrusiveListNode type that can be used to put an instance
into many different lists at once.
As a proof of concept, port the scheduler over to use it. The only
downside here is that the "list" global needs to know the position of
the IntrusiveListNode member, so we have to position things a little
awkwardly to make that happen. We also move the runnable lists to
Thread, to avoid having to publicize the node.
The to_foo() functions are for converting when you might not be sure of the
underlying value type. The as_foo() family assumes that you know exactly
what the underlying value type is.
