Previously y position of boxes in block formatting context
was calculated by looking at y position of previous in-flow
sibling and adding collapsed margin of "collapse through"
boxes lying between box currently being laid out and it's
previous in-flow sibling.
Here introduced BlockMarginState structure that maintains
array of currently collapsible margins hence we no longer
need to look at previous sibling to calculate y position
of a box.
It may be handy to have some sort of storage inspector at some point but
for now, it doesn't make sense to open a database file. So only allow
opening script files, and don't make assumptions on their extension.
This adds a combo box to the action toolbar to allow for entering a
database name manually or selecting from the list of existing databases.
The action to run a script is now disabled while we are not connected to
a database.
It's not particularly useful to see the word count of a SQL script,
except for when displaying the number of selected words. This changes
SQLStudio to behave exactly like HackStudio in this regard. We will use
segment 0 to display the selected text stats (if any) and segment 2 for
the cursor position. Segment 1 will be used in an upcoming commit for
the current SQL connection status. We also now handle displaying action
text the same way as HackStudio.
Most actions do not need to care about whether there is an open editor
tab, as we can (and should) disable those actions when there isn't an
open tab. We can also hide the verify_cast handling inside a helper
function.
The underlying Core::Stream methods require the bytes passed in to be
non-empty. Simply opening the file is enough to ensure the file is
created with empty contents if the editor's text is empty.
In both applications, display the SQL statement that failed to parse.
For the REPL, ensure the REPL prompts the user for another statement.
For SQLStudio, we don't continue executing the script as it likely does
not make sense to run statements that come after a failed statement.
Otherwise the `move(result)` statement inside the lambda does not
actually move anything, because `result` is constant without the mutable
attribute. Caught by clangd.
These are currently hitting the `decltype(nullptr)` constructor, which
marks the response as invalid, resulting in no response being sent to
the waiting client.
This commit changes the init.cpp file to start and initialize the
Scheduler, and actually runs init_stage2. To show that it actually
works, another thread is spawned and executed simultaneously, by context
switching between the two!
This requires two new functions, context_first_init and
restore_context_and_eret. With this code in place, we are now running
the first idle thread! :^)
This changes the stack pointer to the initial_thread stack pointer, and
pushes two pointers onto the stack that point to the initial_thread. The
function then jumps to the ip of the initial_thread, which will be
thread_context_first_enter, and hangs there because that function is not
yet implemented.
This does not handle everything correctly yet, such as setting the
correct state for running userspace applications, however this should be
enough to get kernel scheduling to work.
These are architecture-specific anyway, so they belong in the Arch
directory. This commit also adds ThreadRegisters::set_initial_state to
factor out the logic in Thread.cpp.
I can't think of a reason why copying the Processor class makes sense,
so lets make sure it's not possible to do it by accident by declaring
the copy constructor as deleted.
This removes the x86 specific hlt instruction from the scheduler, and
allows us to run the scheduler code for aarch64 by implementing
Processor::wait_for_interrupt for aarch64.
This file does not contain any architecture specific implementations,
so we can move it to the Kernel base directory. Also update the relevant
include paths.
We are actually storing tpidr_el0, as can be seen in vector_table.S, but
the RegisterState.h incorrectly had tpidr_el1. This will probably save
some annoying debugging later on.
Mark other ErrorOr types as friends, and fix a typo in the &&
constructor, so that we can create an ErrorOr<Core::Object> from an
ErrorOr<GUI::Widget>. Also, add some requires() clauses to these
constructors so the error messages are clearer.
Having a `Point`, `Rect` or `Size` claim it's `null` is silly. We have
`Optional<T>` for that. For `Point`, rename `is_null` to `is_zero` to
better reflect what we're testing. For `Rect` and `Size`, `is_null` is
removed outright.
Also, remove `is_empty` from `Point`. Points can't be empty.
We were calling `Gfx::Rect<T>::is_null` which checked if the width and
height were 0. What we want to do instead, is check if the rect value
was set at all. `Optional<Rect>` is the right way to do this.
This was useful when building both i686 and x86_64 SerenityOS targets as
we could use a single toolchain build for both targets. But now all this
extra job does is create the opportunity for the toolchain to need to be
built twice (i.e. if the pipelines are backed up and the toolchain cache
is busted between these jobs while the x86_64 step is waiting for a VM).
