We were only forwarding the value to the subviews, but not storing it
in m_model_column. This would cause MultiView::model_column() to return
the wrong value.
Thanks to Daniel Bos for spotting this! :^)
In the old model, before bc319d9e88, the parent
(the background thread) would delete us when it exits (i.e. never), so we had to
keep track of our own refcount in order to destroy ourselves when we're done.
With bc319d9e88, the parent keeps additional
reference to us, so:
* There should be no need to explicitly ref() ourselves
* The unref() would not get rid of the last reference to us anymore
The latter is why all the BackgroundAction's were getting leaked. Fix this by
simply unparenting ourselves from the background thread when we're done.
FilePicker was not showing thumbnails correctly because once each
thumbnail rendering BackgroundAction completed, it posted a deferred
invocation event to the *main* event loop.
Since FilePicker runs in a nested event loop, those completion
callbacks never ran until it was too late and the FilePicker was gone.
There is some sort of issue with using a SortingProxyModel together
with ColumnsView. This is a workaround to allow FilePicker to use a
MultiView for now, but this needs to be fixed separately somehow.
A MultiView is a combination of ItemView, TableView and ColumnsView
smashed into a single widget. You can switch between the view modes
by calling MultiView::set_view_mode().
Note that MultiView inherits from StackWidget, not AbstractView.
That's purely for practical reasons, although I'm not entirely sure
if there would be some benefit to having it inherit from AbstractView.
This is a little bit awkward since it's only used for generating
thumbnails on a background thread and it's not like I care about
thumbnails very much in a text editor, but for now let's just pledge
"thread" so I can get on with the thing I wanted to get on with.
Previously it was only possible to change these window attributes when
creating a new window. This patch adds an IPC message that allows you
to change them at runtime.
Each process has a 1-level lookup cache for fast repeated lookups of
the same VM region (which tends to be the majority of lookups.)
The cache is used by the following syscalls: munmap, madvise, mprotect
and set_mmap_name.
After a succesful exec(), there could be a stale Region* in the lookup
cache, and the new executable was able to manipulate it using a number
of use-after-free code paths.
Now the ACPI & PCI code is more safer, because we don't use raw pointers
or references to objects or data that are located in the physical
address space, so an accidental dereference cannot happen easily.
Instead, we use the PhysicalAddress class to represent those addresses.
Now we use the GenericInterruptHandler class instead of IRQHandler in
the CPU functions.
This commit adds an include to the ISR stub macros header file.
Also, this commit adds support for IRQ sharing, so when an IRQHandler
will try to register to already-assigned IRQ number, a SharedIRQHandler
will be created to register both IRQHandlers.
Also, the enable() function is now correct and will use the right
registers and values. In addition to that, write_register() and
read_registers() are not relying on identity mapping anymore.
Also, update the class implementation to use PCI::Device class
accordingly.
The create() helper will now search for an IDE controller in the
PCI bus, allowing to simplify the initialize() method.
This class represents a shared interrupt handler. This class will not be
created automatically but only if two IRQ Handlers are sharing the same
IRQ number.
The GenericInterruptHandler class will be used to represent
an abstract interrupt handler. The InterruptManagement class will
represent a centralized component to manage interrupts.
