The main changes are twofold:
* Buffer flipping is now controlled by the m_screen_can_set_buffer flag
in WSCompositor. This flag, in turn, is impacted by m_can_set_buffer
flag, in WSScreen. m_can_set_buffer is set in the WSScreen constructor
by checking the return value of fb_set_buffer. If the framebuffer
supports this operation, it will succeed, and we record this fact. This
information is then used by WSCompositor to set its own
m_screen_can_set_buffer flag.
* WSScreen now only requests a resolution change of the framebuffer. The
driver itself is ultimately responsible for what resolution or mode is
actually set, so WSScreen has to read the response from that request,
and has no choice but to accept the answer. This allows the driver to
choose a "close enough" value to what was requested, or simply ignore
it.
The result of this is that there is no special configuration necessary
for WindowServer to work with reduced-capability framebuffer devices.
tty0 receives input while WindowServer is up, which meant that having
a shell on tty0 would cause commands typed into Terminal to run twice,
once in Terminal's shell, and once on the tty0 shell.
Long term we shouldn't send input to any VirtualConsole while the
WindowServer is up, so this just fixes the immediate weirdness. :^)
This new version can do three things:
* When invoked as `mount`, it will print out a list of mounted filesystem,
* When invoked as `mount -a`, it will try to mount filesystems
listed in /etc/fstab,
* When invoked as `mount device mountpoint -t fstype`, it will mount that
device on that mountpoint. If not specified, fstype defaults to ext2.
The complication is around /proc/sys/ variables, which were attached
to inodes. Now they're their own thing, and the corresponding inodes
are lazily created (as all other ProcFS inodes are) and simply refer
to them by index.
All programs that have a CEventLoop now allow local socket connections
via /tmp/rpc.PID and will dump a serialized JSON array of all the live
CObjects in the program onto connecting sockets.
Also added a small /bin/rpcdump tool that connects to an RPC socket and
produces a raw dump of the JSON that comes out.
This is kind of a mess, but because IPC client code depending on the
IPC protocol definition artifacts in the server code, we have to build
the IPC servers first. And their dependencies before that, etc.
One more drop in the "maybe we should switch to CMake" bucket..
makeall.sh used to build the AK tests and leave some binary objects laying
around that would get in the way of further incremental builds. There also
wasn't a lot of structure to the order things were built in. This patch
improves both of those things.
It is now possible to unmount file systems from the VFS via `umount`.
It works via looking up the `fsid` of the filesystem from the `Inode`'s
metatdata so I'm not sure how fragile it is. It seems to work for now
though as something to get us going.
Using the new get_process_name() syscall, we can automatically prefix
all userspace debug logging.
Hopefully this is more helpful than annoying. We'll find out! :^)
WindowServer was led to believe that the Terminal window had an alpha
channel that had to be respected by the compositor. This caused us to
always consider it as non-opaque when culling dirty rects in compose.
We can't rely on all hardware to give us a way to flip between the back
and front buffer.
This mode should actually perform slightly better but may show some tearing
as we don't have a way to know when we're in vertical retrace.
We were forced to do this because the page fault code would fall apart
when trying to generate a backtrace for a non-current thread.
This issue has been fixed for a while now, so let's go back to lazily
loading executable pages which should make everything a little better.
This needs to be applied consistently in the rest of this UI, but here
is a start. I can't figure out what the right look should be right now
so I'm just gonna commit it this way and we'll work it out later. :^)
You can now set a GTableCellPaintingDelegate per column in GTableView.
For columns with a painting delegate set, the view will defer to the
delegate for painting each cell in that column.
