\033[1003z will stuff them into stdin like 80,24.
This is a hack, just like the "disable local echo" hack used for
passwords and it should, ideally, be replaced with a proper system as
defined in POSIX for terminal control, but for, it works, and that's
pretty cool because it means `ls` can stop being stupid.
* Some applications now support UTF-8 text through the use of a very
simple decoder.
* The terminal uses a slow, but accurate method to determine the width
of a character the first time it is printed to the screen. Characters
are now stored in the terminal in two bytes, rather than one, and may
in the future be increased to 3 or 4 bytes to ensure support for
Unicode supplemental planes.
* A simple font-fallback method is employed in the applications that
support unicode that will make use of the VL Gothic fonts if the
DejaVu font does not have a character. No guarantees are made for
support of writing systems other than extended Latin and Japanese.
* Finally bring syscall.h up to speed and include all syscalls in the
syscall module of the C library.
* Remove the third-party obfuscated C demos (we have nyancat, good
enough)
* Fix userspace apps to build without complaining about undeclared
strtok_r by disable __STRICT_ANSI__
* Fix .eh_frame by including the proper stuff with libgcc.
- Can now register a userspace file descriptor as the output for kernel
print statements through kprintf()
- Can set logging levels for debug print messages, which are separate
from kernel log events and meant to be more readily visible. Log
events are recorded in a buffer to be viewed later, though nothing
actually using logging at the moment.
- Serial output is disabled by default now. You can enable it yourself
by appending the logtoserial argument to the kernel on boot.
This is an automated system by which we boot qemu headless and use the
serial line to capture output from a testing application that is started
on bootup, running with the VGA terminal shell. This might be expanded
to boot to the graphical display within VNC and perform more advanced
tests with the Python shim using a VNC module for Python; we'll see.
All graphics library commands now take a gfx_context_t pointer, which
points to a simple datastructure describing a rendering context (width,
height, depth, total size, front buffer, backbuffer; where backbuffer =
front buffer when not in double-buffering mode, thus we always render to
backbuffer except on a flip). This may have caused a minor speed
reduction, but I don't really care as it's far more important that we
support multiple graphics contexts.
TODO:
- Shared Memory Fonts library (there are a couple of apps that use these
so-called "shmem fonts" on their own; we need a dedicated library for
them)
- Break off "TTK" GUI toolkit into its own library. Since it's just a
callback-based button framework, this shouldn't be too hard right now.
Also, with the previous tick, I'll be able to put labels on controls
and start using text in more places.
We know longer accept DAMAGE and REDRAW commands. Instead, we just loop
and redraw everything anyway (efficiently, mind you). Pipes still have
issues when closed and various other different things.
Still need to pass sizes around to shm_* functions as a pointer rather
than the actual size so we can retreive sizes; additionally, *p = 0
should mean "we don't want to create this buffer, we only want to
access it if it exists".
It doesn't even need one; eventually, we'll have window management and
you'll just close the app. It should technically just sit in a loop
endlessly after rendering (or waitforsignal, which we don't have yet).
[init] Reenable freetype in the Terminal spawned from init
[nyancat] Fix timing and update animations
[esh] Fix bug with child pid collection and signal handling
This update includes support for shadow passwords, stored in
/etc/master.passwd, as well as support for an /etc/passwd file
containing user/uid/full name/shell associations, which are used by the
shell to get your username for display purposes, as well as by whoami
for the same reason. The login tool does not yet select the right shell
though this is planned.
* root's password is `toor` (a throwback to older times)
* local's password is `local` (because it's obvious)
Without flushing, this doesn't work properly with the terminal pipe.
Technically, this is an issue in the pipe implementation, but as that's
not technically what we're testing there's no sense having it be broken.
For the record, everything works wonderfully.
Completely removes:
* The kernel terminal (both VGA and graphical)
* The kernel ANSI parser (obviously)
* kgets() function
* Dozens of other functions that were made useless
Adds:
* Userspace terminal that should work (relatively) well
* Keyboard device driver (implemented with a "pipe" object)
* Stabalized interrupt interface
* `clear` uses the c library
* All panic screens and kprintf() output goes to the serial line ONLY
* The kernel boots directly into /bin/terminal (no arguments, unless you
want to add them (such as -f))
Need to replace the bitfont with freetype, get the thing attached to a
pipe, start getting other stuff to run under it, and then we can remove
the kernel's graphical terminal.
BIOS execution is provided through the `v8086` module, which provides
software emulation of an 8086 processor. It is not currently working
with some BIOSes and may (read: probably will be) replaced with another
emulator (x86emu comes to mind) at some point in the near future. In the
meantime, the default video mode for QEMU works with this and it's
enough to get us on real VESA instead of fake VBE. The `bochs` module
will be renamed in a future commit. Userspace programs have been
adjusted to work at bitrates other than 32 *POORLY*. If you write pixels
left-to-right, they should work fine. They only work with 24-bpp
otherwise, and then you need to be careful of what pixels you are
writing when, or you will overwrite things in other pixels.
You may pass a commandline argument like the following to set display
modes:
vid=vesa,1024,768
Or for stranger modes under QEMU or Bochs, use the bochs VBE
initializer:
vid=bochs,1280,720
Note that the address of the linear framebuffer is still found via
hackish probing instead of PCI or trusting the VBE information, so if
you have things in the wrong memory ranges (0xE0000000+), be prepared to
have them get read.
Once again, this entire commit is a massive hack. I am happy that it
worked, and I will continue to make it less hacky, but in the meantime,
this is what we've got.
Happy holidays.
- Free process resources when a process exits (reaped in next process
cycle; should probably reap after a wait() or something)
- Free process struct after wait()
- Fix page allocation
- Fix fork() return value for child process (attempted to write to an
invalid point in kernel-stack memory)
We shouldn't be triple faulting randomly anymore!
Continue investigating the fork() return value, as there was a bugged
return at some point during executon of a test run of thrash-process.
