Previously, the font was applied to the Labels but the name wasn't
updated on initial startup.
This meant that the Label's content was only correct in the default
state but not if the user changed the defaults.
When building userland with UBSAN enabled (#7434), we were getting
spammed to death by unaligned access errors.
Fix these by adding 2 bytes of padding to the FontFileHeader struct,
and adjusting all our font files to match the new format. :^)
I introduced a regression in #7184 where `TTY` would report 1 byte read
in canonical mode even if we had no more characters left. This was
caused by counting the '\0' that denotes EOF into the number of
characters that were read.
The fix was simple: exclude the EOF character from the number of bytes.
This still wouldn't be correct by itself, as the EOF and EOL control
characters could change between when the data was written to the TTY and
when it is read. We fix this by signaling out-of-band whether something
is a special character. End-of-file markers have a value of zero and
have their special bits set. Any other bytes with a special flag are
treated as line endings. This is possible, as POSIX doesn't allow
special characters to be 0.
Fixes#7419
Right clicking on back or forward will now show a context menu with
URLs to navigate to. Also added an optional argument for the number of
steps in go_back() and go_forward().
Previously the process' m_profiling flag was ignored for all event
types other than CPU samples.
The kfree tracing code relies on temporarily disabling tracing during
exec. This didn't work for per-process profiles and would instead
panic.
This updates the profiling code so that the m_profiling flag isn't
ignored.
Options shamelessly stolen from this article on systemd's website:
https://systemd.io/TESTING_WITH_SANITIZERS/
We make ASAN more strict and tell UBSAN to print more verbose output on
failure. One of the more interesting ASAN options,
detect_stack_use_after_return, sadly causes both UBSAN and ASAN failures
in test-js.
This replaces the types of m_int_value and m_frac_value with
Checked<u64> which makes it possible to check if the value overflowed
when entering a digit. If that happens, the digit will just be ignored.
This fixes#1263.
This changes the keydown_event handler to use codepoints instead of key
codes for comparison if possible. This is so the functionality still
works as intended with keyboard layouts where e.g. typing '+' actually
results in KeyCode::Key_ExclamationPoint rather than KeyCode::Key_Plus.
This also removes the unnecessary call to atoi().
When loading debug info, we encounter the same filename over and over
(since files usually have many lines!) and we were wasting a ton of
time re-checking if the filename was part of the Toolchain or libgcc,
along with some other checks.
This patch makes prepare_lines() significantly faster by memoizing
the result of these checks per filename.
This makes "bt 12" ~25% faster (from 850ms to 650ms on my machine.) :^)
While playing with conditionally disabling -O2 optimization when
building the Userland subdirectory, I discovered that we can no longer
link errors without -O2. This happens as LibM.so doesn't link to
anything else, resulting in no stack protector implementation. It
appears that optimization somehow avoids this problem?
To fix this inject LibC/ssp.cpp as we do with in dynamic loader.
Previously each malloc size class would keep around a limited number of
unused blocks which were marked with MADV_SET_VOLATILE which could then
be reinitialized when additional blocks were needed.
This changes malloc() so that it also keeps around a number of blocks
without marking them with MADV_SET_VOLATILE. I termed these "hot"
blocks whereas blocks which were marked as MADV_SET_VOLATILE are called
"cold" blocks because they're more expensive to reinitialize.
In the worst case this could increase memory usage per process by
1MB when a program requests a bunch of memory and frees all of it.
Also, in order to make more efficient use of these unused blocks
they're now shared between size classes.
The new one is way more naive and not as fancy as the old one, but it
doesn't crash when trying to draw circles.
This algorithm just sweeps the angles required by the call, makes sure
each segment is at most 1 (pixel) long and just uses the standard
parameterization to find the coordinates of each point on the ellipse.
Move the from handshake() to the constructor (and move the constructor
out-of-line while we're at it.)
This prepares getting rid of the handshake() mechanism since this is
the only remaining user.
Allow the user to pass in a hostname to the connect functionality
instead of just an ip address. The call to gethostbyname(..) handles
both seamlessly for us.
Update the argument name / usage message accordingly.
