We got ourselves into a mess by making widgets override the window
cursor whenever they wanted a custom cursor. This patch introduces a
better solution to that issue: per-widget override cursors.
Each widget now has an override cursor that overrides the window
cursor when that widget is hovered.
Since "rings" typically refer to code execution and user processes
can also execute in ring 0, rename these functions to more accurately
describe what they mean: kernel processes and user processes.
The ring is determined based on the CS register. This fixes crashes
being handled as ring 3 crashes even though EIP/CS clearly showed
that the crash happened in the kernel.
It has been possible for a signal to arrive in between us checking
g_interrupted and exiting - sucessfully, even though we were interrupted.
This way, if a signal arrives before we reset the disposition, we
will reliably check for it later; if it arrives afterwards, it'll
kill us automatically.
It wasn't actually possible to call
const LogStream& operator<<(const LogStream&, ReadonlyBytes);
because it was shadowed by
template<typename T>
const LogStream& operator<<(const LogStream& stream, Span<T> span);
not sure how I didn't find this when I added the overload.
It would be possible to use SFINAE to disable the other overload,
however, I think it is better to use a different method entirely because
the output can be very verbose:
void dump_bytes(ReadonlyBytes);
In addition to being the proper POSIX etiquette, it seems like a bad idea
for issues like the one seen in #3428 to result in a kernel crash. This patch
replaces the current behavior of failing on insufficient buffer size to truncating
SOCK_RAW messages to the buffer size. This will have to change if/when MSG_PEEK
is implemented, but for now this behavior is more compliant and logical than
just bailing.
This makes Terminal::scroll_up() O(1) instead of O(n) in the
size of the history. (It's still O(n) in the size of visible
lines.)
Reduces time to run `disasm /bin/id` with the default terminal
window size from 530ms to 409ms (min-of-5) on my system.
DC and AC table IDs read in the scan header segment weren't validated
against the IDs of Huffman tables read in the DHT segment. This caused
an OOB read when a Huffman table was accessed using the ID read in the
scan header segment. Furthermore, the decoder now replaces the old DC
or AC table if a redefinition has been found prior to the scan header.
Fixes#3439.
If a scrollbar doesn't have a scrubber (because the view it scrolls is
large enough to display all its contents without scrolling), then
it ignores all clicks. We shouldn't draw a hover highlight that suggests
clickability in that case.
This patchset adds the following to the manpage:
- Mention `if` expressions.
- Add section about subshells
- Mention that control structures can now be used as commands
- Update the grammar.
- Fix small header size mistake with "Example"'s being larger than their
containing sections.
This fixes a duplicate message when running `jobs` for the first time
after a job has been moved to the background.
Also actually announces background exits now :^)
This patchset makes the shell capable of lazily resolving and executing
sequences of commands, to allow for putting logical sequences in the
background.
In particular, it enables And/Or/Sequence nodes to be run in the background,
and consequently unmarks them as `would_execute`.
Doing so also fixes job control to an extent, as jobs are now capable of
having 'tails', so sequences can be put in the background while
preserving their following sequences.
No difference in practice, but it's tidier and protects us
if we ever use g_interrupted earlier in main -- in that case,
the compiler might chose to keep the variable in a register without
the volatile.
Found by @bugaevc, thanks!
By being a bit too greedy and only allocating how much we need for
the failing allocation, we can end up in an infinite loop trying
to expand the heap further. That's because there are other allocations
(e.g. logging, vmobjects, regions, ...) that happen before we finally
retry the failed allocation request.
Also fix allocating in page size increments, which lead to an assertion
when the heap had to grow more than the 1 MiB backup.
Rather than trying to find a contiguous set of bits of size 1, just
find one single available bit using a hint.
Also, try to randomize returned physical pages a bit by placing them
into a 256 entry queue rather than making them available immediately.
Then, once the queue is filled, pick a random one, make it available
again and use that slot for the latest page to be returned.
Leverage constexpr and __builtin_ffs for Bitmap::find_first. Also add
a variant Bitmap::find_one_anywhere that can start scanning at a
provided hint.
Also, merge Bitmap::fill_range into the already existing Bitmap::set_range
With this, hitting ctrl-c twice in `for i in $(seq 10) { sleep 1 }`
terminates the loop as expected (...well, I'd expect it to quit after
just one ctrl-c, but serenity's shell makes a single ctrl-c only
quit the current loop iteration).
Part of #3419.
In c3d231616c we added the atomic variable
m_have_any_unmasked_pending_signals tracking the state of pending signals.
Add helper functions that automatically update this variable as needed.
