Some programs explicitly ask for a different initial stack size than
what the OS provides. This is implemented in ELF by having a
PT_GNU_STACK header which has its p_memsz set to the amount that the
program requires. This commit implements this policy by reading the
p_memsz of the header and setting the main thread stack size to that.
ELF::Image::validate_program_headers ensures that the size attribute is
a reasonable value.
This replaces all state-related variables with a single ThreadState.
These are simplified over what the Kernel has, but capture all
userspace-available thread state.
Locking the state behind an atomic and using proper atomic operations
also gets rid of quite some deadlocks and race conditions that have
existed around m_tid and others beforehand.
In terms of behavior, this introduces the following changes:
- All thread state mishandling (e.g. joining a detached thread) crashes
the program. Mishandling thread state is a severe kind of concurrency
bug that might also be indeterministic, so letting it silently
disappear with the return value of pthread_ APIs is a bad idea. The
thread state can always be checked beforehand to ensure that no crash
happens.
- Destructing a still-running thread will crash in AK/Function, so the
Thread destructor issues its own warning for debugging purposes.
- Thread issues warnings before crashes in many places to aid
concurrency debugging (the most difficult kind of debugging).
- Joining dead but not detached threads is legal, as per POSIX APIs.
- The thread ID is never reset to 0 after the thread has been started
and subsequently been assigned a valid thread ID. The thread's exit
state is still obtainable.
- Detaching threads that are about to exit is considered a programming
bug and will often (not always, as we can't catch all execution
sequences involved in such a situation) crash the program on purpose.
If you want to detach a thread that will definitely exit on its own,
you have to prevent it from exiting before detach() was called (e.g.
with an "exit requested" flag).
This is just used to key the socket fd for system server takeover. On
macOS, this file path has a space in it, which trips up the parsing as
it splits on spaces. That parsing should be fixed (probably shouldn't
rely on spaces as a delimter), but for now, we can change the key to
avoid spaces.
Previously, the word was highlighted red in case it was not found in the
dictionary. That color was repurposed as a general "invalid input" color
to nudge the player that something was wrong with the last input.
Accordingly, the field m_last_word_not_in_dictionary was renamed to
m_last_word_invalid
In the "inspiration" for this game, messages are displayed on top of the
game area in case an invalid guess is inputted. After a few seconds,
they disappear. In a similar fashion, a statusbar is created on the game
window and similar messages are outputted there.
Even if block has all children inline there need to be a check
if it creates BFC because otherwise IFC will be looking in
wrong parent BFC to calculate space used by floats.
We previously passed both OperandSize and AddressSize to the
constructor.
Both values were only ever 32-bit at construction.
We used AddressSize::Size64 to signify Long mode which was needlessly
complicated.
As the existing near-by comment says, the default size of displacements
& immediates is 32 bits even in Long mode.
This makes `disasm` work on our binaries in x86-64 builds.
This adds shortcut for inserting a new empty indented line
above/below current cursor position.
- <Ctrl-Return> for inserting line below.
- <Ctrl-Shift-Return> for inserting line above.
Added in d522a6f and 1e604b7, their purpose snuffed out in 11bb88f
like the faint pulse of a pleading candle, two lives of short excess,
doomed to itemize their sins to no effect and for all eternity...
This deduplicates argument handling logic from Help and man and makes it
more modular for future use cases. The argument handling works as
before: two arguments specify section and page (in this order), one
argument specifies either a page (the first section that it's found in
is used) or a path to a manpage markdown file.
- Calculate the full name on demand
- Make section and name protected
- Reorder some members logically
- Change the name getter to be fallible, as some implementors need to
allocate
This is a first step in deduplicating code within and across Help and
man.
Because LibManual also doesn't contain any DeprecatedString, some
adjustments to Help's string handling is included, just to interoperate
with LibManual better. Further work in this area mostly requires String
APIs in LibGUI.
On Serenity, SQLServer is started by SystemServer. But on Lagom, it is
manually started by e.g. Ladybird when the application is started, and
killed when the application exits. This means every Ladybird process
starts its own SQLServer, which defeats the purpose of SQLServer acting
as the single process interacting with the database files.
This patch will allow SQLClient to start up a single SQLServer instance,
first checking if one already exists. If it does exist, SQLClient will
simply connect to SQLServer's socket. If it does not exist, SQLClient
will launch SQLServer much like SystemServer would (with a local socket
file, etc.).
The child SQLServer process is double-forked; the grandchild process
becomes the SQLServer process, which the middle child process simply
exits to "detach" the grandchild process from the SQLClient process.
Currently, we create a new SQL::Database object for each database we are
requested to open. When multiple clients connect to the same database,
the same underlying database file is opened and cached each time. This
results in updates from one client not being propagated to others.
To prevent this, when a database is requested to be open, check if it is
already open. We can then re-use that SQL::Database object for the new
connection.
This used to be the other way around. If we just inserted input with
document.write, this would always be true and not allow document.write
to immediately parse its input (given that there's no pending parsing
blocking script)
Instead of having two separate context menus and popping up either the
"file" or "directory" one depending on the selected node, we now have a
single context menu and update it (before popping it up) to show the
context-appropriate actions.
This is achieved by simply updating the visibility of the actions.
This takes care of one TODO! :^)
This patch adds a visibility state to GUI::Action. All actions default
to being visible. When invisible, they do not show up in toolbars on
menus (and importantly, they don't occupy any space).
This can be used to hide/show context-sensitive actions dynamically
without rebuilding menus and toolbars.
Thanks to Tim Slater for assuming that action visibility was a thing,
which gave me a reason to implement it! :^)
When you select a text area in "bottom-up" way (e.g. from line 10
to line 5), then type the shortcut, the text editor will not
comment those text for you.
Normalize the text range can easily fix this minor bug.
The indexes are into the _node_, not in the fragment, so when a node is
split into multiple fragments, simply taking the length of the fragment
is incorrect. This patch corrects this mistake.
This action was originally added so that Magnifier's window would
always be on top by default, but it's a redundant menu item and wasn't
actually setting itself at start-up. Instead, rely on the same menu
item provided to all Modeless windows by default for a more consistent
UX, and set the option after show() so it takes effect.
