A typo in the changes to our userland timekeeping classes caused us to
make a syscall every time we want to check whether a timer is ready to
fire in `EventLoopManagerUnix::wait_for_events()`. Instead, only use
coarse time, and get it immediately before it is used in both cases.
This reduces CPU usage by an (eyeballed) 20-30% while playing back
video with VideoPlayer.
Since Core::Object properties are really only used by GML now that the
Inspector is long gone, there's no need for these to pollute
Core::Object.
This patch adds a new GUI::Object class to hold properties, and makes
it the new base class of GUI::Window, GUI::Widget and GUI::Layout.
The "instantiate an object by name" mechanism that GML uses is also
hoisted into GUI::Object as well.
Since the existing Promise class is designed with deferred tasks on the
main thread only, we need a new class that will ensure we can handle
promises that are resolved/rejected off the main thread.
This new class ensures that the callbacks are only called on the same
thread that the promise is fulfilled from. If the callbacks are not set
before the thread tries to fulfill the promise, it will spin until they
are so that they will run on that thread.
This is based on Jakt::File::current_executable_path() and all the other
sources I looked at to figure out the per-platform way to do this. My
goodness, every platform has its own bespoke way.
In order to follow spec text to achieve this, we need to change the
underlying representation of a host in AK::URL to deserialized format.
Before this, we were parsing the host and then immediately serializing
it again.
Making that change resulted in a whole bunch of fallout.
After this change, callers can access the serialized data through
this concept-host-serializer. The functional end result of this
change is that IPv6 hosts are now correctly serialized to be
surrounded with '[' and ']'.
Returning a reference resulted in Mail's use of Promise causing a crash.
Also, awaiting an already-awaited promise is an odd thing to do anyway,
so let's just make it release the resolved/rejected value instead of
returning a reference to it.
Co-Authored-By: Valtteri Koskivuori <vkoskiv@gmail.com>
The Test262RunnerHandler class in test-test262 was made in order to
spawn a subprocess, connect to its input/output error pipes, and obtain
its return value.
Later on, a copy of this implementation was added to TestSed with
modifications, such as adding support for reading from the output pipes
as well.
Unify these two implementations into a new Core::Command class. This new
implementation is more closely modeled from the TestSed implementation
due to the extra functionality it implemented.
This allows us to get rid of another mime-type list in the codebase.
To do so, the `get_description_from_mime_type` function is introduced in
this patch.
We used to have two separate lists for mime-type detection, one for
detection based on the file extension, and the other one for detection
based on byte sniffing.
This list also contains a general description that will be of use in
`file.cpp`.
To create this list, I had to fill in some gaps as the two lists started
to diverge.
These three lines were added in commits 41d5531, 0f7a651 and 97aca8f all
the way back in June 2020, and went unnoticed until Lucas pointed this
out during my refactoring.
These too should be sorted alphabetically, as evidenced by the fact that
text/markdown was in there twice before this change. :^)
Also broke out tables of sufffixes and basenames we consider plaintext,
and sorted those alphabetically as well.
It makes far more sense to sort by the standard mime type strings,
rather than the ad-hoc variable names associated with each enumeration.
This way, the sort looks nicer, and also matches the corresponding
description enumerations in the file utility.
The previous iteration of this API was somewhat odd and rough in random
places, which degraded usability and made less than perfect sense.
This commit reworks the API to be a little closer to more
conventional promise APIs (a la javascript promises).
Also adds a test to ensure the class even works.
Much like the previous commit, this commit makes the
ArgParser::Arg::accept_value callback return an ErrorOr<bool> instead of
just a bool.
The aim of this is to make argument parsing more robust, especially
with the newer String api that returns an ErrorOr for many functions.
This commit adds the ability to use the String class with `add_option`
and `add_positional_argument`.
This should help with the transition away from DeprecatedString.
This is a preparation before we can create a usable mechanism to use
filesystem-specific mount flags.
To keep some compatibility with userland code, LibC and LibCore mount
functions are kept being usable, but now instead of doing an "atomic"
syscall, they do multiple syscalls to perform the complete procedure of
mounting a filesystem.
The FileBackedFileSystem IntrusiveList in the VFS code is now changed to
be protected by a Mutex, because when we mount a new filesystem, we need
to check if a filesystem is already created for a given source_fd so we
do a scan for that OpenFileDescription in that list. If we fail to find
an already-created filesystem we create a new one and register it in the
list if we successfully mounted it. We use a Mutex because we might need
to initiate disk access during the filesystem creation, which will take
other mutexes in other parts of the kernel, therefore making it not
possible to take a spinlock while doing this.
Similar to commit bcfa4a3, this enables navigating to a file:// URL in
Ladybird with audio file extensions.
The file extensions used here were taken from
https://docs.fileformat.com/audio/ (except QOA, which isn't listed).
This commit changes the variables used to represent the size and
progress of downloads from u32 to u64. This allows `pro` and
`Browser` to report the total size and progress of a download
correctly for downloads larger than 4GiB.
Calling `from_utf8` with a DeprecatedString will hide the fact that we
have a DeprecatedString, while using `from_deprecated_string` with a
StringView will silently and needlessly allocate a DeprecatedString,
so let's forbid that.
When called with `SearchInPath::Yes`, calls to `Core::System::exec()`
would fail. The value returned from
`resolve_executable_from_environment()` was assigned to a StringView,
but the original reference was not kept, causing the StringView to
appear empty.
Finally, the end of our journey to fallible I/O!
Now all that's left to do is to rewrite the code around each and every
.release_value_but_fixme_should_propagate_errors() call. :^)
Closes#17129.
DeprecatedFile doesn't properly handle I/O or OOM errors, and only
provides a rudimentary interface to reasonably handle it. We have long
since learned how to do it "properly" with ErrorOr<> and similar
interfaces. See also d43a7eae54.
Note that this leaves behind an invocation to DeprecatedFile in
Ladybird/AndroidPlatform.cpp. However, that part of the system has
compilation errors since at least January, and will get removed or
rewritten as part of issue #19085. Let's not wait for this Android port
to remove this deprecated API.
The if statement for setting the O_NONBLOCK flag was written the other
way around, causing Core::File to open files in non-blocking mode, when
we didn't actually specify it.
Some applications may not want to have the ability to create a
file if it doesn't exist, but still be able to read and write
from it. The easy solution here would be just to not apply
O_CREAT when creating the flags, but to prevent breaking a ton
of applications, having a `DontCreate` mode is the best for now.
`Core::System::fstatat()` is similar to our standard `Core::System`
wrappers.
`Core::Directory::stat(relative_path, flags)` is a convenience method if
you already have a Directory, to stat a file relative to it.
This closely mirrors Socket::set_blocking. Note that it does not make
sense to make this a virtual method on a base class, since
SeekableStream and FixedMemoryStream cannot possible be anything except
than blocking.
That's what this class really is; in fact that's what the first line of
the comment says it is.
This commit does not rename the main files, since those will contain
other time-related classes in a little bit.
This gives us free error-propagation in Core::command(...) and
HackStudio::ProjectBuilder::for_each_library_dependencies.
The comment about "String will be in the null state" has been misleading
for a long time, so it is removed.
These 2 are an actual separate types of syscalls, so let's stop using
special flags for bind mounting or re-mounting and instead let userspace
calling directly for this kind of actions.
It's unnecessary to allocate a string when we only want to compare it
with another string.
This change also adds a helper for string literals, so that we won't
need to add -sv suffix everywhere. :^)
This class, in a similar fashion to what has been done with
`InputBufferedStream`, postpones write to the stream until an internal
buffer is full.
This patch also adds the `OutputBufferedFile` alias.
This shouldn't have been moved to EventLoopManager, as the manager is
global and one-per-process, and the implementation is one-per-loop.
This makes cross-thread event posting work again, and unbreaks
SoundPlayer (and probably other things as well.)
Things such as timers and notifiers aren't specific to one instance of
Core::EventLoop, so let's not tie them down to EventLoopImplementation.
Instead, move those APIs + signals & a few other things to a new
EventLoopManager interface. EventLoopManager also knows how to create a
new EventLoopImplementation object.
Using QEventLoop works for everything but it breaks *one* little feature
that we care about: automatically quitting the app when all windows have
been closed.
That only works if you drive the outermost main event loop with a
QCoreApplication instead of a QEventLoop. This is unfortunate, as it
complicates our API a little bit, but I'm sure we can think of a way to
make this nicer someday.
In order for QCoreApplication::exec() to process our own
ThreadEventQueue, we now have a zero-timer that we kick whenever new
events are posted to the thread queue.
The EventLoop is now a wrapper around an EventLoopImplementation.
Our old EventLoop code has moved into EventLoopImplementationUnix and
continues to work as before.
The main difference is that all the separate thread_local variables have
been collected into a file-local ThreadData data structure.
The goal here is to allow running Core::EventLoop with a totally
different backend, such as Qt for Ladybird.
This program has never lived up to its original idea, and has been
broken for years (property editing, etc). It's also unmaintained and
off-by-default since forever.
At this point, Inspector is more of a maintenance burden than a feature,
so this commit removes it from the system, along with the mechanism in
Core::EventLoop that enables it.
If we decide we want the feature again in the future, it can be
reimplemented better. :^)
Not a single client of this API actually used the event mask feature to
listen for readability AND writability.
Let's simplify the API and have only one hook: on_activation.
Instead of juggling events between individual instances of
Core::EventLoop, move queueing and processing to a separate per-thread
queue (ThreadEventQueue).
This was used in exactly one place, to avoid sending multiple
CustomEvents to the enqueuer thread in Audio::ConnectionToServer.
Instead of this, we now just send a CustomEvent and wake the enqueuer
thread. If it wakes up and has multiple CustomEvents, they get delivered
and ignored in no time anyway. Since they only get ignored if there's
no work to be done, this seems harmless.
This is quite useful for userspace applications that can't cope with the
restriction, but it's still useful to impose other non-configurable
restrictions by using jails.
`process.fds()` is protected by a Mutex, which causes issues when we try
to acquire it while holding a Spinlock. Since nothing seems to use this
value, let's just remove it entirely for now.
Rather than the very C-like API we currently have, accepting a void* and
a length, let's take a Bytes object instead. In almost all existing
cases, the compiler figures out the length.
This was fine before as the last entry was a null string (which could be
printed), but we no longer use C-style sentinel-terminated arrays for
arguments.
If it is default-initialized to 0, mktime will assume that DST is not in
effect for the specified time. Setting it to a negative value instructs
mktime to determine for itself whether DST is in effect.
- We were using primitive versions of mkstemp and mkdtemp, they have
been converted to use LibCore::System.
- If an error occurred whilst creating a temporary directory or file, it
was thrown and the program would crash. Now, we use ErrorOr<T> so that
the caller can handle the error accordingly
- The `Type` enumeration has been made private, and `create_temp` has
been "split" (although rewritten) into create_temp_directory and
create_temp_file. The old pattern of TempFile::create_temp(Type::File)
felt a bit awkward, and TempFile::create_temp_file() feels a bit nicer
to use! :^)
Once the Core::Filesystem PR is merged (#17789), it would be better for
this helper to be merged in with that. But until then, this is a nice
improvement.
Similar to POSIX read, the basic read and write functions of AK::Stream
do not have a lower limit of how much data they read or write (apart
from "none at all").
Rename the functions to "read some [data]" and "write some [data]" (with
"data" being omitted, since everything here is reading and writing data)
to make them sufficiently distinct from the functions that ensure to
use the entire buffer (which should be the go-to function for most
usages).
No functional changes, just a lot of new FIXMEs.
In this context, the promises are considered "jobs", and such jobs
depend in some way on the event loop. Therefore, they can be added to
the event loop, and the event loop will cancel all of its pending jobs
when it ends.
This class had slightly confusing semantics and the added weirdness
doesn't seem worth it just so we can say "." instead of "->" when
iterating over a vector of NNRPs.
This patch replaces NonnullRefPtrVector<T> with Vector<NNRP<T>>.
`Core::Directory::for_each_entry()` takes a callback which is passed the
DirectoryEntry and the parent Directory. It returns any error from
creating the iterator, iterating the entries, or returned from the
callback.
As a simple example, this:
```c++
Core::DirIterator piece_set_iterator { "/res/icons/chess/sets/",
Core::DirIterator::SkipParentAndBaseDir };
while (piece_set_iterator.has_next())
m_piece_sets.append(piece_set_iterator.next_path());
```
becomes this:
```c++
TRY(Core::Directory::for_each_entry("/res/icons/chess/sets/"sv,
Core::DirIterator::SkipParentAndBaseDir,
[&](auto const& entry, auto&) -> ErrorOr<IterationDecision> {
TRY(m_piece_sets.try_append(entry.name));
return IterationDecision::Continue;
}));
```
`Directory::path()` returning `ErrorOr` makes it awkward to use, and all
current users create a Directory with a path. If we find we need
pathless directories later, we can come up with a clever solution
then. :^)
This also removes DirIterator::error_string(), since the same strerror()
string will be included when you print the Error itself. Except in `ls`
which is still using fprintf() for now.
Our `find` utility makes use of the `dirent::d_type` field for filtering
results, which `Core::DirIterator` didn't expose. So, now it does. :^)
We now store the name and type of the entry as the "next" value instead
of just the name. The type is exposed as a `DirectoryEntry::Type`
instead of a `DT_FOO` constant, so that we're not tied to posixy
systems, and because proper enums are nice. :^)
This is not guaranteed to always work correctly as ArgsParser deals in
StringViews and might have a non-properly-null-terminated string as a
value. As a bonus, using StringView (and DeprecatedString where
necessary) leads to nicer looking code too :^)
This commit moves the implementation of getopt into AK, and converts its
API to understand and use StringView instead of char*.
Everything else is caught in the crossfire of making
Option::accept_value() take a StringView instead of a char const*.
With this, we must now pass a Span<StringView> to ArgsParser::parse(),
applications using LibMain are unaffected, but anything not using that
or taking its own argc/argv has to construct a Vector<StringView> for
this method.
This currently allocates in .parse(), but that's better than making the
caller do the exact same before passing us the values.
Note that this is only temporary to aid in conversion, a future commit
will remove this and switch to requiring the users to allocate the
vector instead.
At the moment, this processes the RIFF chunk structure and extracts
the ICCP chunk, so that `icc` can now print ICC profiles embedded
in webp files. (And are image files really more than containers
of icc profiles?)
It doesn't even decode image dimensions yet.
The lossy format is a VP8 video frame. Once we get to that, we
might want to move all the image decoders into a new LibImageDecoders
that depends on both LibGfx and LibVideo. (Other newer image formats
like heic and av1f also use video frames for image data.)
This was called from LibCore and passed raw StringView data that may
not be null terminated, then incorrectly passed those strings to
getenv() and also tried printing them with just the %s format
specifier.
Similar to the return values earlier, a signed value doesn't really make
sense here. Relying on the much more standard `size_t` makes it easier
to use Stream in all contexts.
`Stream` will be qualified as `AK::Stream` until we remove the
`Core::Stream` namespace. `IODevice` now reuses the `SeekMode` that is
defined by `SeekableStream`, since defining its own would require us to
qualify it with `AK::SeekMode` everywhere.
Having an alias function that only wraps another one is silly, and
keeping the more obvious name should flush out more uses of deprecated
strings.
No behavior change.
`Process::get_name()` and `Process::set_name()` are basically the same
as `get_process_name()` and `set_process_name()`, except making use of
convenient Serenity standard types and returning ErrorOr, instead of
char* and errno shenanigans.
`Process::set_name()` has an optional `SetThreadName` parameter, for
when you also want to set the thread's name to the same thing. That's
true for the two places that use `set_process_name()`.
Besides from a general check if a file's directory has write
permissions, this also checks if the directory has set a sticky bit,
meaning that only file owners and the directory owner can remove or move
files in such directory. It's being used in /tmp for example.
At the moment, there is no immediate advantage compared to just calling
the underlying functions directly, but having a common interface feels
more ergonomic (users don't have to care about how a type serializes)
and maybe we'll find a way to hide the actual implementation from direct
access some time in the future.
The macOS FileWatcher depends on macOS dispatch queues, which run on a
different thread than the Core::EventLoop. This implementation handles
filesystem events on its dispatch queue, then forwards the event back to
the main Core::EventLoop for notifying the FileWatcher owner.
This will be handy for platforms which need to be able to store extra
OS-specific members. For example, macOS needs to store a dispatch queue,
and event stream, etc.
A negative return value doesn't make sense for any of those functions.
The return types were inherited from POSIX, where they also need to have
an indicator for an error (negative values).
InodeWatcherFlags is an enumeration from the Kernel. To avoid using it
outside of Serenity, add a FileWatcherFlags for FileWatcher, much like
we already have FileWatcherEvent::Type.
This is currently being implicitly including by InodeWatcherEvent.h by
way of FileWatcher.h. The former will soon be removed from the latter,
which would otherwise cause a compile error in these files.
This implements FileWatcher using inotify filesystem events. Serenity's
InodeWatcher is remarkably similar to inotify, so this is almost an
identical implementation.
The existing TestLibCoreFileWatcher test is added to Lagom (currently
just for Linux).
This does not implement BlockingFileWatcher as that is currently not
used anywhere but on Serenity.
This saves us an actual seek and rereading already stored buffer data in
cases where the seek is entirely covered by the currently buffered data.
This is especially important since we implement `discard` using `seek`
for seekable streams.
While at it, rename the `read_trivial_value` and `write_trivial_value`
functions to `read_value` and `write_value` respectively, since we'll
add compatibility for non-trivial types down the line.
We don't need to call `can_read_line()` as the buffer will be populated
by `find_and_populate_until_any_of()`. The change is also beneficial as
the buffer will be populated until a candidate is found and not
necessarily a new line.
The search used to go through the buffer from the start, even if we just
appended a small number of bytes at the end. It now remembers the last
stop and resume the search from it.
This parameter allows to start searching after an offset. For example,
to resume a search.
It is unfortunately a breaking change in API so this patch also modifies
one user and one test.
Ladybird currently doesn't render any webpages on FreeBSD and throws
hundreds of errors,beginning with this:
IPC::ConnectionBase (0x0000000805bf2b00) had an error (File descriptor
passing not supported on this platform), disconnecting.
WebContent process crashed!
There is no particular reason why we shouldn't allow zero-sized reads or
writes here, and this actually might cause issues with our common
stream-to-stream copy pattern if we end up at an unfortunate offset
where the next read would be zero-sized and trigger EOF only after that.
This was unintuitive, and only useful in a few cases. In the majority,
users had to immediately call `stop()`, and several who did want the
timer started would call `start()` on it immediately anyway. Case in
point: There are only two places I had to add a manual `start()`.
Before this patch, Core::SessionManagement::parse_path_with_sid() would
figure out the root session ID by sifting through /sys/kernel/processes.
That file can take quite a while to generate (sometimes up to 40ms on my
machine, which is a problem on its own!) and with no caching, many of
our programs were effectively doing this multiple times on startup when
unveiling something in /tmp/session/%sid/
While we should find ways to make generating /sys/kernel/processes fast
again, this patch addresses the specific problem by introducing a new
syscall: sys$get_root_session_id(). This extracts the root session ID
by looking directly at the process table and takes <1ms instead of 40ms.
This cuts WebContent process startup time by ~100ms on my machine. :^)
Also add some tests that ensure that the input and output streams match
each other, because I can't wrap my head around what the internal
representation looks like.
This removes the direct dependency on sys/time.h from ElapsedTimer, and
makes the code a lot cleaner by using the helpers from AK::Time for
time math and getting the current timestamp.
In doing so, this removes all uses of the Encoder's stream operator,
except for where it is currently still used in the generated IPC code.
So the stream operator currently discards any errors, which is the
existing behavior. A subsequent commit will propagate the errors.
These instances were detected by searching for files that include
stdlib.h, but don't match the regex:
\\b(_abort|abort|abs|aligned_alloc|arc4random|arc4random_buf|arc4random_
uniform|atexit|atof|atoi|atol|atoll|bsearch|calloc|clearenv|div|div_t|ex
it|_Exit|EXIT_FAILURE|EXIT_SUCCESS|free|getenv|getprogname|grantpt|labs|
ldiv|ldiv_t|llabs|lldiv|lldiv_t|malloc|malloc_good_size|malloc_size|mble
n|mbstowcs|mbtowc|mkdtemp|mkstemp|mkstemps|mktemp|posix_memalign|posix_o
penpt|ptsname|ptsname_r|putenv|qsort|qsort_r|rand|RAND_MAX|random|reallo
c|realpath|secure_getenv|serenity_dump_malloc_stats|serenity_setenv|sete
nv|setprogname|srand|srandom|strtod|strtof|strtol|strtold|strtoll|strtou
l|strtoull|system|unlockpt|unsetenv|wcstombs|wctomb)\\b
(Without the linebreaks.)
This regex is pessimistic, so there might be more files that don't
actually use anything from the stdlib.
In theory, one might use LibCPP to detect things like this
automatically, but let's do this one step after another.
These instances were detected by searching for files that include
MemMem.h, but don't match the regex:
\\b(MemMem(?!\.h>)|bitap_bitwise|memmem|memmem_optional)\\b
These are the only symbols defined by MemMem.h.
In theory, one might use LibCPP to detect things like this
automatically, but let's do this one step after another.
In 7c5e30daaa, the focus was "only" on
Userland/Libraries/, whereas this commit cleans up the remaining
headers in the repo, and any new badly-formatted include.
The existing `load_from_gml()` methods look the same as before from the
outside. Inside though, they now forward to `try_load_from_gml()` which
returns Error when things go wrong. It also now calls the `try_create()`
factory method for Objects instead of the `construct()` one.
