With this change, System::foo() becomes Core::System::foo().
Since LibCore builds on other systems than SerenityOS, we now have to
make sure that wrappers work with just a standard C library underneath.
This isn't a complete conversion to ErrorOr<void>, but a good chunk.
The end goal here is to propagate buffer allocation failures to the
caller, and allow the use of TRY() with formatting functions.
Also add slightly richer parse errors now that we can include a string
literal with returned errors.
This will allow us to use TRY() when working with JSON data.
Previously, trailing whitespaces were not removed from values in
config files. This could cause errors with poorly formatted files.
This commit fixes this by trimming whitespaces from values in
ConfigFile::reparse().
Derivatives of Core::Object should be constructed through
ClassName::construct(), to avoid handling ref-counted objects with
refcount zero. Fixing the visibility means that misuses like this are
more difficult.
There is also make_ref_counted(), which does not call did_construct(),
so the method was not guaranteed to be run. Since there is only a single
user, and `WindowServer::Window` is a final class anyway (so there is no
need to separate the constructor and post-constructor phases), let's get
rid of this concept.
(The following commits reduce the opportunities to call
make_ref_counted, but still.)
ArgsParser will now automatically look for an environment variable
'ARGSPARSER_EMIT_MARKDOWN', and if it is set to exactly the string "1"
(i.e. mere presence or values like "ON" or "yes" are not enough), then
ArgsParser will emit a Markdown-formatted help message, instead of the
regular help message designed for consumption through a terminal.
Just like month, the day value here is one-based. This resulted in the
following situation, which is obviously unexpected:
Core::DateTime::create(1970); // 1970-01-00 -> 1969-12-31
Returns the size in bytes for a file path given its filename. Useful
when file size is needed without having to open the file to query it
using fstat() or seeking to the end.
The main event loop pushes itself onto the event loop stack, and so it
should also pop itself when destroyed.
This will surface attempts to use the event loop stack after the main
event loop has been destroyed.
For example, consider the following SecretString construction:
String foo = "foo";
auto ss = SecretString::take_ownership(foo.to_byte_buffer());
The ByteBuffer created by to_byte_buffer() will not contain the NUL
terminator. Therefore, the value returned by SecretString::characters
will not be NUL-terminated either.
Currently, the only use of SecretString is to pass its character data to
crypt(), which requires a NUL-terminated string. To ensure this cannot
result in a buffer overrun, make SecretString append a NUL terminator to
its buffer if there isn't one already.
This matches the API of Account::authenticate. The only caller to this
API is the passwd utility, which already has the new password stored as
a SecretString.
When a socket's user doesn't need it to be active, but wants to keep it
open, the socket's notifiers should not be enabled to avoid hogging the
CPU with effectively useless notifications.
This API can be used to disable said notifiers until the user needs the
notifications.
This restriction doesn't make much sense, a user should be free to
consume the buffered data with as small a max_size as they desire.
This fixes a possible RequestServer spin when talking to HTTP servers.
If the server responds with this header, we _must_ close the connection,
as the server is allowed to ignore the socket and not respond to
anything past that response.
Fixes some RequestServer spins.
Prior this change, if you had a program which frequently reads small
amount of bytes, then it would constantly fire up syscalls.
This patch sets that the minimum size that is passed to the read syscall
is 1024 and then it saves these additional bytes in a buffer for next
reads, which greatly improves the cpu usage on such cases.
In other words: reading flacs is now very efficient.
This was used to work around a possible bug where select() would mark a
socket readable, but another user of the same socket would read before
the notifier's callback is run; let's remove this and fix any issues
regarding that specific situation later when they pop up.
