This wasn't much of a problem before because copying the ByteBuffer
merely copied the RefPtr but now that ByteBuffer behaves like Vector
this causes unnecessary allocations.
Previously ByteBuffer would internally hold a RefPtr to the byte
buffer and would behave like a reference type, i.e. copying a
ByteBuffer would not create a duplicate byte buffer, but rather
two objects which refer to the same internal buffer.
This also changes ByteBuffer so that it has some internal capacity
much like the Vector<T> type. Unlike Vector<T> however a byte
buffer's data may be uninitialized.
With this commit ByteBuffer makes use of the kmalloc_good_size()
API to pick an optimal allocation size for its internal buffer.
Previously deleting an animated image wouldn't make the animation timer
stop. This resulted in the animation still running in the ViewWidget.
Moreover the timer wasn't stopped when loading different images, which
led to high CPU usage when going from an animated image to a
non-animated one.
Previously some actions like Rotate/Flip/Set as Desktop Wallpaper would
make the application crash if no image was loaded. Now image actions are
enabled/disabled based on whether an image has been loaded or not.
This feels like a better name since the "autocomplete engine" can, in
addition to providing autocomplete suggestions, also find declarations
of symbols and report back the symbols that are defined in a document.
Also, Cpp/ParserAutoComplete has been renamed to CppComprehensionEngine
and Shell/AutoComplete has been renamed to ShellComprehensionEngine.
While the waste stack and the playable card on top of the waste stack
are collectively referred to as the "waste", it's programatically nice
to separate them to enable 3-card-draw mode. In that mode, the playable
stack will contain 3 cards with a slight x-axis shift, while the waste
stack underneath will remain unshifted. So rather than introducing some
ugly logic to CardStack to handle this, it's more convenient to have a
separate stack on top of the waste stack.
This non-POSIX header is used in Linux/BSD systems for storing the
default termios settings. This lets us setup new TTYs' `m_termios.c_cc`
in a nicer way than using a magic string.
Bytes in the 0x80..0x9F range were treated as C1 control codes,
which prevented them from being parsed as UTF-8 bytes.
This caused some characters (like U+DF, encoded as 0xC3 0x9F)
from being recognized as printable characters.
Since we now store intermediate characters separately, the intermediates
should be checked for the presence of the '?' DEC private marker, not
the first parameter.
This commit replaces the former, hand-written parser with a new one that
can be generated automatically according to a state change diagram.
The new `EscapeSequenceParser` class provides a more ergonomic interface
to dealing with escape sequences. This interface has been inspired by
Alacritty's [vte library](https://github.com/alacritty/vte/).
I tried to avoid changing the application logic inside the `Terminal`
class. While this code has not been thoroughly tested, I can't find
regressions in the basic command line utilities or `vttest`.
`Terminal` now displays nicer debug messages when it encounters an
unknown escape sequence. Defensive programming and bounds checks have
been added where we access parameters, and as a result, we can now
endure 4-5 seconds of `cat /dev/urandom`. :D
We generate EscapeSequenceStateMachine.h when building the in-kernel
LibVT, and we assume that the file is already in place when the userland
library is being built. This will probably cause problems later on, but
I can't find a way to do it nicely.
This program turns a description of a state machine that takes its input
byte-by-byte into C++ code. The state machine is described in a custom
format as specified below:
```
// Comments are started by two slashes, and cause the rest of the line
// to be ignored
@name ExampleStateMachine // sets the name of the generated class
@namespace Test // sets the namespace (optional)
@begin Begin // sets the state the parser will start in
// The rest of the file contains one or more states and an optional
// @anywhere directive. Each of these is a curly bracket delimited set
// of state transitions. State transitions contain a selector, the
// literal "=>" and a (new_state, action) tuple. Examples:
// 0x0a => (Begin, PrintLine)
// [0x00..0x1f] => (_, Warn) // '_' means no change
// [0x41..0x5a] => (BeginWord, _) // '_' means no action
// Rules common to all states. These take precedence over rules in the
// specific states.
@anywhere {
0x0a => (Begin, PrintLine)
[0x00..0x1f] => (_, Warn)
}
Begin {
[0x41..0x5a] => (Word, _)
[0x61..0x7a] => (Word, _)
// For missing values, the transition (_, _) is implied
}
Word {
// The entry action is run when we transition to this state from a
// *different* state. @anywhere can't have this
@entry IncreaseWordCount
0x09 => (Begin, _)
0x20 => (Begin, _)
// The exit action is run before we transition to any *other* state
// from here. @anywhere can't have this
@exit EndOfWord
}
```
The generated code consists of a single class which takes a
`Function<Action, u8>` as a parameter in its constructor. This gets
called whenever an action is to be done. This is because some input
might not produce an action, but others might produce up to 3 (exit,
state transition, entry). The actions allow us to build a more
advanced parser over the simple state machine.
The sole public method, `void advance(u8)`, handles the input
byte-by-byte, managing the state changes and requesting the appropriate
Action from the handler.
Internally, the state transitions are resolved via a lookup table. This
is a bit wasteful for more complex state machines, therefore the
generator is designed to be easily extendable with a switch-based
resolver; only the private `lookup_state_transition` method needs to be
re-implemented.
My goal for this tool is to use it for implementing a standard-compliant
ANSI escape sequence parser for LibVT, as described on
<https://vt100.net/emu/dec_ansi_parser>
By constraining two implementations, the compiler will select the best
fitting one. All this will require is duplicating the implementation and
simplifying for the `void` case.
This constraining also informs both the caller and compiler by passing
the callback parameter types as part of the constraint
(e.g.: `IterationFunction<int>`).
Some `for_each` functions in LibELF only take functions which return
`void`. This is a minimal correctness check, as it removes one way for a
function to incompletely do something.
There seems to be a possible idiom where inside a lambda, a `return;` is
the same as `continue;` in a for-loop.
Previously, all sorts of weird stuff would happen when the editor was at
the last line of the terminal (or when the printed line would be at the
last line), this commit makes the editor scroll the terminal up before
trying to write to a row that doesn't actually exist (yet).
This fixes ^R search making a mess when initiated at the last line
(especially with multiline prompts).
This implements different blend modes in the SoftwareRasterizer by
first setting up the blend factors then rendering the pixels into a
temporary buffer and finally mixing the contents of the temporary buffer
with the contents of the backbuffer based on the blend factors.
