- Checks to make sure lines count coming from the pty are within a proper
range before doing scrolling.
- Sanitizes scroll region when being set.
- Changes panic for unimplemented screen clear to a print statement.
The first two changes ensure scrolling won't crash us. By sanitizing the
region on set we don't have to complicate the scroll code with limits,
mins, or maxes to ensure the scroll operation is within the range.
Checking if the lines is greater than the total region allows us to
simply clear the region and avoid subtracting large numbers from small
ones.
Most of the crashes on resize were due to columns and lines being set to
zero. This causes all sorts of other checks within the code to ensure
these values are greater than zero before running calculations. To avoid
this we just need to ensure that lines and columns are some non zero
value. This is seems to be what gnome terminal does. I have selected
2 lines and two columns for min terminal size for now.
- Remove the use of limit.
- Reduce the number of comparisons.
When using numbers provided by the PTY for subtractions there is a extra
step of ensuring that we won't trigger failure on testing when trying to
subtract form zero.
** NOTE **
This commit fails fails the tmux_git_log test. I am submitting a PR to
talk about the test. I think the test was generated before a few things
were fixed the final test gird still has cells that should have been
scrolled off the screen. Also, comparing output from gnome-terminal
there is no difference. So this PR is here to discuss and gather
information on balding test and discussing the possibility that this test
may be flawed.
** NOTE **
There was a lot of complexity around the threadsafe `Flag` type and
waking up the event loop. The idea was to prevent unnecessary calls to
the glutin window's wakeup_event_loop() method which can be expensive.
This complexity made it difficult to get synchronization between the pty
reader and the render thread correct. Now, the `dirty` flag on the
terminal is also used to prevent spurious wakeups. It is only changed
when the mutex is held, so race conditions associated with that flag
shouldn't happen.
This adds the ability to click and drag with the mouse and have the
effect of visually selecting text. The ability to copy the selection
into a clipboard buffer is not yet implemented.
This adds a trait OnResize and a separate method handle_resize to the
display. Instead of having a callback to receive resize events, a list
of &mut OnResize are passed to this new method. Doing this allowed the
only RefCell usage in the codebase to be removed :).
The cell under the cursor is now tracked in the input processor at
`self.mouse.line` and `self.mouse.column`. This could probably be
optimized to only compute the cell when in certain states, but the
calculation is cheap.
This is part of an ongoing decoupling effort across the codebase and
tidying effort in main.rs. Everything to do with showing the window with
a grid of characters is now managed by the `Display` type. It owns the
window, the font rasterizer, and the renderer. The only info needed from
it are dimensions of characters and the window itself for sizing the
terminal properly. Additionally, the I/O loop has access to wake it up
when new data arrives.
ANSI escape sequences like `\x1b[48;5;10m` were not supported until now.
Specifically, the second attribute, 5, says that the following attribute
is a color index.
The ref tests were updated since `enum Color` variants changed.
Michael Brumlow