Add a renderer for applications running on the primary screen. This
renderer uses relative cursor positioning, and works with all the same
primitives as the alternate screen.
Fix a bug where repositioning was never turned back to false. This was a
nasty one with huge perf implications.
Set internal cells to a space and default on init. This prevents us from
writing them. As a result, we now issue a hardware clear on resize.
Previously, stopping the read thread would not fully clean up the tty.
This could result in issues such as #9 where the terminal state was
borked upon return to shell or if temporarily exiting the TUI.
Fixes: #9
Refactor the main API to split Loop and Vaxis types. This enables the
core Vaxis type to be more easily referenced in other types, since it
doesn't require a comptime type parameter. This will make the switch to
the zg unicode library easier.
Request a device status report to trigger a read. Wait for the thread to
join in the main loop to ensure that the read thread has fully closed.
This should prevent the need for polling and a quit_fd, and a separate
mechanism on macos.
the man page for macOS `poll(2)` says, under "bugs", that polling
devices does not work, and indeed, although the current implementation
based on poll(2) does produce output, it does not appear to actually
be polling for this reason; instead the wait is blocking on the `read`
call, causing the macOS user to need to input another character to
exit.
this change introduces a wrapper over macOS's implementation of
`select`, modeled after `std.io.poll` as `select.zig`. it is a compile
error to use `select.zig` when `builtin.os.tag.isDarwin()` is false. a
lightly altered version of `Tty.zig` accompanies this change. it's
certainly possible to incorporate the two into one file; i didn't just
to leave the other one in its original state.
with this change, writing to the pipe correctly exits.
additionally, on macOS, `Thread.setName` must be called from the
thread whose name you wish to set, so I have just commented out that
line.
When changing to the new zig termios api, I missed setting two flags.
This left SIGINTs being sent to the terminal and applications not able
to handle ctrl+c.
- Added standard .gitattributes file for Zig projects.
- Reworked build.zig a little, hopefully it's a bit clearer. Also, now zig build will run all steps.
- outer: while in examples was redundant since there's only one loop to break from. switch expressions don't allow breaking from them, so breaking is only for loops, i.e. while and for.
- When returning a struct instance from a function, the compiler infers the return type from function signature, so instead of return MyType{...}; , it's more idiomatic to write return .{...};.
- Logging adds a new line by default, so you don't usually need to write \n like here: log.debug("event: {}\r\n", .{event});.
When linking lib_c, os.system gets overwritten with std.c. The linux
implementation of std.c doesn't contain the constants we need.
Signed-off-by: Tim Culverhouse <tim@timculverhouse.com>
Implement our own grapheme measuring function which switches on whether
the terminal supports mode 2027
Signed-off-by: Tim Culverhouse <tim@timculverhouse.com>
Now that DA1 parsing is done, block the queryTerminal function until the
DA1 response is received, or a 1 second timeout elapses. With this
functionality, move certain events into Vaxis's realm of handling: IE
enabling kitty keyboard, unicode mode, etc
Signed-off-by: Tim Culverhouse <tim@timculverhouse.com>
I'm not positive this is the final approach. We fill in the `text` field
if there was multi-codepoint text generated from the key_press.
Signed-off-by: Tim Culverhouse <tim@timculverhouse.com>
This is so similar to my vaxis project for go that we'll retain the name
here as well. The two can keep similar APIs and feature sets, but are
used in different languages. Maybe someday libvaxis will export a C api
for even broader use
Signed-off-by: Tim Culverhouse <tim@timculverhouse.com>
