refactor(config): parse config values from raw bytes like Zig
Zig parses config values (cli/args.zig, Config.zig field parsers, config/string.zig) from raw byte slices, and each field type decides whether it requires UTF-8 (InvalidValue otherwise) or stores bytes. The port escape-encoded raw bytes into Strings at the line reader and let fields parse the escaped text, so fields could observe encoding artifacts; successive audits kept finding such leaks (RV5-RV8). Field parsers now receive the original bytes: text fields validate UTF-8 where Zig does, byte/path fields store bytes (EncodedOs/OsString). This resolves all 9 RV8 findings (docs/rewrite/notes/review-rv8-codex.md), each with a regression test: Windows path conversion fails with BadPathName instead of substituting U+FFFD; Windows process launch uses strict UTF-8 like Zig's Command; the WTF-8 decoder accepts paired- surrogate encodings like Zig; literal U+FDD0 Unicode key triggers; selection-word-chars and clipboard-codepoint-map reject invalid UTF-8; font style / codepoint-map family names use original bytes; C title actions forward bytes; D-Bus new-window/Flatpak conversions match Zig. Design, sweep and validation: docs/rewrite/notes/final-bytes-refactor.md. Also moves the bundled iTerm2-Color-Schemes license out of resources/themes/ so +list-themes no longer lists it as a theme (output is now byte-identical to Ghostty 1.3.1). Done by Codex (gpt-6-astra), reviewed and verified by the orchestrator. Test Plan: - cargo test --workspace --no-fail-fast (3428 passed, exit 0) - just check-macos (both), just check-windows, just check-wasm - +show-config --default --docs, +list-keybinds, +list-actions --docs, +list-themes --plain compared with /usr/bin/ghostty
This commit is contained in:
1 parent
70c250cf2d
commit
6f9b3129e1
100 files changed
+2310
-1054
No files matched your search
@@ -127,8 +127,8 @@ pub fn action_to_c(action: &Action, keep: &mut Keep) -> ghostty_action_s {
|
||||
v.desktop_notification.title = keep.cstr(&n.title);
|
||||
v.desktop_notification.body = keep.cstr(&n.body);
|
||||
}
|
||||
Action::SetTitle(t) => v.set_title.title = keep.cstr(&t.title),
|
||||
Action::SetTabTitle(t) => v.set_tab_title.title = keep.cstr(&t.title),
|
||||
Action::SetTitle(t) => v.set_title.title = keep.cstr_bytes(&t.title.to_bytes()),
|
||||
Action::SetTabTitle(t) => v.set_tab_title.title = keep.cstr_bytes(&t.title.to_bytes()),
|
||||
Action::PromptTitle(p) => v.prompt_title = *p as c_int,
|
||||
Action::Pwd(p) => {
|
||||
v.pwd.pwd = keep.cstr_bytes(&p.pwd.to_bytes());
|
||||
@@ -158,8 +158,9 @@ pub fn action_to_c(action: &Action, keep: &mut Keep) -> ghostty_action_s {
|
||||
}
|
||||
Action::KeyTable(KeyTable::Activate(name)) => {
|
||||
v.key_table.tag = 0;
|
||||
v.key_table.value.activate.name = keep.cstr(name);
|
||||
v.key_table.value.activate.len = name.len();
|
||||
let bytes = name.to_bytes();
|
||||
v.key_table.value.activate.name = keep.cstr_bytes(&bytes);
|
||||
v.key_table.value.activate.len = bytes.len();
|
||||
}
|
||||
Action::KeyTable(KeyTable::Deactivate) => v.key_table.tag = 1,
|
||||
Action::KeyTable(KeyTable::DeactivateAll) => v.key_table.tag = 2,
|
||||
@@ -240,6 +241,39 @@ mod tests {
|
||||
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn c_title_actions_and_key_table_names_forward_original_bytes() {
|
||||
use ghostty_os::osbytes::EncodedOs;
|
||||
let raw = b"name\xff\xef\xb7\x90\xc3\xa9";
|
||||
let mut keep = Keep::default();
|
||||
for tab in [false, true] {
|
||||
let title = SetTitle {
|
||||
title: EncodedOs::from_bytes(raw),
|
||||
};
|
||||
let action = if tab {
|
||||
Action::SetTabTitle(title)
|
||||
} else {
|
||||
Action::SetTitle(title)
|
||||
};
|
||||
let c = action_to_c(&action, &mut keep);
|
||||
// SAFETY: both title union members have identical layout, pointer owned by keep.
|
||||
let title = unsafe { std::ffi::CStr::from_ptr(c.action.set_title.title) };
|
||||
assert_eq!(title.to_bytes(), raw);
|
||||
}
|
||||
let c = action_to_c(
|
||||
&Action::KeyTable(KeyTable::Activate(EncodedOs::from_bytes(raw))),
|
||||
&mut keep,
|
||||
);
|
||||
// SAFETY: the tag is key_table.activate and keep owns the buffer.
|
||||
let name = unsafe {
|
||||
std::slice::from_raw_parts(
|
||||
c.action.key_table.value.activate.name.cast::<u8>(),
|
||||
c.action.key_table.value.activate.len,
|
||||
)
|
||||
};
|
||||
assert_eq!(name, raw);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn target_cval() {
|
||||
let t = target_to_c(&Target::App);
|
||||
|
||||
@@ -147,8 +147,8 @@ fn build_backing(value: &CValue) -> Backing {
|
||||
for c in cmds {
|
||||
b.strings.push(cstring_value(&c.action_key));
|
||||
b.strings.push(cstring_value(&c.action));
|
||||
b.strings.push(cstring_value(&c.title));
|
||||
b.strings.push(cstring_value(&c.description));
|
||||
b.strings.push(cstring_value(c.title.as_encoded()));
|
||||
b.strings.push(cstring_value(c.description.as_encoded()));
|
||||
}
|
||||
// The CStrings' heap buffers don't move when the Vec grows.
|
||||
b.commands = b
|
||||
@@ -265,7 +265,7 @@ pub unsafe extern "C" fn ghostty_config_load_file(ptr: ghostty_config_t, path: *
|
||||
if let Err(err) = c.config.load_file(&path) {
|
||||
log::error!(
|
||||
"error loading config from file path={} err={err}",
|
||||
path.as_encoded()
|
||||
path.to_text_lossy()
|
||||
);
|
||||
}
|
||||
}
|
||||
@@ -511,11 +511,7 @@ pub unsafe extern "C" fn ghostty_config_trigger(
|
||||
}
|
||||
// SAFETY: `str` points to `len` bytes.
|
||||
let s = unsafe { std::slice::from_raw_parts(str.cast::<u8>(), len) };
|
||||
let Ok(s) = std::str::from_utf8(s) else {
|
||||
log::error!("error finding trigger err=InvalidFormat");
|
||||
return default_trigger();
|
||||
};
|
||||
match BindingAction::parse(s) {
|
||||
match BindingAction::parse_bytes(s) {
|
||||
Ok(action) => {
|
||||
let trigger = c
|
||||
.config
|
||||
|
||||
@@ -17,7 +17,6 @@ use std::{
|
||||
cell::{Cell, OnceCell, RefCell},
|
||||
collections::HashMap,
|
||||
ffi::{CStr, CString, c_char, c_int, c_void},
|
||||
fmt::Write as _,
|
||||
rc::{Rc, Weak},
|
||||
sync::Arc,
|
||||
};
|
||||
@@ -547,7 +546,7 @@ pub struct SurfaceRt {
|
||||
cursor_pos: Cell<CursorPos>,
|
||||
/// The current title of the surface. The embedded apprt saves this so
|
||||
/// that `get_title` works without the implementer needing to save it.
|
||||
title: RefCell<Option<String>>,
|
||||
title: RefCell<Option<ghostty_os::osbytes::EncodedOs>>,
|
||||
}
|
||||
|
||||
impl SurfaceRt {
|
||||
@@ -570,7 +569,7 @@ impl RtSurface for SurfaceRt {
|
||||
None
|
||||
}
|
||||
|
||||
fn get_title(&self) -> Option<String> {
|
||||
fn get_title(&self) -> Option<ghostty_os::osbytes::EncodedOs> {
|
||||
self.title.borrow().clone()
|
||||
}
|
||||
|
||||
@@ -854,36 +853,21 @@ unsafe fn platform_view_from_c(
|
||||
}
|
||||
}
|
||||
|
||||
/// Zig's `std.zig.stringEscape` with non-ASCII code points written as
|
||||
/// `\u{...}` (the config value is parsed as a Zig string; the bytes of
|
||||
/// valid UTF-8 round trip either way). Bytes that are not valid UTF-8 are
|
||||
/// emitted in the reversible OS-byte encoding, which the string parser
|
||||
/// copies through unchanged, so `string::parse_bytes(string_escape(b)) == b`.
|
||||
fn string_escape(input: &[u8]) -> String {
|
||||
let mut out = String::with_capacity(input.len());
|
||||
for chunk in input.utf8_chunks() {
|
||||
for c in chunk.valid().chars() {
|
||||
match c {
|
||||
'\n' => out.push_str("\\n"),
|
||||
'\r' => out.push_str("\\r"),
|
||||
'\t' => out.push_str("\\t"),
|
||||
'\\' => out.push_str("\\\\"),
|
||||
'"' => out.push_str("\\\""),
|
||||
' ' | '!' | '#'..='&' | '\'' | '('..='[' | ']'..='~' => out.push(c),
|
||||
c if c.is_ascii() => {
|
||||
let _ = write!(out, "\\x{:02x}", c as u32);
|
||||
}
|
||||
c => {
|
||||
let _ = write!(out, "\\u{{{:x}}}", c as u32);
|
||||
}
|
||||
}
|
||||
}
|
||||
for &b in chunk.invalid() {
|
||||
out.push('\u{FDD0}');
|
||||
out.push(char::from(b));
|
||||
/// Escape initial input for config/string while retaining arbitrary original bytes.
|
||||
fn string_escape(input: &[u8]) -> ghostty_os::osbytes::EncodedOs {
|
||||
let mut out = Vec::with_capacity(input.len());
|
||||
for &byte in input {
|
||||
match byte {
|
||||
b'\n' => out.extend_from_slice(b"\\n"),
|
||||
b'\r' => out.extend_from_slice(b"\\r"),
|
||||
b'\t' => out.extend_from_slice(b"\\t"),
|
||||
b'\\' => out.extend_from_slice(b"\\\\"),
|
||||
b'"' => out.extend_from_slice(b"\\\""),
|
||||
0..=0x1f | 0x7f => out.extend_from_slice(format!("\\x{byte:02x}").as_bytes()),
|
||||
_ => out.push(byte),
|
||||
}
|
||||
}
|
||||
out
|
||||
ghostty_os::osbytes::EncodedOs::from_bytes(&out)
|
||||
}
|
||||
|
||||
/// Apply the surface options to the config (the first half of Zig's
|
||||
@@ -921,7 +905,7 @@ unsafe fn apply_surface_options(config: &mut Config, opts: &ghostty_surface_conf
|
||||
if let Some(cmd) = unsafe { cstr(opts.command) }.map(CStr::to_bytes)
|
||||
&& !cmd.is_empty()
|
||||
{
|
||||
config.command = Some(Command::Shell(osbytes::bytes_to_config_string(cmd)));
|
||||
config.command = Some(Command::Shell(osbytes::EncodedOs::from_bytes(cmd)));
|
||||
config.wait_after_command = true;
|
||||
}
|
||||
|
||||
@@ -934,10 +918,7 @@ unsafe fn apply_surface_options(config: &mut Config, opts: &ghostty_surface_conf
|
||||
let (Some(k), Some(v)) = (unsafe { cstr(var.key) }, unsafe { cstr(var.value) }) else {
|
||||
continue;
|
||||
};
|
||||
config.env.put(
|
||||
&osbytes::bytes_to_config_string(k.to_bytes()),
|
||||
&osbytes::bytes_to_config_string(v.to_bytes()),
|
||||
);
|
||||
config.env.put_bytes(k.to_bytes(), v.to_bytes());
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1946,11 +1927,10 @@ pub unsafe extern "C" fn ghostty_surface_binding_action(
|
||||
let bytes = unsafe { std::slice::from_raw_parts(action.cast::<u8>(), len) };
|
||||
// The action text is bytes (Zig `[]const u8`); the parser works on the
|
||||
// reversible OS-byte encoding so a `text:` payload keeps invalid UTF-8.
|
||||
let action_str = ghostty_os::osbytes::bytes_to_config_string(bytes);
|
||||
let action = match BindingAction::parse(&action_str) {
|
||||
let action = match BindingAction::parse_bytes(bytes) {
|
||||
Ok(a) => a,
|
||||
Err(err) => {
|
||||
log::error!("error parsing binding action action={action_str} err={err:?}");
|
||||
log::error!("error parsing binding action action={bytes:?} err={err:?}");
|
||||
return false;
|
||||
}
|
||||
};
|
||||
@@ -2194,7 +2174,7 @@ mod tests {
|
||||
] {
|
||||
let esc = string_escape(raw);
|
||||
assert_eq!(
|
||||
ghostty_config::string::parse_bytes(&esc).unwrap(),
|
||||
ghostty_config::string::parse_bytes(esc.to_bytes()).unwrap(),
|
||||
raw,
|
||||
"{esc:?}"
|
||||
);
|
||||
@@ -2241,10 +2221,10 @@ mod tests {
|
||||
let Some(Command::Shell(c)) = &config.command else {
|
||||
panic!("no command");
|
||||
};
|
||||
assert_eq!(osbytes::config_str_to_bytes(c), b"echo \xfe");
|
||||
assert_eq!(c.to_bytes(), b"echo \xfe");
|
||||
let (k, v) = config.env.iter().next().unwrap();
|
||||
assert_eq!(osbytes::config_str_to_bytes(k), b"K\xfd");
|
||||
assert_eq!(osbytes::config_str_to_bytes(v), b"v\xfc");
|
||||
assert_eq!(k.to_bytes(), b"K\xfd");
|
||||
assert_eq!(v.to_bytes(), b"v\xfc");
|
||||
assert!(matches!(
|
||||
config.input.list.first(),
|
||||
Some(ReadableIO::Raw(_))
|
||||
|
||||
Reference in new issue
Block a user