diff --git a/.appveyor.yml b/.appveyor.yml
deleted file mode 100644
index 66989c948995377f9359618c47aac586ae6e94d9..0000000000000000000000000000000000000000
--- a/.appveyor.yml
+++ /dev/null
@@ -1,32 +0,0 @@
-os: Visual Studio 2015
-
-environment:
- matrix:
- # Stable 64-bit MSVC
- - channel: stable
- target: x86_64-pc-windows-msvc
- # Nightly 32-bit MSVC
- - channel: nightly
- target: i686-pc-windows-msvc
- # Beta 32-bit GNU
- - channel: beta
- target: i686-pc-windows-gnu
-
-matrix:
- allow_failures:
- - channel: nightly
- - channel: beta
-
-install:
- - appveyor DownloadFile https://win.rustup.rs/ -FileName rustup-init.exe
- - rustup-init -yv --default-toolchain %channel% --default-host %target%
- - set PATH=%PATH%;%USERPROFILE%\.cargo\bin
- - rustc -vV
- - cargo -vV
-
-# Uses 'cargo test' to run tests and build. Alternatively, the project may call compiled programs
-# directly or perform other testing commands. Rust will automatically be placed in the PATH
-# environment variable.
-build: false
-test_script:
- - cargo test --all --verbose
diff --git a/.cargo_vcs_info.json b/.cargo_vcs_info.json
index 2e372ed9553e6a0145c9b5f96e8f21bf0000b2a4..9448ae4f6e9984b06e4c5a1616ea6f28713d27c1 100644
--- a/.cargo_vcs_info.json
+++ b/.cargo_vcs_info.json
@@ -1,5 +1,6 @@
{
"git": {
- "sha1": "bb3ffd6bf97436c0363a7f63fb00cc1620272ee9"
- }
-}
+ "sha1": "38b6cd950ff359724f4f36685bb4b5465cda06b4"
+ },
+ "path_in_vcs": ""
+}
\ No newline at end of file
diff --git a/.gitignore b/.gitignore
deleted file mode 100755
index 6aa106405a4b4a00d779beab77c90c9473a6d203..0000000000000000000000000000000000000000
--- a/.gitignore
+++ /dev/null
@@ -1,3 +0,0 @@
-/target/
-**/*.rs.bk
-Cargo.lock
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
deleted file mode 100644
index 200d884cb64a9cbe4fff03aa442365ac96f7ecb0..0000000000000000000000000000000000000000
--- a/.gitlab-ci.yml
+++ /dev/null
@@ -1,10 +0,0 @@
-# This file is a template, and might need editing before it works on your project.
-# Official language image. Look for the different tagged releases at:
-# https://hub.docker.com/r/library/rust/tags/
-image: "rust:latest"
-
-# Use cargo to test the project
-test:cargo:
- script:
- - rustc --version && cargo --version # Print version info for debugging
- - cargo test --verbose
diff --git a/Cargo.toml b/Cargo.toml
index 8a610afe56b887d9ea5645acb501bd65a30a2988..6a920c9e786b68753a56440a12b648116b42d4d8 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -3,31 +3,60 @@
# When uploading crates to the registry Cargo will automatically
# "normalize" Cargo.toml files for maximal compatibility
# with all versions of Cargo and also rewrite `path` dependencies
-# to registry (e.g. crates.io) dependencies
+# to registry (e.g., crates.io) dependencies.
#
-# If you believe there's an error in this file please file an
-# issue against the rust-lang/cargo repository. If you're
-# editing this file be aware that the upstream Cargo.toml
-# will likely look very different (and much more reasonable)
+# If you are reading this file be aware that the original Cargo.toml
+# will likely look very different (and much more reasonable).
+# See Cargo.toml.orig for the original contents.
[package]
+edition = "2018"
name = "wild"
-version = "2.0.2"
+version = "2.1.0"
authors = ["Kornel <kornel@geekhood.net>"]
+include = [
+ "/src/*.rs",
+ "/Cargo.toml",
+ "/LICENSE",
+ "/README.md",
+]
description = "Glob (wildcard) expanded command-line arguments on Windows"
-homepage = "https://crates.rs/crates/wild"
+homepage = "https://lib.rs/crates/wild"
documentation = "https://docs.rs/wild"
readme = "README.md"
-keywords = ["wildcards", "glob", "windows", "shell", "CommandLineToArgvW"]
-categories = ["command-line-interface"]
-license = "MIT"
+keywords = [
+ "wildcards",
+ "glob",
+ "windows",
+ "shell",
+ "CommandLineToArgvW",
+]
+categories = [
+ "command-line-interface",
+ "os::windows-apis",
+]
+license = "Apache-2.0 OR MIT"
repository = "https://gitlab.com/kornelski/wild"
+
+[package.metadata.docs.rs]
+targets = ["x86_64-unknown-linux-gnu"]
+all-features = true
+rustdoc-args = [
+ "--cfg",
+ "docsrs",
+]
+
[dev-dependencies.glob]
version = "0.3"
+
[target."cfg(windows)".dependencies.glob]
version = "0.3"
+
[badges.appveyor]
repository = "pornel/wild"
[badges.gitlab]
repository = "kornelski/wild"
+
+[badges.maintenance]
+status = "passively-maintained"
diff --git a/Cargo.toml.orig b/Cargo.toml.orig
index eb384f661126b3ded0b631323d671a36890b9c7a..3915a4140f209a614a03865d840482c3624a42bd 100644
--- a/Cargo.toml.orig
+++ b/Cargo.toml.orig
@@ -1,33 +1,30 @@
-# THIS FILE IS AUTOMATICALLY GENERATED BY CARGO
-#
-# When uploading crates to the registry Cargo will automatically
-# "normalize" Cargo.toml files for maximal compatibility
-# with all versions of Cargo and also rewrite `path` dependencies
-# to registry (e.g., crates.io) dependencies
-#
-# If you believe there's an error in this file please file an
-# issue against the rust-lang/cargo repository. If you're
-# editing this file be aware that the upstream Cargo.toml
-# will likely look very different (and much more reasonable)
-
[package]
-name = "wild"
-version = "2.0.2"
authors = ["Kornel <kornel@geekhood.net>"]
+categories = ["command-line-interface", "os::windows-apis"]
description = "Glob (wildcard) expanded command-line arguments on Windows"
-homepage = "https://crates.rs/crates/wild"
documentation = "https://docs.rs/wild"
-readme = "README.md"
+homepage = "https://lib.rs/crates/wild"
keywords = ["wildcards", "glob", "windows", "shell", "CommandLineToArgvW"]
-categories = ["command-line-interface"]
-license = "MIT"
+license = "Apache-2.0 OR MIT"
+name = "wild"
+readme = "README.md"
repository = "https://gitlab.com/kornelski/wild"
-[dev-dependencies.glob]
-version = "0.3"
-[target."cfg(windows)".dependencies.glob]
-version = "0.3"
-[badges.appveyor]
-repository = "pornel/wild"
+version = "2.1.0"
+edition = "2018"
+include = ["/src/*.rs", "/Cargo.toml", "/LICENSE", "/README.md"]
+
+[badges]
+gitlab = { repository = "kornelski/wild" }
+appveyor = { repository = "pornel/wild" }
+maintenance = { status = "passively-maintained" }
+
+[target.'cfg(windows)'.dependencies]
+glob = "0.3"
+
+[dev-dependencies]
+glob = "0.3"
-[badges.gitlab]
-repository = "kornelski/wild"
+[package.metadata.docs.rs]
+targets = ["x86_64-unknown-linux-gnu"]
+all-features = true
+rustdoc-args = ["--cfg", "docsrs"]
diff --git a/README.md b/README.md
index 663113573e58eb0ce545b2f3c4fc47d9f74dccce..85ea5806fc4af083bfba477d77326d57b0c97609 100644
--- a/README.md
+++ b/README.md
@@ -1,10 +1,10 @@
-# [`Wild::args`](https://crates.rs/crates/wild) for [Rust](https://www.rust-lang.org)
+# [`Wild::args`](https://lib.rs/crates/wild) for [Rust](https://www.rust-lang.org)
Allows Rust applications support wildcard arguments (`*foo*`, `file.???`, `*.log.[0-9]`, etc.) on command-line, uniformly on all platforms, including Windows.
Unix shells automatically interpret wildcard arguments and pass them expanded (already converted to file names) to applications, but Windows' `cmd.exe` doesn't do that. For consistent cross-platform behavior, this crate emulates Unix-like expansion on Windows. You only need to use `wild::args()` instead of `std::env::args()`.
-It is more robust than using [`glob()`](https://crates.rs/crates/glob) on values from `std::env::args()`, because this crate is aware of argument quoting, and special characteres in quotes (`"*"`) are intentionally not expanded.
+It is more robust than using [`glob()`](https://lib.rs/crates/glob) on values from `std::env::args()`, because this crate is aware of argument quoting, and special characteres in quotes (`"*"`) are intentionally not expanded.
The glob syntax on Windows is limited to `*`, `?`, and `[a-z]`/`[!a-z]` ranges, as supported by the glob crate. Parsing of quoted arguments precisely follows Windows' native syntax ([`CommandLineToArgvW`][1], specifically).
@@ -14,16 +14,19 @@ The glob syntax on Windows is limited to `*`, `?`, and `[a-z]`/`[!a-z]` ranges,
`wild::args()` is a drop-in replacement for `std::env::args()`.
-```rust
-extern crate wild;
+```toml
+[dependencies]
+wild = "2"
+```
+```rust
fn main() {
let args = wild::args();
println!("The args are: {:?}", args.collect::<Vec<_>>());
}
```
-## Usage with [Clap](https://crates.rs/crates/clap)
+## Usage with [Clap](https://lib.rs/crates/clap)
```rust
let matches = clap::App::new("your_app")
diff --git a/debian/cargo-checksum.json b/debian/cargo-checksum.json
index c08973a457a505e44f50155a0a88e2fd676180fa..60d032f7651ebab3db6ee8f0b6b92dbfc75872af 100644
--- a/debian/cargo-checksum.json
+++ b/debian/cargo-checksum.json
@@ -1 +1 @@
-{"package":"97d34fecce28871e5c0e059deae21ef7f7d13b98a5964b24c58b3735c8052fc8","files":{}}
+{"package":"05b116685a6be0c52f5a103334cbff26db643826c7b3735fc0a3ba9871310a74","files":{}}
diff --git a/debian/changelog b/debian/changelog
index 311b1a4fbc114ff82a5af94d58f95995b7fd4cee..c78c92f9746ffe9906a079e4b5ced05d276d53d5 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,10 @@
+rust-wild (2.1.0-1) unstable; urgency=medium
+
+ * Team upload.
+ * Package wild 2.1.0 from crates.io using debcargo 2.5.0
+
+ -- Sylvestre Ledru <sylvestre@debian.org> Wed, 14 Sep 2022 09:24:42 +0200
+
rust-wild (2.0.2-1) unstable; urgency=medium
* Team upload.
diff --git a/debian/compat b/debian/compat
index b4de3947675361a7770d29b8982c407b0ec6b2a0..48082f72f087ce7e6fa75b9c41d7387daecd447b 100644
--- a/debian/compat
+++ b/debian/compat
@@ -1 +1 @@
-11
+12
diff --git a/debian/control b/debian/control
index 1527a196156287f99106e955a3456acfddb61802..fedc36f8396c252f40501e409b4b1bb2e8f26a34 100644
--- a/debian/control
+++ b/debian/control
@@ -1,8 +1,8 @@
Source: rust-wild
Section: rust
Priority: optional
-Build-Depends: debhelper (>= 11),
- dh-cargo (>= 15),
+Build-Depends: debhelper (>= 12),
+ dh-cargo (>= 25),
cargo:native <!nocheck>,
rustc:native <!nocheck>,
libstd-rust-dev <!nocheck>,
@@ -10,10 +10,11 @@ Build-Depends: debhelper (>= 11),
Maintainer: Debian Rust Maintainers <pkg-rust-maintainers@alioth-lists.debian.net>
Uploaders:
Helen Koike <helen@koikeco.de>
-Standards-Version: 4.2.0
+Standards-Version: 4.5.1
Vcs-Git: https://salsa.debian.org/rust-team/debcargo-conf.git [src/wild]
Vcs-Browser: https://salsa.debian.org/rust-team/debcargo-conf/tree/master/src/wild
-Homepage: https://crates.rs/crates/wild
+Homepage: https://lib.rs/crates/wild
+Rules-Requires-Root: no
Package: librust-wild-dev
Architecture: any
@@ -25,10 +26,10 @@ Provides:
librust-wild+default-dev (= ${binary:Version}),
librust-wild-2-dev (= ${binary:Version}),
librust-wild-2+default-dev (= ${binary:Version}),
- librust-wild-2.0-dev (= ${binary:Version}),
- librust-wild-2.0+default-dev (= ${binary:Version}),
- librust-wild-2.0.2-dev (= ${binary:Version}),
- librust-wild-2.0.2+default-dev (= ${binary:Version})
+ librust-wild-2.1-dev (= ${binary:Version}),
+ librust-wild-2.1+default-dev (= ${binary:Version}),
+ librust-wild-2.1.0-dev (= ${binary:Version}),
+ librust-wild-2.1.0+default-dev (= ${binary:Version})
Description: Glob (wildcard) expanded command-line arguments on Windows - Rust source code
This package contains the source for the Rust wild crate, packaged by debcargo
for use with cargo and dh-cargo.
diff --git a/debian/copyright.debcargo.hint b/debian/copyright.debcargo.hint
index f09c43763439118c520c6a92e3a149899c615100..16392124e57d99ad11f0f4b85dd3cc9c10ffd35c 100644
--- a/debian/copyright.debcargo.hint
+++ b/debian/copyright.debcargo.hint
@@ -5,7 +5,7 @@ Source: https://gitlab.com/kornelski/wild
Files: *
Copyright: FIXME (overlay) UNKNOWN-YEARS Kornel <kornel@geekhood.net>
-License: MIT
+License: Apache-2.0 or MIT
Comment:
FIXME (overlay): Since upstream copyright years are not available in
Cargo.toml, they were extracted from the upstream Git repository. This may not
@@ -21,9 +21,13 @@ Comment:
Files: debian/*
Copyright:
- 2019 Debian Rust Maintainers <pkg-rust-maintainers@alioth-lists.debian.net>
- 2019 Helen Koike <helen@koikeco.de>
-License: MIT
+ 2019-2022 Debian Rust Maintainers <pkg-rust-maintainers@alioth-lists.debian.net>
+ 2019-2022 Helen Koike <helen@koikeco.de>
+License: Apache-2.0 or MIT
+
+License: Apache-2.0
+ Debian systems provide the Apache 2.0 license in
+ /usr/share/common-licenses/Apache-2.0
License: MIT
Permission is hereby granted, free of charge, to any person obtaining a copy
diff --git a/debian/tests/control b/debian/tests/control
new file mode 100644
index 0000000000000000000000000000000000000000..d6d84dadb7109f321eb8a989bc53f6f4c1a45ac9
--- /dev/null
+++ b/debian/tests/control
@@ -0,0 +1,14 @@
+Test-Command: /usr/share/cargo/bin/cargo-auto-test wild 2.1.0 --all-targets --all-features
+Features: test-name=rust-wild:@
+Depends: dh-cargo (>= 18), librust-glob-0.3+default-dev, @
+Restrictions: allow-stderr, skip-not-installable
+
+Test-Command: /usr/share/cargo/bin/cargo-auto-test wild 2.1.0 --all-targets
+Features: test-name=librust-wild-dev:default
+Depends: dh-cargo (>= 18), librust-glob-0.3+default-dev, @
+Restrictions: allow-stderr, skip-not-installable
+
+Test-Command: /usr/share/cargo/bin/cargo-auto-test wild 2.1.0 --all-targets --no-default-features
+Features: test-name=librust-wild-dev:
+Depends: dh-cargo (>= 18), librust-glob-0.3+default-dev, @
+Restrictions: allow-stderr, skip-not-installable
diff --git a/debian/watch b/debian/watch
index d75195bd28c291ca9679638983970ad1ab45286f..ae23bf0cd9c04166b8143e7659909f34837603e8 100644
--- a/debian/watch
+++ b/debian/watch
@@ -2,4 +2,3 @@ version=4
opts=filenamemangle=s/.*\/(.*)\/download/wild-$1\.tar\.gz/g,\
uversionmangle=s/(\d)[_\.\-\+]?((RC|rc|pre|dev|beta|alpha)\d*)$/$1~$2/ \
https://qa.debian.org/cgi-bin/fakeupstream.cgi?upstream=crates.io/wild .*/crates/wild/@ANY_VERSION@/download
-
diff --git a/src/argsiter.rs b/src/argsiter.rs
index bbe17ed5b87a4aee6b3f2f55e9d36c5b5dd163c0..e9539335c7061112f0ee8cfb379acc63cfca08d0 100644
--- a/src/argsiter.rs
+++ b/src/argsiter.rs
@@ -1,6 +1,5 @@
-use globiter::*;
+use crate::globiter::GlobArgs;
use std::ffi::OsString;
-use glob;
use std::fmt;
/// Windows replacement for `std::env::ArgsOs`
@@ -75,13 +74,13 @@ impl Iterator for ArgsOs {
}
impl fmt::Debug for Args {
- fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
self.iter.fmt(f)
}
}
impl fmt::Debug for ArgsOs {
- fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
self.args.as_ref().map(|args| args.fmt(f))
.unwrap_or_else(|| "".fmt(f))
}
diff --git a/src/globiter.rs b/src/globiter.rs
index 037cb29cdf70978e009af72644ebd413a6e1587e..8a0f055137c1ec0a1a342c3f494c8bac103e2374 100644
--- a/src/globiter.rs
+++ b/src/globiter.rs
@@ -1,5 +1,6 @@
+use crate::parser::CommandLineWParser;
+use crate::parser::CharCode;
use std::ffi::OsString;
-use parser;
use std::fmt;
pub(crate) struct ArgOs {
@@ -10,13 +11,13 @@ pub(crate) struct ArgOs {
/// Iterator retuning glob-escaped arguments. Call `args()` to obtain it.
#[must_use]
-pub(crate) struct GlobArgs<'a> {
- line: &'a [u16],
+pub(crate) struct GlobArgs<'argsline> {
+ parser: CommandLineWParser<'argsline>,
}
impl<'a> fmt::Debug for GlobArgs<'a> {
- fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
- String::from_utf16_lossy(self.line).fmt(f)
+ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+ self.parser.fmt(f)
}
}
@@ -37,37 +38,53 @@ impl LossyOsStringExt for OsString {}
impl<'a> Iterator for GlobArgs<'a> {
type Item = ArgOs;
fn next(&mut self) -> Option<Self::Item> {
- let state = (vec![], vec![], false);
- let (state, rest) = parser::next_arg(self.line, state, |&mut (ref mut arg, ref mut text, ref mut contains_glob), c, quoted| {
+ let mut pattern = vec![];
+ let mut text = vec![];
+ let mut contains_glob = false;
+ let has_arg = self.parser.accumulate_next(|c| {
+ let (quoted, c) = match c {
+ CharCode::Quoted(c) => (true, c),
+ CharCode::Unquoted(c) => (false, c),
+ };
text.push(c);
- match c as u8 {
- b'?' | b'*' | b'[' | b']' if c < 256 => {
+ const Q: u16 = b'?' as u16;
+ const A: u16 = b'*' as u16;
+ const L: u16 = b'[' as u16;
+ const R: u16 = b']' as u16;
+ match c {
+ Q | A | L | R => {
if quoted {
- arg.push(u16::from(b'['));
- arg.push(c);
- arg.push(u16::from(b']'));
+ pattern.extend([
+ u16::from(b'['),
+ c,
+ u16::from(b']'),
+ ].iter().copied());
} else {
- arg.push(c);
- *contains_glob = true;
+ pattern.push(c);
+ contains_glob = true;
}
},
- _ => arg.push(c),
+ _ => pattern.push(c),
};
});
- self.line = rest;
- state.map(|(pattern, text, contains_glob)| ArgOs {
- pattern: OsString::from_wide(&pattern),
- text: OsString::from_wide(&text),
- contains_glob,
- })
+ if has_arg {
+ Some(ArgOs {
+ pattern: OsString::from_wide(&pattern),
+ text: OsString::from_wide(&text),
+ contains_glob,
+ })
+ } else {
+ None
+ }
}
}
-impl<'a> GlobArgs<'a> {
+impl<'argsline> GlobArgs<'argsline> {
/// UTF-16/UCS2 string from `GetCommandLineW`
#[allow(dead_code)]
- pub(crate) fn new(line: &'a [u16]) -> Self {
- Self { line }
+ pub(crate) fn new(command_line_args_ucs2: &'argsline [u16]) -> Self {
+ Self {
+ parser: CommandLineWParser::new(command_line_args_ucs2),
+ }
}
}
-
diff --git a/src/lib.rs b/src/lib.rs
index ffd980a441da83cc4980b6fd37c485ee8ebe6406..7d6be438a01d3eb595b6e57256d1f91a54a35a0a 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -1,3 +1,4 @@
+#![cfg_attr(docsrs, feature(doc_cfg))]
//! Emulates glob (wildcard) argument expansion on Windows. No-op on other platforms.
//!
//! Unix shells expand command-line arguments like `a*`, `file.???` and pass them expanded to applications.
@@ -14,29 +15,28 @@
//!
//! Use `wild::args()` instead of `std::env::args()` (or `wild::args_os()` instead of `std::env::args_os()`).
//!
-//! If you use [Clap](https://crates.rs/crates/clap), use `.get_matches_from(wild::args())` instead of `.get_matches()`.
+//! If you use [Clap](https://lib.rs/crates/clap), use `.get_matches_from(wild::args())` instead of `.get_matches()`.
-#[cfg(any(test,windows))]
-extern crate glob;
+/// An optional, experimental low-level interface for parsing command-line strings from other sources. In most cases [`args`] and [`args_os`] are more appropriate.
+#[cfg(any(test, windows))]
+pub mod parser;
-#[cfg(any(test,windows))]
-mod parser;
+#[cfg(any(test, windows))]
+mod globiter;
-#[cfg(any(test,windows))]
+#[cfg(any(test, windows))]
mod argsiter;
#[cfg(windows)]
-pub use argsiter::*;
-
-#[cfg(any(test,windows))]
-mod globiter;
+pub use crate::argsiter::*;
-/// Iterator of arguments. Equivalent to `std::env::Args`. See `args()` for details.
+/// Iterator of arguments. Equivalent to `std::env::Args`. See [`args`] for details.
///
/// On unix it's an alias for `std::env::Args`.
/// On Windows it's a custom iterator that implements glog expansion.
#[cfg(not(windows))]
pub type Args = std::env::Args;
+/// Same as [`Args`], but keeps invalid Unicode intact.
#[cfg(not(windows))]
pub type ArgsOs = std::env::ArgsOs;
@@ -50,11 +50,13 @@ pub type ArgsOs = std::env::ArgsOs;
/// the file system as it parses. This allows reading potentially huge lists of
/// filenames, but it's not an atomic snapshot (use `.collect()` if you need that).
#[cfg(not(windows))]
+#[must_use]
pub fn args() -> Args {
std::env::args()
}
#[cfg(not(windows))]
+#[must_use]
pub fn args_os() -> ArgsOs {
std::env::args_os()
}
@@ -69,14 +71,14 @@ pub fn args_os() -> ArgsOs {
/// On non-Windows platforms it returns `env::args()` as-is,
/// assuming expansion has already been done by the shell.
#[cfg(windows)]
+#[must_use]
pub fn args() -> Args {
- Args {
- iter: args_os(),
- }
+ Args { iter: args_os() }
}
-/// Same as `args()`, but returns `OsString`
+/// Same as [`args`], but returns `OsString`
#[cfg(windows)]
+#[must_use]
pub fn args_os() -> ArgsOs {
ArgsOs {
args: globs(),
@@ -107,7 +109,7 @@ fn raw_command_line() -> Option<&'static [u16]> {
return None;
}
let mut len = 0;
- while *line_ptr.offset(len as isize) != 0 {
+ while *line_ptr.add(len) != 0 {
len += 1;
}
Some(std::slice::from_raw_parts(line_ptr, len))
diff --git a/src/parser.rs b/src/parser.rs
index fb82cd7ac4caeba98be0bf41ea745079aeed7b91..aa41c266e73771c5778d0b6ddb2a859b7f16a39a 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -1,3 +1,27 @@
+use std::fmt;
+
+/// An experimental, low-level access to each individual character of raw arguments.
+#[must_use]
+pub struct CommandLineWParser<'argsline> {
+ line: &'argsline [u16],
+}
+
+impl<'argsline> CommandLineWParser<'argsline> {
+ #[inline]
+ #[must_use]
+ pub fn new(command_line_args_ucs2: &'argsline [u16]) -> Self {
+ Self {
+ line: command_line_args_ucs2,
+ }
+ }
+}
+
+impl<'a> fmt::Debug for CommandLineWParser<'a> {
+ #[cold]
+ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+ String::from_utf16_lossy(self.line).fmt(f)
+ }
+}
#[derive(Debug)]
enum State {
@@ -7,101 +31,117 @@ enum State {
Backslashes(usize, bool),
}
+/// A single code unit, which may be UCS-2 or half-broken UTF-16. Not a character.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum CharCode {
+ /// This code unit was found inside quotes (it's just text)
+ Quoted(u16),
+ /// This code unit was found outside quotes (you could interpret it as a glob)
+ Unquoted(u16),
+}
+
/// Given UCS2/potentially-broken-UTF-16 string parses one argument, following
/// the absolutely bizarre quoting rules of `CommandLineToArgvW`, and returns
-/// parsed argument as well as a slice of the remaining arguments.
+/// a bool indicating whether there's anything moe left.
///
-/// Calling this repeatedly until rest is empty will parse all arguments.
+/// Calling this repeatedly until it returns false will parse all arguments.
///
-/// `arg` is an empty pre-allocated argument to be returned, and the callback adds a new code unit to it.
-/// The last callback argument is whether the unit was quoted or not.
+/// The callback is expected to accumulate code units itself.
///
/// This parses u16 code units, rather than code points.
/// This allows supporting unpaired surrogates and ensures they won't "eat" any control characters.
-pub fn next_arg<AddC, ArgVec>(line: &[u16], mut arg: ArgVec, push: AddC) -> (Option<ArgVec>, &[u16])
-where
- AddC: Fn(&mut ArgVec, u16, bool),
-{
- use self::State::*;
- let mut state = BetweenArgs;
- for (i, &cu) in line.iter().enumerate() {
- state = match state {
- BetweenArgs => match cu {
- c if c == u16::from(b' ') => BetweenArgs,
- c if c == u16::from(b'"') => InArg(true),
- c if c == u16::from(b'\\') => Backslashes(1, false),
- c => {
- push(&mut arg, c, false);
- InArg(false)
- },
- },
- InArg(quoted) => match cu {
- c if c == u16::from(b'\\') => Backslashes(1, quoted),
- c if quoted && c == u16::from(b'"') => OnQuote,
- c if !quoted && c == u16::from(b'"') => InArg(true),
- c if !quoted && c == u16::from(b' ') => {
- return (Some(arg), &line[i+1..]);
+impl<'argsline> CommandLineWParser<'argsline> {
+ pub fn accumulate_next<CharacterAccumulator>(&mut self, mut push: CharacterAccumulator) -> bool
+ where CharacterAccumulator: FnMut(CharCode)
+ {
+ use self::State::*;
+ let mut state = BetweenArgs;
+ for (i, &cu) in self.line.iter().enumerate() {
+ state = match state {
+ BetweenArgs => match cu {
+ c if c == u16::from(b' ') => BetweenArgs,
+ c if c == u16::from(b'"') => InArg(true),
+ c if c == u16::from(b'\\') => Backslashes(1, false),
+ c => {
+ push(CharCode::Unquoted(c));
+ InArg(false)
+ },
},
- c => {
- push(&mut arg, c, quoted);
- InArg(quoted)
- },
- },
- OnQuote => match cu {
- c if c == u16::from(b'"') => {
- // In quoted arg "" means literal quote and the end of the quoted string (but not arg)
- push(&mut arg, u16::from(b'"'), true);
- InArg(false)
- },
- c if c == u16::from(b' ') => {
- return (Some(arg), &line[i+1..]);
+ InArg(quoted) => match cu {
+ c if c == u16::from(b'\\') => Backslashes(1, quoted),
+ c if quoted && c == u16::from(b'"') => OnQuote,
+ c if !quoted && c == u16::from(b'"') => InArg(true),
+ c if !quoted && c == u16::from(b' ') => {
+ self.line = &self.line[i+1..];
+ return true;
+ },
+ c => {
+ push(if quoted { CharCode::Quoted(c) } else { CharCode::Unquoted(c) });
+ InArg(quoted)
+ },
},
- c => {
- push(&mut arg, c, false);
- InArg(false)
+ OnQuote => match cu {
+ c if c == u16::from(b'"') => {
+ // In quoted arg "" means literal quote and the end of the quoted string (but not arg)
+ push(CharCode::Quoted(u16::from(b'"')));
+ InArg(false)
+ },
+ c if c == u16::from(b' ') => {
+ self.line = &self.line[i+1..];
+ return true;
+ },
+ c => {
+ push(CharCode::Unquoted(c));
+ InArg(false)
+ },
},
- },
- Backslashes(count, quoted) => match cu {
- c if c == u16::from(b'\\') => Backslashes(count + 1, quoted),
- c if c == u16::from(b'"') => {
- // backslashes followed by a quotation mark are treated as pairs of protected backslashes
- for _ in 0..count/2 {
- push(&mut arg, u16::from(b'\\'), quoted);
- }
+ Backslashes(count, quoted) => match cu {
+ c if c == u16::from(b'\\') => Backslashes(count + 1, quoted),
+ c if c == u16::from(b'"') => {
+ // backslashes followed by a quotation mark are treated as pairs of protected backslashes
+ let b = if quoted { CharCode::Quoted(u16::from(b'\\')) } else { CharCode::Unquoted(u16::from(b'\\')) };
+ for _ in 0..count/2 {
+ push(b);
+ }
- if count & 1 != 0 {
- // An odd number of backslashes is treated as followed by a protected quotation mark.
- push(&mut arg, u16::from(b'"'), quoted);
- InArg(quoted)
- } else if quoted {
- // An even number of backslashes is treated as followed by a word terminator.
- return (Some(arg), &line[i+1..]);
- } else {
+ if count & 1 != 0 {
+ // An odd number of backslashes is treated as followed by a protected quotation mark.
+ let c = u16::from(b'"');
+ push(if quoted { CharCode::Quoted(c) } else { CharCode::Unquoted(c) });
+ InArg(quoted)
+ } else if quoted {
+ // An even number of backslashes is treated as followed by a word terminator.
+ self.line = &self.line[i+1..];
+ return true;
+ } else {
+ InArg(quoted)
+ }
+ },
+ c => {
+ // A string of backslashes not followed by a quotation mark has no special meaning.
+ let b = if quoted { CharCode::Quoted(u16::from(b'\\')) } else { CharCode::Unquoted(u16::from(b'\\')) };
+ for _ in 0..count {
+ push(b);
+ }
+ push(if quoted { CharCode::Quoted(c) } else { CharCode::Unquoted(c) });
InArg(quoted)
- }
+ },
},
- c => {
- // A string of backslashes not followed by a quotation mark has no special meaning.
- for _ in 0..count {
- push(&mut arg, u16::from(b'\\'), quoted);
- }
- push(&mut arg, c, quoted);
- InArg(quoted)
- },
- },
+ }
}
+ let arg = match state {
+ BetweenArgs => false,
+ OnQuote | InArg(..) => true,
+ Backslashes(count, quoted) => {
+ // A string of backslashes not followed by a quotation mark has no special meaning.
+ let b = if quoted { CharCode::Quoted(u16::from(b'\\')) } else { CharCode::Unquoted(u16::from(b'\\')) };
+ for _ in 0..count {
+ push(b);
+ }
+ true
+ },
+ };
+ self.line = &self.line[..0];
+ arg
}
- let arg = match state {
- BetweenArgs => None,
- OnQuote | InArg(..) => Some(arg),
- Backslashes(count, quoted) => {
- // A string of backslashes not followed by a quotation mark has no special meaning.
- for _ in 0..count {
- push(&mut arg, u16::from(b'\\'), quoted);
- }
-
- Some(arg)
- },
- };
- (arg, &line[..0])
}