Add context and normal diff modes

This commit is contained in:
Michael Howell
2021-03-09 16:23:46 -07:00
parent 2e84164d2f
commit e72ea046b7
15 changed files with 1526 additions and 36 deletions
+3
View File
@@ -1,3 +1,6 @@
/target
/lib/normal-diff/target
/lib/context-diff/target
/lib/unified-diff/target
Cargo.lock
*.swp
+6 -14
View File
@@ -1,15 +1,7 @@
[package]
name = "unified-diff"
version = "0.2.1"
authors = [
"Michael Howell <michael@notriddle.com>",
"The Rust Project Developers"
[workspace]
members = [
"lib/unified-diff",
"lib/context-diff",
"lib/normal-diff",
"bin/diff",
]
edition = "2018"
description = "An implementation of the GNU unified diff format"
license = "MIT OR Apache-2.0"
repository = "https://github.com/notriddle/rust-unified-diff"
exclude = [ "fuzz" ]
[dependencies]
diff = "0.1.10"
+39 -22
View File
@@ -1,17 +1,34 @@
A GNU unified diff generator. Oracle tested against GNU patch 2.7.6
A package (currently just `diff`, but eventually more) of programs related to finding differences between files.
Based on the incomplete diff generator in https://github.com/rust-lang/rust/blob/master/src/tools/compiletest/src/runtest.rs,
but it implements a different format.
Based on the incomplete diff generator in https://github.com/rust-lang/rust/blob/master/src/tools/compiletest/src/runtest.rs, and made to be compatible with GNU's diff and patch tools.
```
~/unified-diff$ cargo run Cargo.lock Cargo.toml
~/diffutils$ cargo run -- -u3 Cargo.lock Cargo.toml
Finished dev [unoptimized + debuginfo] target(s) in 0.00s
Running `target/debug/unified-diff Cargo.lock Cargo.toml`
Running `target/debug/diff -u3 Cargo.lock Cargo.toml`
--- Cargo.lock
+++ Cargo.toml
@@ -1,14 +1,14 @@
@@ -1,39 +1,7 @@
-# This file is automatically @generated by Cargo.
-# It is not intended for manual editing.
-version = 3
-
-[[package]]
-name = "context-diff"
-version = "0.1.0"
-dependencies = [
- "diff 0.1.12",
-]
-
-[[package]]
-name = "diff"
-version = "0.1.0"
-dependencies = [
- "context-diff",
- "normal-diff",
- "unified-diff",
-]
-
-[[package]]
-name = "diff"
-version = "0.1.12"
@@ -19,23 +36,23 @@ but it implements a different format.
-checksum = "0e25ea47919b1560c4e3b7fe0aaab9becf5b84a10325ddf7db0f0ba5e1026499"
-
-[[package]]
+[package]
name = "unified-diff"
version = "0.1.0"
-name = "normal-diff"
-version = "0.1.0"
-dependencies = [
- "diff",
+authors = [
+ "Michael Howell <michael@notriddle.com>",
+ "The Rust Project Developers"
- "diff 0.1.12",
-]
-
-[[package]]
-name = "unified-diff"
-version = "0.3.0"
-dependencies = [
- "diff 0.1.12",
+[workspace]
+members = [
+ "lib/unified-diff",
+ "lib/context-diff",
+ "lib/normal-diff",
+ "bin/diff",
]
+edition = "2018"
+
+[[bin]]
+name = "unified-diff"
+
+[dependencies]
+diff = "0.1.10"
~/unified-diff$ rustup override set nightly
~/unified-diff$ cargo fuzz run fuzz_patch
```
+12
View File
@@ -0,0 +1,12 @@
[package]
name = "diff"
version = "0.1.0"
authors = ["Michael Howell <michael@notriddle.com>"]
edition = "2018"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
context-diff = { path = "../../lib/context-diff" }
normal-diff = { path = "../../lib/normal-diff" }
unified-diff = { path = "../../lib/unified-diff" }
+50
View File
@@ -0,0 +1,50 @@
use crate::params::*;
use std::env;
use std::fs;
use std::io::{self, Write};
mod params;
fn main() -> Result<(), String> {
let opts = env::args_os();
let Params {
from,
to,
context_count,
format,
} = parse_params(opts)?;
// read files
let from_content = match fs::read(&from) {
Ok(from_content) => from_content,
Err(e) => {
return Err(format!("Failed to read from-file: {}", e));
}
};
let to_content = match fs::read(&to) {
Ok(to_content) => to_content,
Err(e) => {
return Err(format!("Failed to read from-file: {}", e));
}
};
// run diff
let result: Vec<u8> = match format {
Format::Normal => normal_diff::diff(&from_content, &to_content),
Format::Unified => unified_diff::diff(
&from_content,
&from.to_string_lossy(),
&to_content,
&to.to_string_lossy(),
context_count,
),
Format::Context => context_diff::diff(
&from_content,
&from.to_string_lossy(),
&to_content,
&to.to_string_lossy(),
context_count,
),
};
io::stdout().write_all(&result).unwrap();
Ok(())
}
+232
View File
@@ -0,0 +1,232 @@
use std::ffi::{OsStr, OsString};
#[derive(Clone, Copy, Debug, Eq, PartialEq)]
pub enum Format {
Normal,
Unified,
Context,
}
#[cfg(unix)]
fn osstr_bytes(osstr: &OsStr) -> &[u8] {
use std::os::unix::ffi::OsStrExt;
osstr.as_bytes()
}
#[cfg(not(unix))]
fn osstr_bytes(osstr: &OsStr) -> Vec<u8> {
osstr.to_string_lossy().bytes().collect()
}
#[derive(Clone, Debug, Eq, PartialEq)]
pub struct Params {
pub from: OsString,
pub to: OsString,
pub format: Format,
pub context_count: usize,
}
pub fn parse_params<I: IntoIterator<Item = OsString>>(opts: I) -> Result<Params, String> {
let mut opts = opts.into_iter();
// parse CLI
let exe = match opts.next() {
Some(from) => from,
None => {
return Err(format!("Usage: <exe> <from> <to>"));
}
};
let mut from = None;
let mut to = None;
let mut format = None;
let mut context_count = 3;
while let Some(param) = opts.next() {
if param == "--" {
break;
}
if param == "-" {
if from.is_none() {
from = Some(OsString::from("/dev/stdin"));
} else if to.is_none() {
to = Some(OsString::from("/dev/stdin"));
} else {
return Err(format!("Usage: {} <from> <to>", exe.to_string_lossy()));
}
continue;
}
let p = osstr_bytes(&param);
if p.get(0) == Some(&b'-') && p.get(1) != Some(&b'-') {
let mut bit = p[1..].into_iter().copied().peekable();
// Can't use a for loop because `diff -30u` is supposed to make a diff
// with 30 lines of context.
while let Some(b) = bit.next() {
match b {
b'0'..=b'9' => {
context_count = (b - b'0') as usize;
while let Some(b'0'..=b'9') = bit.peek() {
context_count = context_count * 10;
context_count += (bit.next().unwrap() - b'0') as usize;
}
}
b'c' => {
if format.is_some() && format != Some(Format::Context) {
return Err(format!("Conflicting output style options"));
}
format = Some(Format::Context);
}
b'u' => {
if format.is_some() && format != Some(Format::Unified) {
return Err(format!("Conflicting output style options"));
}
format = Some(Format::Unified);
}
b'U' => {
if format.is_some() && format != Some(Format::Unified) {
return Err(format!("Conflicting output style options"));
}
format = Some(Format::Unified);
let context_count_maybe = if bit.peek().is_some() {
String::from_utf8(bit.collect::<Vec<u8>>()).ok()
} else {
opts.next().map(|x| x.to_string_lossy().into_owned())
};
if let Some(context_count_maybe) =
context_count_maybe.and_then(|x| x.parse().ok())
{
context_count = context_count_maybe;
break;
} else {
return Err(format!("Invalid context count"));
}
}
_ => return Err(format!("Unknown option: {}", String::from_utf8_lossy(&[b]))),
}
}
} else if from.is_none() {
from = Some(param);
} else if to.is_none() {
to = Some(param)
} else {
return Err(format!("Usage: {} <from> <to>", exe.to_string_lossy()));
}
}
let from = if let Some(from) = from {
from
} else if let Some(param) = opts.next() {
param
} else {
return Err(format!("Usage: {} <from> <to>", exe.to_string_lossy()));
};
let to = if let Some(to) = to {
to
} else if let Some(param) = opts.next() {
param
} else {
return Err(format!("Usage: {} <from> <to>", exe.to_string_lossy()));
};
let format = format.unwrap_or(Format::Normal);
Ok(Params {
from,
to,
format,
context_count,
})
}
#[cfg(test)]
mod tests {
use super::*;
fn os(s: &str) -> OsString {
OsString::from(s)
}
#[test]
fn basics() {
assert_eq!(
Ok(Params {
from: os("foo"),
to: os("bar"),
format: Format::Normal,
context_count: 3,
}),
parse_params([os("diff"), os("foo"), os("bar")].iter().cloned())
);
}
#[test]
fn context_count() {
assert_eq!(
Ok(Params {
from: os("foo"),
to: os("bar"),
format: Format::Unified,
context_count: 54,
}),
parse_params(
[os("diff"), os("-u54"), os("foo"), os("bar")]
.iter()
.cloned()
)
);
assert_eq!(
Ok(Params {
from: os("foo"),
to: os("bar"),
format: Format::Unified,
context_count: 54,
}),
parse_params(
[os("diff"), os("-U54"), os("foo"), os("bar")]
.iter()
.cloned()
)
);
assert_eq!(
Ok(Params {
from: os("foo"),
to: os("bar"),
format: Format::Unified,
context_count: 54,
}),
parse_params(
[os("diff"), os("-U"), os("54"), os("foo"), os("bar")]
.iter()
.cloned()
)
);
assert_eq!(
Ok(Params {
from: os("foo"),
to: os("bar"),
format: Format::Context,
context_count: 54,
}),
parse_params(
[os("diff"), os("-c54"), os("foo"), os("bar")]
.iter()
.cloned()
)
);
}
#[test]
fn double_dash() {
assert_eq!(
Ok(Params {
from: os("-g"),
to: os("-h"),
format: Format::Normal,
context_count: 3,
}),
parse_params([os("diff"), os("--"), os("-g"), os("-h")].iter().cloned())
);
}
#[test]
fn unknown_argument() {
assert!(
parse_params([os("diff"), os("-g"), os("foo"), os("bar")].iter().cloned()).is_err()
);
assert!(parse_params([os("diff"), os("-g"), os("bar")].iter().cloned()).is_err());
assert!(parse_params([os("diff"), os("-g")].iter().cloned()).is_err());
}
#[test]
fn empty() {
assert!(parse_params([].iter().cloned()).is_err());
}
}
+15
View File
@@ -0,0 +1,15 @@
[package]
name = "context-diff"
version = "0.1.0"
authors = [
"Michael Howell <michael@notriddle.com>",
"The Rust Project Developers"
]
edition = "2018"
description = "An implementation of the GNU unified diff format"
license = "MIT OR Apache-2.0"
repository = "https://github.com/notriddle/diffutils"
exclude = [ "fuzz" ]
[dependencies]
diff = "0.1.10"
File diff suppressed because it is too large Load Diff
+15
View File
@@ -0,0 +1,15 @@
[package]
name = "normal-diff"
version = "0.1.0"
authors = [
"Michael Howell <michael@notriddle.com>",
"The Rust Project Developers"
]
edition = "2018"
description = "An implementation of the GNU unified diff format"
license = "MIT OR Apache-2.0"
repository = "https://github.com/notriddle/diffutils"
exclude = [ "fuzz" ]
[dependencies]
diff = "0.1.10"
+362
View File
@@ -0,0 +1,362 @@
use std::io::Write;
#[derive(Debug, PartialEq)]
struct Mismatch {
pub line_number_expected: usize,
pub line_number_actual: usize,
pub expected: Vec<Vec<u8>>,
pub actual: Vec<Vec<u8>>,
}
impl Mismatch {
fn new(line_number_expected: usize, line_number_actual: usize) -> Mismatch {
Mismatch {
line_number_expected,
line_number_actual,
expected: Vec::new(),
actual: Vec::new(),
}
}
}
// Produces a diff between the expected output and actual output.
fn make_diff(expected: &[u8], actual: &[u8]) -> Vec<Mismatch> {
let mut line_number_expected = 1;
let mut line_number_actual = 1;
let mut results = Vec::new();
let mut mismatch = Mismatch::new(line_number_expected, line_number_actual);
let mut expected_lines: Vec<&[u8]> = expected.split(|&c| c == b'\n').collect();
let mut actual_lines: Vec<&[u8]> = actual.split(|&c| c == b'\n').collect();
debug_assert_eq!(b"".split(|&c| c == b'\n').count(), 1);
// ^ means that underflow here is impossible
let _expected_lines_count = expected_lines.len() - 1;
let _actual_lines_count = actual_lines.len() - 1;
if expected_lines.last() == Some(&&b""[..]) {
expected_lines.pop();
}
if actual_lines.last() == Some(&&b""[..]) {
actual_lines.pop();
}
for result in diff::slice(&expected_lines, &actual_lines) {
match result {
diff::Result::Left(str) => {
if mismatch.actual.len() != 0 {
results.push(mismatch);
mismatch = Mismatch::new(line_number_expected, line_number_actual);
}
mismatch.expected.push(str.to_vec());
line_number_expected += 1;
}
diff::Result::Right(str) => {
mismatch.actual.push(str.to_vec());
line_number_actual += 1;
}
diff::Result::Both(_str, _) => {
line_number_expected += 1;
line_number_actual += 1;
if mismatch.actual.len() != 0 || mismatch.expected.len() != 0 {
results.push(mismatch);
mismatch = Mismatch::new(line_number_expected, line_number_actual);
} else {
mismatch.line_number_expected = line_number_expected;
mismatch.line_number_actual = line_number_actual;
}
}
}
}
if mismatch.actual.len() != 0 || mismatch.expected.len() != 0 {
results.push(mismatch);
}
results
}
pub fn diff(expected: &[u8], actual: &[u8]) -> Vec<u8> {
let mut output = Vec::new();
let diff_results = make_diff(expected, actual);
for result in diff_results {
let line_number_expected = result.line_number_expected;
let line_number_actual = result.line_number_actual;
let expected_count = result.expected.len();
let actual_count = result.actual.len();
match (expected_count, actual_count) {
(0, 0) => unreachable!(),
(0, _) => writeln!(
&mut output,
"{}a{},{}",
line_number_expected - 1,
line_number_actual,
line_number_actual + actual_count - 1
)
.unwrap(),
(_, 0) => writeln!(
&mut output,
"{},{}d{}",
line_number_expected,
expected_count + line_number_expected - 1,
line_number_actual - 1
)
.unwrap(),
_ => writeln!(
&mut output,
"{},{}c{},{}",
line_number_expected,
expected_count + line_number_expected - 1,
line_number_actual,
actual_count + line_number_actual - 1
)
.unwrap(),
}
for expected in &result.expected {
write!(&mut output, "< ").unwrap();
output.write_all(expected).unwrap();
writeln!(&mut output, "").unwrap();
}
if expected_count != 0 && actual_count != 0 {
writeln!(&mut output, "---").unwrap();
}
for actual in &result.actual {
write!(&mut output, "> ").unwrap();
output.write_all(actual).unwrap();
writeln!(&mut output, "").unwrap();
}
}
output
}
#[test]
fn test_permutations() {
// test all possible six-line files.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
for &d in &[0, 1, 2] {
for &e in &[0, 1, 2] {
for &f in &[0, 1, 2] {
use std::fs::{self, File};
use std::io::Write;
use std::process::Command;
let mut alef = Vec::new();
let mut bet = Vec::new();
alef.write_all(if a == 0 { b"a\n" } else { b"b\n" })
.unwrap();
if a != 2 {
bet.write_all(b"b\n").unwrap();
}
alef.write_all(if b == 0 { b"c\n" } else { b"d\n" })
.unwrap();
if b != 2 {
bet.write_all(b"d\n").unwrap();
}
alef.write_all(if c == 0 { b"e\n" } else { b"f\n" })
.unwrap();
if c != 2 {
bet.write_all(b"f\n").unwrap();
}
alef.write_all(if d == 0 { b"g\n" } else { b"h\n" })
.unwrap();
if d != 2 {
bet.write_all(b"h\n").unwrap();
}
alef.write_all(if e == 0 { b"i\n" } else { b"j\n" })
.unwrap();
if e != 2 {
bet.write_all(b"j\n").unwrap();
}
alef.write_all(if f == 0 { b"k\n" } else { b"l\n" })
.unwrap();
if f != 2 {
bet.write_all(b"l\n").unwrap();
}
// This test diff is intentionally reversed.
// We want it to turn the alef into bet.
let diff = diff(&alef, &bet);
File::create("target/ab.diff")
.unwrap()
.write_all(&diff)
.unwrap();
let mut fa = File::create("target/alef").unwrap();
fa.write_all(&alef[..]).unwrap();
let mut fb = File::create("target/bet").unwrap();
fb.write_all(&bet[..]).unwrap();
let _ = fa;
let _ = fb;
let output = Command::new("patch")
.arg("-p0")
.arg("target/alef")
.stdin(File::open("target/ab.diff").unwrap())
.output()
.unwrap();
if !output.status.success() {
panic!("{:?}", output);
}
//println!("{}", String::from_utf8_lossy(&output.stdout));
//println!("{}", String::from_utf8_lossy(&output.stderr));
let alef = fs::read("target/alef").unwrap();
assert_eq!(alef, bet);
}
}
}
}
}
}
}
#[test]
fn test_permutations_empty_lines() {
// test all possible six-line files with missing newlines.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
for &d in &[0, 1, 2] {
for &e in &[0, 1, 2] {
for &f in &[0, 1, 2] {
use std::fs::{self, File};
use std::io::Write;
use std::process::Command;
let mut alef = Vec::new();
let mut bet = Vec::new();
alef.write_all(if a == 0 { b"\n" } else { b"b\n" }).unwrap();
if a != 2 {
bet.write_all(b"b\n").unwrap();
}
alef.write_all(if b == 0 { b"\n" } else { b"d\n" }).unwrap();
if b != 2 {
bet.write_all(b"d\n").unwrap();
}
alef.write_all(if c == 0 { b"\n" } else { b"f\n" }).unwrap();
if c != 2 {
bet.write_all(b"f\n").unwrap();
}
alef.write_all(if d == 0 { b"\n" } else { b"h\n" }).unwrap();
if d != 2 {
bet.write_all(b"h\n").unwrap();
}
alef.write_all(if e == 0 { b"\n" } else { b"j\n" }).unwrap();
if e != 2 {
bet.write_all(b"j\n").unwrap();
}
alef.write_all(if f == 0 { b"\n" } else { b"l\n" }).unwrap();
if f != 2 {
bet.write_all(b"l\n").unwrap();
}
// This test diff is intentionally reversed.
// We want it to turn the alef into bet.
let diff = diff(&alef, &bet);
File::create("target/ab_.diff")
.unwrap()
.write_all(&diff)
.unwrap();
let mut fa = File::create("target/alef_").unwrap();
fa.write_all(&alef[..]).unwrap();
let mut fb = File::create("target/bet_").unwrap();
fb.write_all(&bet[..]).unwrap();
let _ = fa;
let _ = fb;
let output = Command::new("patch")
.arg("-p0")
.arg("target/alef_")
.stdin(File::open("target/ab_.diff").unwrap())
.output()
.unwrap();
if !output.status.success() {
panic!("{:?}", output);
}
//println!("{}", String::from_utf8_lossy(&output.stdout));
//println!("{}", String::from_utf8_lossy(&output.stderr));
let alef = fs::read("target/alef_").unwrap();
assert_eq!(alef, bet);
}
}
}
}
}
}
}
#[test]
fn test_permutations_reverse() {
// test all possible six-line files.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
for &d in &[0, 1, 2] {
for &e in &[0, 1, 2] {
for &f in &[0, 1, 2] {
use std::fs::{self, File};
use std::io::Write;
use std::process::Command;
let mut alef = Vec::new();
let mut bet = Vec::new();
alef.write_all(if a == 0 { b"a\n" } else { b"f\n" })
.unwrap();
if a != 2 {
bet.write_all(b"a\n").unwrap();
}
alef.write_all(if b == 0 { b"b\n" } else { b"e\n" })
.unwrap();
if b != 2 {
bet.write_all(b"b\n").unwrap();
}
alef.write_all(if c == 0 { b"c\n" } else { b"d\n" })
.unwrap();
if c != 2 {
bet.write_all(b"c\n").unwrap();
}
alef.write_all(if d == 0 { b"d\n" } else { b"c\n" })
.unwrap();
if d != 2 {
bet.write_all(b"d\n").unwrap();
}
alef.write_all(if e == 0 { b"e\n" } else { b"b\n" })
.unwrap();
if e != 2 {
bet.write_all(b"e\n").unwrap();
}
alef.write_all(if f == 0 { b"f\n" } else { b"a\n" })
.unwrap();
if f != 2 {
bet.write_all(b"f\n").unwrap();
}
// This test diff is intentionally reversed.
// We want it to turn the alef into bet.
let diff = diff(&alef, &bet);
File::create("target/abr.diff")
.unwrap()
.write_all(&diff)
.unwrap();
let mut fa = File::create("target/alefr").unwrap();
fa.write_all(&alef[..]).unwrap();
let mut fb = File::create("target/betr").unwrap();
fb.write_all(&bet[..]).unwrap();
let _ = fa;
let _ = fb;
let output = Command::new("patch")
.arg("-p0")
.arg("target/alefr")
.stdin(File::open("target/abr.diff").unwrap())
.output()
.unwrap();
if !output.status.success() {
panic!("{:?}", output);
}
//println!("{}", String::from_utf8_lossy(&output.stdout));
//println!("{}", String::from_utf8_lossy(&output.stderr));
let alef = fs::read("target/alefr").unwrap();
assert_eq!(alef, bet);
}
}
}
}
}
}
}
+15
View File
@@ -0,0 +1,15 @@
[package]
name = "unified-diff"
version = "0.3.0"
authors = [
"Michael Howell <michael@notriddle.com>",
"The Rust Project Developers"
]
edition = "2018"
description = "An implementation of the GNU unified diff format"
license = "MIT OR Apache-2.0"
repository = "https://github.com/notriddle/diffutils"
exclude = [ "fuzz" ]
[dependencies]
diff = "0.1.10"
@@ -380,6 +380,7 @@ pub fn diff(
#[test]
fn test_permutations() {
// test all possible six-line files.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
@@ -457,6 +458,7 @@ fn test_permutations() {
#[test]
fn test_permutations_missing_line_ending() {
// test all possible six-line files with missing newlines.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
@@ -549,6 +551,7 @@ fn test_permutations_missing_line_ending() {
#[test]
fn test_permutations_empty_lines() {
// test all possible six-line files with missing newlines.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
@@ -636,6 +639,7 @@ fn test_permutations_empty_lines() {
#[test]
fn test_permutations_missing_lines() {
// test all possible six-line files.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {
@@ -707,6 +711,7 @@ fn test_permutations_missing_lines() {
#[test]
fn test_permutations_reverse() {
// test all possible six-line files.
let _ = std::fs::create_dir("target");
for &a in &[0, 1, 2] {
for &b in &[0, 1, 2] {
for &c in &[0, 1, 2] {