From 8e5aee0bbf01db91795305d7e4d34f66a261e6f7 Mon Sep 17 00:00:00 2001
From: xenofem <xenofem@xeno.science>
Date: Wed, 6 Mar 2024 21:13:17 -0500
Subject: [PATCH] bytediff

---
 Cargo.lock          |   7 +++
 Cargo.toml          |   1 +
 src/bin/bytediff.rs | 136 ++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 144 insertions(+)
 create mode 100644 src/bin/bytediff.rs

diff --git a/Cargo.lock b/Cargo.lock
index e692060..54b001e 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -59,6 +59,12 @@ dependencies = [
  "windows-sys",
 ]
 
+[[package]]
+name = "anyhow"
+version = "1.0.80"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5ad32ce52e4161730f7098c077cd2ed6229b5804ccf99e5366be1ab72a98b4e1"
+
 [[package]]
 name = "clap"
 version = "4.4.2"
@@ -109,6 +115,7 @@ checksum = "acbf1af155f9b9ef647e42cdc158db4b64a1b61f743629225fde6f3e0be2a7c7"
 name = "evenmoreutils"
 version = "0.2.0"
 dependencies = [
+ "anyhow",
  "clap",
  "lazy_static",
  "regex",
diff --git a/Cargo.toml b/Cargo.toml
index 72bdbd1..b938abb 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -6,6 +6,7 @@ authors = ["xenofem <xenofem@xeno.science>"]
 license = "MIT"
 
 [dependencies]
+anyhow = "1.0"
 clap = { version = "4.3", features = ["derive"] }
 lazy_static = "1.4"
 regex = "1.7"
diff --git a/src/bin/bytediff.rs b/src/bin/bytediff.rs
new file mode 100644
index 0000000..e63c600
--- /dev/null
+++ b/src/bin/bytediff.rs
@@ -0,0 +1,136 @@
+use std::fs::File;
+use std::io::{prelude::*, BufReader};
+
+use anyhow::{anyhow, Context, Result};
+
+struct ZipLonger<A: Iterator, B: Iterator> {
+    a: A,
+    b: B,
+}
+
+impl<A: Iterator, B: Iterator> Iterator for ZipLonger<A, B> {
+    type Item = (Option<A::Item>, Option<B::Item>);
+
+    fn next(&mut self) -> Option<Self::Item> {
+        let a_next = self.a.next();
+        let b_next = self.b.next();
+        if a_next.is_none() && b_next.is_none() {
+            return None;
+        }
+        Some((a_next, b_next))
+    }
+}
+
+impl<A: Iterator, B: Iterator> ZipLonger<A, B> {
+    fn new(a: A, b: B) -> Self {
+        Self { a, b }
+    }
+}
+
+struct Chunks<I: Iterator> {
+    n: usize,
+    inner: I,
+}
+
+impl<T, F, I: Iterator<Item = Result<T, F>>> Iterator for Chunks<I> {
+    type Item = Result<Vec<T>, F>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        let mut values = Vec::new();
+        for _ in 0..(self.n) {
+            let next = self.inner.next();
+            match next {
+                Some(Ok(v)) => {
+                    values.push(v);
+                }
+                Some(Err(e)) => {
+                    return Some(Err(e));
+                }
+                None => {
+                    break;
+                }
+            }
+        }
+        if values.is_empty() {
+            return None;
+        }
+        Some(Ok(values))
+    }
+}
+
+impl<I: Iterator> Chunks<I> {
+    fn new(inner: I, n: usize) -> Self {
+        Self { inner, n }
+    }
+}
+
+fn color_byte_display(this: &[u8], that: &[u8], color: u8) -> String {
+    let mut result = String::new();
+
+    for (i, b) in this.iter().enumerate() {
+        if i % 2 == 0 && i > 0 {
+            result.push(' ');
+        }
+
+        if that.get(i) == Some(b) {
+            result.push_str(&format!("{:02x}", b));
+        } else {
+            result.push_str(&format!("\x1b[1;{}m{:02x}\x1b[0m", color, b));
+        }
+    }
+
+    result
+}
+
+fn main() -> Result<()> {
+    let mut args = std::env::args();
+    let name = args
+        .next()
+        .ok_or_else(|| anyhow!("Missing executable name in args"))?;
+    let (a, b) = match (args.next(), args.next()) {
+        (Some(a), Some(b)) => (a, b),
+        _ => return Err(anyhow!("Usage: {} file file", name)),
+    };
+
+    let a_bytes = BufReader::new(
+        File::open(a.clone()).with_context(|| format!("Failed to open input file {}", a))?,
+    )
+    .bytes();
+    let b_bytes = BufReader::new(
+        File::open(b.clone()).with_context(|| format!("Failed to open input file {}", b))?,
+    )
+    .bytes();
+
+    let mut in_differing_region = false;
+
+    for (i, (rva, rvb)) in
+        ZipLonger::new(Chunks::new(a_bytes, 16), Chunks::new(b_bytes, 16)).enumerate()
+    {
+        let va: Vec<u8> = rva
+            .transpose()
+            .with_context(|| format!("Error reading from input file {}", a))?
+            .unwrap_or_default();
+        let vb: Vec<u8> = rvb
+            .transpose()
+            .with_context(|| format!("Error reading from input file {}", b))?
+            .unwrap_or_default();
+
+        if va == vb {
+            if in_differing_region {
+                in_differing_region = false;
+                println!("---");
+            }
+            continue;
+        }
+        in_differing_region = true;
+
+        println!(
+            "{0:08x}: {1}\n{0:08x}: {2}",
+            i * 16,
+            color_byte_display(&va, &vb, 31),
+            color_byte_display(&vb, &va, 32),
+        );
+    }
+
+    Ok(())
+}