implement better hashing

- call `file::Checked::hash` in `CacheFile::from_args` instead

.cargo/config.toml

@@ -1,2 +1,9 @@
 [build]
 target = "x86_64-unknown-linux-musl"
+
+# rustflags = [
+#   # emit instructions tuned to the current CPU
+#   "-C", "target-cpu=native",
+#   # assume CPU features
+#   "-C", "target-feature=+avx2,+sse4.1,+ssse3,+aes",
+# ]

Cargo.lock

@@ -67,6 +67,18 @@ dependencies = [
  "windows-sys 0.59.0",
 ]
 
+[[package]]
+name = "arrayref"
+version = "0.3.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "76a2e8124351fda1ef8aaaa3bbd7ebbcb486bbcd4225aca0aa0d84bb2db8fecb"
+
+[[package]]
+name = "arrayvec"
+version = "0.7.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7c02d123df017efcdfbd739ef81735b36c5ba83ec3c59c80a9d7ecc718f92e50"
+
 [[package]]
 name = "base64"
 version = "0.22.1"
@@ -86,21 +98,14 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1b8e56985ec62d17e9c1001dc89c88ecd7dc08e47eba5ec7c29c7b5eeecde967"
 
 [[package]]
-name = "blake2"
-version = "0.10.6"
+name = "blake2b_simd"
+version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "46502ad458c9a52b69d4d4d32775c788b7a1b85e8bc9d482d92250fc0e3f8efe"
+checksum = "06e903a20b159e944f91ec8499fe1e55651480c541ea0a584f5d967c49ad9d99"
 dependencies = [
- "digest",
-]
-
-[[package]]
-name = "block-buffer"
-version = "0.10.4"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3078c7629b62d3f0439517fa394996acacc5cbc91c5a20d8c658e77abd503a71"
-dependencies = [
- "generic-array",
+ "arrayref",
+ "arrayvec",
+ "constant_time_eq",
 ]
 
 [[package]]
@@ -195,6 +200,12 @@ dependencies = [
  "windows-sys 0.59.0",
 ]
 
+[[package]]
+name = "constant_time_eq"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7c74b8349d32d297c9134b8c88677813a227df8f779daa29bfc29c183fe3dca6"
+
 [[package]]
 name = "cookie"
 version = "0.18.1"
@@ -233,16 +244,6 @@ dependencies = [
  "cfg-if",
 ]
 
-[[package]]
-name = "crypto-common"
-version = "0.1.6"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1bfb12502f3fc46cca1bb51ac28df9d618d813cdc3d2f25b9fe775a34af26bb3"
-dependencies = [
- "generic-array",
- "typenum",
-]
-
 [[package]]
 name = "ctrlc"
 version = "3.4.7"
@@ -273,17 +274,6 @@ dependencies = [
  "thiserror 1.0.69",
 ]
 
-[[package]]
-name = "digest"
-version = "0.10.7"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9ed9a281f7bc9b7576e61468ba615a66a5c8cfdff42420a70aa82701a3b1e292"
-dependencies = [
- "block-buffer",
- "crypto-common",
- "subtle",
-]
-
 [[package]]
 name = "dirs-next"
 version = "2.0.0"
@@ -385,16 +375,6 @@ dependencies = [
  "percent-encoding",
 ]
 
-[[package]]
-name = "generic-array"
-version = "0.14.7"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "85649ca51fd72272d7821adaf274ad91c288277713d9c18820d8499a7ff69e9a"
-dependencies = [
- "typenum",
- "version_check",
-]
-
 [[package]]
 name = "getrandom"
 version = "0.2.16"
@@ -902,7 +882,7 @@ name = "shrupl"
 version = "0.1.0-alpha"
 dependencies = [
  "base64ct",
- "blake2",
+ "blake2b_simd",
  "clap",
  "console",
  "ctrlc",
@@ -1045,12 +1025,6 @@ dependencies = [
  "zerovec",
 ]
 
-[[package]]
-name = "typenum"
-version = "1.18.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1dccffe3ce07af9386bfd29e80c0ab1a8205a2fc34e4bcd40364df902cfa8f3f"
-
 [[package]]
 name = "unicode-ident"
 version = "1.0.18"

Cargo.toml

@@ -6,7 +6,7 @@ description = "ShrUpl is a tool to upload files to a Sharry Instance through a p
 
 [dependencies]
 base64ct = { version = "1.8.0", default-features = false, features = ["alloc"] }
-blake2 = { version = "0.10.6", default-features = false }
+blake2b_simd = "1.0.3"
 clap = { version = "4.5.38", features = ["derive"] }
 console = { version = "0.15.11", default-features = false }
 ctrlc = { version = "3.4.7", features = ["termination"] }
@@ -22,5 +22,11 @@ thiserror = "2.0.12"
 ureq = { version = "3.0.11", features = ["json"] }
 
 [profile.release]
+# Optimize for speed even more aggressively
+opt-level = "z"
+# better inlining
+codegen-units = 1
+# linker‐time optimization
 lto = true
+debug = false
 panic = "abort"

src/appstate.rs

@@ -1,13 +1,13 @@
 use std::{fmt, io, time::Duration};
 
-use console::style;
-use indicatif::{ProgressBar, ProgressDrawTarget, ProgressStyle};
+use indicatif::{ProgressBar, ProgressDrawTarget};
 use log::{debug, warn};
 
 use crate::{
     cachefile::CacheFile,
     cli::Cli,
     file::{Chunk, FileTrait},
+    output::new_progressbar,
     sharry::{self, Client},
 };
 
@@ -46,6 +46,8 @@ impl AppState {
             .inspect_err(|e| debug!("could not resume from hash {:?}: {e}", args.get_hash()))
             .ok()?;
 
+        inner.peek_uploading();
+
         Some(Self::new(new_http(args.get_timeout()), inner))
     }
 
@@ -54,23 +56,11 @@ impl AppState {
 
         let share_id = http.share_create(&args.get_uri(), &args.alias, args.get_share_request())?;
 
-        Ok(Self::new(http, CacheFile::from_args(args, share_id)))
+        Ok(Self::new(http, CacheFile::from_args(args, share_id)?))
     }
 
     fn with_progressbar(&mut self, f: impl FnOnce(&ProgressBar), drop_bar: bool) {
-        let bar = &*self.progress.get_or_insert_with(|| {
-            ProgressBar::hidden().with_style(
-                ProgressStyle::with_template(&format!(
-                    concat!(
-                        "{{bar:50.cyan/blue}} {{msg:.magenta}}: ",
-                        "{{binary_bytes:.yellow}}{}{{binary_total_bytes:.yellow}} ",
-                        "({{eta}})",
-                    ),
-                    style("/").magenta(),
-                ))
-                .expect("style template is not valid"),
-            )
-        });
+        let bar = &*self.progress.get_or_insert_with(new_progressbar);
 
         if let Some(upl) = self.inner.peek_uploading() {
             if bar.length().is_none() {
@@ -154,7 +144,7 @@ impl AppState {
             self.http
                 .share_create(&args.get_uri(), &args.alias, args.get_share_request())?;
 
-        Ok(Self::new(self.http, CacheFile::from_args(args, share_id)))
+        Ok(Self::new(self.http, CacheFile::from_args(args, share_id)?))
     }
 
     pub fn save(&self) -> io::Result<()> {

src/cachefile.rs

@@ -3,14 +3,17 @@ use std::{
     fs,
     io::{self, Write},
     path::PathBuf,
+    time::Duration,
 };
 
-use log::trace;
+use indicatif::ProgressDrawTarget;
+use log::{debug, trace};
 use serde::{Deserialize, Serialize};
 
 use crate::{
     cli::Cli,
-    file::{self, Chunk},
+    file::{self, Chunk, FileTrait},
+    output::new_progressbar,
     sharry::{self, Client, Uri},
 };
 
@@ -56,15 +59,35 @@ impl CacheFile {
         Ok(Self { file_name, ..state })
     }
 
-    pub fn from_args(args: &Cli, share_id: String) -> Self {
-        Self {
+    pub fn from_args(args: &Cli, share_id: String) -> io::Result<Self> {
+        // TODO CLI switch begin
+
+        let mut files = args.files.clone();
+
+        let bar = new_progressbar();
+        bar.set_draw_target(ProgressDrawTarget::stderr());
+        // BOOKMARK assumption: total file size < 2 EiB
+        bar.set_length(files.iter().map(|f| f.get_size()).sum());
+        bar.enable_steady_tick(Duration::from_millis(50));
+
+        for chk in &mut files {
+            bar.set_message(format!("hashing {:?}", chk.get_name()));
+            chk.hash(|bytes| bar.inc(bytes))?;
+            debug!("{chk:?}");
+        }
+
+        bar.finish_with_message("finished hashing files");
+
+        // TODO CLI switch end
+
+        Ok(Self {
             file_name: Self::cache_file(args),
             uri: args.get_uri(),
             alias_id: args.alias.clone(),
             share_id,
             uploading: None,
-            files: args.files.clone().into(),
-        }
+            files: files.into(),
+        })
     }
 
     pub fn queue_empty(&self) -> bool {

src/cli.rs

@@ -1,7 +1,7 @@
 use std::{convert::Infallible, fmt, io, time::Duration};
 
 use base64ct::{Base64UrlUnpadded, Encoding};
-use blake2::{Blake2b, Digest, digest::consts::U16};
+use blake2b_simd::Params as Blake2b;
 use clap::{
     Parser,
     builder::{PossibleValuesParser, TypedValueParser},
@@ -96,8 +96,6 @@ fn parse_sharry_file(data: &str) -> io::Result<Checked> {
     Checked::new(data)
 }
 
-type Blake2b128 = Blake2b<U16>;
-
 fn sorted<T>(values: &[T]) -> Vec<&T>
 where
     T: Ord,
@@ -143,14 +141,15 @@ impl Cli {
     }
 
     pub fn get_hash(&self) -> String {
-        let mut hasher = Blake2b128::new();
-        hasher.update(self.get_uri());
-        hasher.update(&self.alias);
+        let mut hasher = Blake2b::new().hash_length(16).to_state();
+
+        hasher.update(self.get_uri().as_ref());
+        hasher.update(self.alias.as_bytes());
 
         for chk in sorted(&self.files) {
-            hasher.update(chk);
+            hasher.update(chk.as_ref());
         }
 
-        Base64UrlUnpadded::encode_string(&hasher.finalize())
+        Base64UrlUnpadded::encode_string(hasher.finalize().as_bytes())
     }
 }

src/file/checked.rs

@@ -20,6 +20,8 @@ pub struct Checked {
     pub(super) path: PathBuf,
     /// size of that file
     pub(super) size: u64,
+    /// hash of that file
+    pub(super) hash: Option<String>,
 }
 
 impl AsRef<[u8]> for Checked {
@@ -41,6 +43,7 @@ impl Checked {
             Ok(Self {
                 path: fs::canonicalize(&value)?,
                 size: meta.len(),
+                hash: None,
             })
         } else {
             Err(io::Error::new(
@@ -50,6 +53,19 @@ impl Checked {
         }
     }
 
+    pub fn hash(&mut self, f: impl Fn(u64)) -> io::Result<()> {
+        if self.hash.is_some() {
+            return Err(io::Error::other(format!(
+                "file {:?} is already hashed!",
+                self.path.display()
+            )));
+        }
+
+        self.hash = Some(super::compute_file_hash(&self.path, self.size, f)?);
+
+        Ok(())
+    }
+
     /// start uploading this file
     ///
     /// - tries to create a new entry in a share
@@ -68,7 +84,7 @@ impl Checked {
     ) -> sharry::Result<Uploading> {
         let file_id = client.file_create(uri, alias_id, share_id, &self)?;
 
-        Ok(Uploading::new(self.path, self.size, file_id))
+        Ok(Uploading::new(self.path, self.size, self.hash, file_id))
     }
 }
 
@@ -84,4 +100,8 @@ impl<'t> FileTrait<'t> for Checked {
     fn get_size(&self) -> u64 {
         self.size
     }
+
+    fn check_hash(&self, on_progress: impl Fn(u64)) -> io::Result<bool> {
+        super::check_file_hash(&self.path, self.size, &self.hash, on_progress)
+    }
 }

src/file/mod.rs

@@ -2,12 +2,62 @@ mod checked;
 mod chunk;
 mod uploading;
 
-use std::{ffi::OsStr, path::Path};
+use std::{
+    ffi::OsStr,
+    fs,
+    io::{self, Read},
+    path::Path,
+};
+
+use base64ct::{Base64, Encoding};
+use blake2b_simd::Params as Blake2b;
 
 pub use checked::Checked;
 pub use chunk::Chunk;
 pub use uploading::Uploading;
 
+fn compute_file_hash<P>(path: P, size: u64, on_progress: impl Fn(u64)) -> io::Result<String>
+where
+    P: AsRef<Path>,
+{
+    let mut file = fs::File::open(path)?;
+    let mut hasher = Blake2b::new().hash_length(64).to_state();
+
+    let mut buf = vec![0u8; 4 * 1024 * 1024];
+    let mut bytes_read = 0;
+
+    loop {
+        let n = file.read(&mut buf)?;
+        if n == 0 {
+            break;
+        }
+        hasher.update(&buf[..n]);
+
+        bytes_read += n as u64;
+        on_progress(n as u64);
+    }
+
+    if bytes_read != size {
+        return Err(io::Error::other(format!(
+            "Hashed {bytes_read:?} bytes, known file size {:?}!",
+            size
+        )));
+    }
+
+    Ok(Base64::encode_string(hasher.finalize().as_bytes()))
+}
+
+fn check_file_hash(
+    path: impl AsRef<Path>,
+    size: u64,
+    hash: &Option<String>,
+    on_progress: impl Fn(u64),
+) -> io::Result<bool> {
+    let Some(hash) = hash else { return Ok(false) };
+
+    Ok(*hash == compute_file_hash(path, size, on_progress)?)
+}
+
 pub trait FileTrait<'t> {
     /// extract the filename part of a `Path` reference
     ///
@@ -25,4 +75,6 @@ pub trait FileTrait<'t> {
 
     /// get the file's size
     fn get_size(&self) -> u64;
+
+    fn check_hash(&self, on_progress: impl Fn(u64)) -> io::Result<bool>;
 }

src/file/uploading.rs

@@ -11,8 +11,12 @@ use super::{Checked, Chunk, FileTrait};
 
 #[derive(Serialize, Deserialize, Debug)]
 pub struct Uploading {
+    /// canonical path to a regular file
     path: PathBuf,
+    /// size of that file
     size: u64,
+    /// hash of that file
+    hash: Option<String>,
     file_id: String,
     #[serde(skip)]
     last_offset: Option<u64>,
@@ -20,10 +24,11 @@ pub struct Uploading {
 }
 
 impl Uploading {
-    pub(super) fn new(path: PathBuf, size: u64, file_id: String) -> Self {
+    pub(super) fn new(path: PathBuf, size: u64, hash: Option<String>, file_id: String) -> Self {
         Self {
             path,
             size,
+            hash,
             file_id,
             last_offset: None,
             offset: 0,
@@ -79,6 +84,7 @@ impl Uploading {
         Checked {
             path: self.path,
             size: self.size,
+            hash: self.hash,
         }
     }
 }
@@ -94,4 +100,8 @@ impl<'t> FileTrait<'t> for Uploading {
     fn get_size(&self) -> u64 {
         self.size
     }
+
+    fn check_hash(&self, on_progress: impl Fn(u64)) -> io::Result<bool> {
+        super::check_file_hash(&self.path, self.size, &self.hash, on_progress)
+    }
 }

src/main.rs

@@ -24,6 +24,32 @@ use output::{Log, SHRUPL};
 use sharry::{ClientError, Parameter};
 
 fn main() {
+    let args = Cli::parse();
+
+    env_logger::Builder::new()
+        .filter_module("shrupl", args.get_level_filter())
+        .parse_default_env()
+        .init();
+
+    info!("args: {args:#?}");
+
+    println!("{} to {}!", style("Welcome").magenta().bold(), *SHRUPL);
+
+    let mut state = AppState::try_resume(&args)
+        .and_then(|state| output::prompt_continue().then_some(state))
+        .unwrap_or_else(|| match AppState::from_args(&args) {
+            Ok(state) => {
+                state.save().unwrap_or_else(|e| {
+                    Log::warning(format_args!("Failed to save state: {e}"));
+                });
+                state
+            }
+            Err(e) => {
+                Log::handle(&e);
+                Log::error(format_args!("Failed to create state: {e}"));
+            }
+        });
+
     let check_ctrlc = {
         let stop = Arc::new(AtomicBool::new(false));
         let stop_ctrlc = stop.clone();
@@ -41,36 +67,6 @@ fn main() {
         }
     };
 
-    let args = Cli::parse();
-
-    env_logger::Builder::new()
-        .filter_module("shrupl", args.get_level_filter())
-        .parse_default_env()
-        .init();
-
-    info!("args: {args:#?}");
-
-    println!("{} to {}!", style("Welcome").magenta().bold(), *SHRUPL);
-
-    let mut state = AppState::try_resume(&args)
-        .and_then(|state| output::prompt_continue().then_some(state))
-        .unwrap_or_else(|| {
-            check_ctrlc();
-
-            match AppState::from_args(&args) {
-                Ok(state) => {
-                    state.save().unwrap_or_else(|e| {
-                        Log::warning(format_args!("Failed to save state: {e}"));
-                    });
-                    state
-                }
-                Err(e) => {
-                    Log::handle(&e);
-                    Log::error(format_args!("Failed to create state: {e}"));
-                }
-            }
-        });
-
     info!("continuing with state: {state:#?}");
 
     let fns_magenta = output::style_all(&args.file_names(), StyledObject::magenta).join(", ");

src/output.rs

@@ -2,6 +2,7 @@ use std::{fmt, process, sync::LazyLock};
 
 use console::{StyledObject, style};
 use dialoguer::{Select, theme::ColorfulTheme};
+use indicatif::{ProgressBar, ProgressStyle};
 use log::{error, info};
 
 use crate::sharry;
@@ -43,6 +44,20 @@ where
     strs.iter().map(|&s| f(style(s)).to_string()).collect()
 }
 
+pub fn new_progressbar() -> ProgressBar {
+    ProgressBar::hidden().with_style(
+        ProgressStyle::with_template(&format!(
+            concat!(
+                "{{bar:50.cyan/blue}} {{msg:.magenta}}: ",
+                "{{binary_bytes:.yellow}}{}{{binary_total_bytes:.yellow}} ",
+                "({{eta}})",
+            ),
+            style("/").magenta(),
+        ))
+        .expect("invalid style template"),
+    )
+}
+
 pub enum Log {}
 
 impl Log {