From 0fad4f44f45ff8efcb522d239bd20d69fd7bd36a Mon Sep 17 00:00:00 2001 From: Jordan Doyle Date: Sat, 28 Sep 2024 23:27:53 +0400 Subject: [PATCH] Use simdutf8 for interacting with files within Git --- Cargo.lock | 7 +++++++ Cargo.toml | 1 + src/git.rs | 39 ++++++++++++++++++++++++++++----------- 3 files changed, 33 insertions(+), 14 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index afcba74..b529169 100644 --- a/Cargo.lock +++ a/Cargo.lock @@ -2814,6 +2814,7 @@ "rsass", "rust-ini", "serde", + "simdutf8", "syntect", "tar", "time", @@ -3061,6 +3062,12 @@ dependencies = [ "libc", ] + +[[package]] +name = "simdutf8" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e3a9fe34e3e7a50316060351f37187a3f546bce95496156754b601a5fa71b76e" [[package]] name = "slab" diff --git a/Cargo.toml b/Cargo.toml index 5afb434..cee9e25 100644 --- a/Cargo.toml +++ a/Cargo.toml @@ -37,6 +37,7 @@ rocksdb = { version = "0.22", default-features = false, features = ["snappy"] } rust-ini = "0.21.1" serde = { version = "1.0", features = ["derive", "rc"] } +simdutf8 = "0.1.5" syntect = "5" tar = "0.4" time = { version = "0.3", features = ["serde"] } diff --git a/src/git.rs b/src/git.rs index b7bc468..9086544 100644 --- a/src/git.rs +++ a/src/git.rs @@ -1,6 +1,6 @@ use std::{ borrow::Cow, - collections::{BTreeMap, VecDeque}, + collections::VecDeque, ffi::OsStr, fmt::{self, Arguments, Write}, io::ErrorKind, @@ -157,15 +157,17 @@ .or_else(|| path.file_name()) .map_or_else(|| Cow::Borrowed(""), OsStr::to_string_lossy); - let content = match (formatted, String::from_utf8(blob.take_data())) { + let content = match (formatted, simdutf8::basic::from_utf8(&blob.data)) { (true, Err(_)) => Content::Binary(vec![]), (true, Ok(data)) => Content::Text(Cow::Owned(format_file( - &data, + data, &extension, &self.git.syntax_set, )?)), - (false, Err(e)) => Content::Binary(e.into_bytes()), - (false, Ok(data)) => Content::Text(Cow::Owned(data)), + (false, Err(_)) => Content::Binary(blob.take_data()), + (false, Ok(_data)) => Content::Text(Cow::Owned(unsafe { + String::from_utf8_unchecked(blob.take_data()) + })), }; return Ok(PathDestination::File(FileWithContent { @@ -295,7 +297,7 @@ continue; }; - let Ok(content) = std::str::from_utf8(&blob.data) else { + let Ok(content) = simdutf8::basic::from_utf8(&blob.data) else { continue; }; @@ -757,7 +759,7 @@ .transpose()? .unwrap_or_else(|| repo.empty_tree()); - let mut diffs = BTreeMap::<_, FileDiff>::new(); + let mut diffs = Vec::new(); let mut diff_output = String::new(); let mut resource_cache = repo.diff_resource_cache_for_tree_diff()?; @@ -795,9 +797,9 @@ diffs.iter().fold( (0, 0, 0, 0, 0), |(max_file_name_length, max_change_length, files_changed, insertions, deletions), - (f, stats)| { + stats| { ( - max_file_name_length.max(f.len()), + max_file_name_length.max(stats.path.len()), max_change_length .max(((stats.insertions + stats.deletions).ilog10() + 1) as usize), files_changed + 1, @@ -811,7 +813,7 @@ let total_changes = insertions + deletions; - for (file, diff) in &diffs { + for diff in &diffs { let local_changes = diff.insertions + diff.deletions; let width = WIDTH.min(local_changes); @@ -829,6 +831,7 @@ let plus_str = "+".repeat(adjusted_addition_width); let minus_str = "-".repeat(adjusted_deletion_width); + let file = diff.path.as_str(); writeln!(diff_stats, " {file:max_file_name_length$} | {local_changes:max_change_length$} {plus_str}{minus_str}").unwrap(); } @@ -864,6 +867,7 @@ #[derive(Default, Debug)] struct FileDiff { + path: String, insertions: usize, deletions: usize, } @@ -1039,11 +1043,12 @@ struct DiffBuilder<'a, F> { output: &'a mut String, resource_cache: &'a mut gix::diff::blob::Platform, - diffs: &'a mut BTreeMap, + diffs: &'a mut Vec, formatter: F, } impl<'a, F: DiffFormatter + Callback> DiffBuilder<'a, F> { + #[allow(clippy::too_many_lines)] fn handle( &mut self, change: gix::object::tree::diff::Change<'_, '_, '_>, @@ -1052,7 +1057,11 @@ return Ok(gix::object::tree::diff::Action::Continue); } - let diff = self.diffs.entry(change.location.to_string()).or_default(); + let mut diff = FileDiff { + path: change.location.to_string(), + insertions: 0, + deletions: 0, + }; let change = change.diff(self.resource_cache)?; let prep = change.resource_cache.prepare_diff()?; @@ -1129,10 +1138,10 @@ .file_header(self.output, format_args!("+++ {new_path}")); let old_source = gix::diff::blob::sources::lines_with_terminator( - std::str::from_utf8(prep.old.data.as_slice().unwrap_or_default())?, + simdutf8::basic::from_utf8(prep.old.data.as_slice().unwrap_or_default())?, ); let new_source = gix::diff::blob::sources::lines_with_terminator( - std::str::from_utf8(prep.new.data.as_slice().unwrap_or_default())?, + simdutf8::basic::from_utf8(prep.new.data.as_slice().unwrap_or_default())?, ); let input = gix::diff::blob::intern::InternedInput::new(old_source, new_source); @@ -1165,6 +1174,8 @@ ); } } + + self.diffs.push(diff); self.resource_cache.clear_resource_cache_keep_allocation(); Ok(gix::object::tree::diff::Action::Continue) -- rgit 0.1.3