Merge branch 'gix-status'

Author: Byron

Cargo.lock

@@ -242,6 +242,7 @@ members = [
     "gix-diff",
     "gix-date",
     "gix-traverse",
+    "gix-dir",
     "gix-index",
     "gix-bitmap",
     "gix-worktree",

Cargo.toml

@@ -133,6 +133,7 @@ is usable to some extent.
   * `gitoxide-core`
 * **very early**  _(possibly without any documentation and many rough edges)_
   * [gix-date](https://github.com/Byron/gitoxide/blob/main/crate-status.md#gix-date)
+  * [gix-dir](https://github.com/Byron/gitoxide/blob/main/crate-status.md#gix-dir)
 * **idea** _(just a name placeholder)_
   * [gix-note](https://github.com/Byron/gitoxide/blob/main/crate-status.md#gix-note)
   * [gix-fetchhead](https://github.com/Byron/gitoxide/blob/main/crate-status.md#gix-fetchhead)

README.md

@@ -608,6 +608,18 @@ A plumbing crate with shared functionality regarding EWAH compressed bitmaps, as
   * [x] decode on-disk representation
   * [ ] encode on-disk representation
 
+### gix-dir
+
+A git directory walk.
+
+* [ ] list untracked files
+    - [ ] `normal` - files and directories
+    - [ ] `all` - expand to untracked files in untracked directories
+* [ ] list ignored files
+    - [ ] `matching` mode (show every ignored file, do not aggregate into parent directory)
+    - [ ] `traditional` mode (aggregate all ignored files of a folder into ignoring the folder itself)
+* [ ] accelerated walk with `untracked`-cache (as provided by `UNTR` extension of `gix_index::File`)
+
 ### gix-index
 
 The git staging area.
@@ -629,6 +641,7 @@ The git staging area.
     * [x] 'link' base indices to take information from, split index
     * [x] 'sdir' [sparse directory entries](https://github.blog/2021-08-16-highlights-from-git-2-33/) - marker
   * [x] verification of entries and extensions as well as checksum
+  * [ ] expand sparse directory entries using information of the tree itself
 * write
   * [x] V2
   * [x] V3 - extension bits
@@ -655,7 +668,7 @@ The git staging area.
     * [ ] IEOT index entry offset table
     * [ ] 'link' base indices to take information from, split index
     * [ ] 'sdir' sparse directory entries
-* add and remove entries
+* [ ] add and remove entries
 * [x] API documentation
     * [ ] Some examples
 

crate-status.md

@@ -0,0 +1,6 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

gix-dir/CHANGELOG.md

@@ -0,0 +1,14 @@
+[package]
+name = "gix-dir"
+version = "0.0.0"
+repository = "https://github.com/Byron/gitoxide"
+license = "MIT OR Apache-2.0"
+description = "A crate of the gitoxide project dealing with directory walks"
+authors = ["Sebastian Thiel <[email protected]>"]
+edition = "2021"
+rust-version = "1.65"
+
+[lib]
+doctest = false
+
+[dependencies]

gix-dir/Cargo.toml

@@ -0,0 +1 @@
+../LICENSE-APACHE

gix-dir/LICENSE-APACHE

@@ -0,0 +1 @@
+../LICENSE-MIT

gix-dir/LICENSE-MIT

@@ -0,0 +1,3 @@
+//! A crate for handling a git-style directory walk.
+#![deny(rust_2018_idioms)]
+#![forbid(unsafe_code)]

gix-dir/src/lib.rs

@@ -19,6 +19,7 @@ serde = ["dep:serde", "bstr/serde", "gix-glob/serde"]
 [dependencies]
 gix-glob = { version = "^0.15.1", path = "../gix-glob" }
 gix-path = { version = "^0.10.3", path = "../gix-path" }
+gix-trace = { version = "^0.1.6", path = "../gix-trace" }
 
 bstr = { version = "1.3.0", default-features = false, features = ["std", "unicode"]}
 unicode-bom = "2.0.2"

gix-ignore/Cargo.toml

@@ -25,6 +25,25 @@ pub struct Search {
     pub patterns: Vec<gix_glob::search::pattern::List<search::Ignore>>,
 }
 
+/// The kind of *ignored* item.
+///
+/// This classification is obtained when checking if a path matches an ignore pattern.
+#[derive(Default, Copy, Clone, Ord, PartialOrd, Eq, PartialEq, Hash, Debug)]
+pub enum Kind {
+    /// The item is ignored and will be removed to make place for tracked items that are to be checked out.
+    ///
+    /// This is the default for ignored items.
+    /// Another way of thinking about this class is to consider these files *trashable*, or talk about them as `ignored-and-expendable`.
+    #[default]
+    Expendable,
+    /// An ignored file was additionally marked as *precious* using the `$` prefix to indicate the file shall be kept.
+    ///
+    /// This means that precious files are treated like untracked files, which also must not be removed, but won't show up by default
+    /// as they are also ignored.
+    /// One can also talk about them as `ignored-and-precious`.
+    Precious,
+}
+
 ///
 pub mod parse;
 

gix-ignore/src/lib.rs

@@ -18,17 +18,43 @@ impl<'a> Lines<'a> {
 }
 
 impl<'a> Iterator for Lines<'a> {
-    type Item = (gix_glob::Pattern, usize);
+    type Item = (gix_glob::Pattern, usize, crate::Kind);
 
     fn next(&mut self) -> Option<Self::Item> {
-        for line in self.lines.by_ref() {
+        for mut line in self.lines.by_ref() {
             self.line_no += 1;
-            if line.first() == Some(&b'#') {
-                continue;
-            }
-            match gix_glob::Pattern::from_bytes(truncate_non_escaped_trailing_spaces(line)) {
+            let first = match line.first().copied() {
+                Some(b'#') | None => continue,
+                Some(c) => c,
+            };
+            let (kind, can_negate) = if first == b'$' {
+                line = &line[1..];
+                (crate::Kind::Precious, false)
+            } else {
+                let second = line.get(1);
+                if first == b'!' && second == Some(&b'$') {
+                    gix_trace::error!(
+                        "Line {} starts with !$ which is not allowed ('{}')",
+                        self.line_no,
+                        line.as_bstr()
+                    );
+                    continue;
+                }
+                if first == b'\\' && second == Some(&b'$') {
+                    line = &line[1..];
+                }
+                (crate::Kind::Expendable, true)
+            };
+
+            line = truncate_non_escaped_trailing_spaces(line);
+            let res = if can_negate {
+                gix_glob::Pattern::from_bytes(line)
+            } else {
+                gix_glob::Pattern::from_bytes_without_negation(line)
+            };
+            match res {
                 None => continue,
-                Some(pattern) => return Some((pattern, self.line_no)),
+                Some(pattern) => return Some((pattern, self.line_no, kind)),
             }
         }
         None

gix-ignore/src/parse.rs

@@ -15,6 +15,8 @@ pub struct Match<'a> {
     pub pattern: &'a gix_glob::Pattern,
     /// The path to the source from which the pattern was loaded, or `None` if it was specified by other means.
     pub source: Option<&'a Path>,
+    /// The kind of pattern this match represents.
+    pub kind: crate::Kind,
     /// The line at which the pattern was found in its `source` file, or the occurrence in which it was provided.
     pub sequence_number: usize,
 }
@@ -24,13 +26,13 @@ pub struct Match<'a> {
 pub struct Ignore;
 
 impl Pattern for Ignore {
-    type Value = ();
+    type Value = crate::Kind;
 
     fn bytes_to_patterns(bytes: &[u8], _source: &std::path::Path) -> Vec<pattern::Mapping<Self::Value>> {
         crate::parse(bytes)
-            .map(|(pattern, line_number)| pattern::Mapping {
+            .map(|(pattern, line_number, kind)| pattern::Mapping {
                 pattern,
-                value: (),
+                value: kind,
                 sequence_number: line_number,
             })
             .collect()
@@ -61,7 +63,7 @@ impl Search {
         Ok(group)
     }
 
-    /// Parse a list of patterns, using slashes as path separators
+    /// Parse a list of ignore patterns, using slashes as path separators.
     pub fn from_overrides(patterns: impl IntoIterator<Item = impl Into<OsString>>) -> Self {
         Self::from_overrides_inner(&mut patterns.into_iter().map(Into::into))
     }
@@ -73,11 +75,13 @@ impl Search {
                     .enumerate()
                     .filter_map(|(seq_id, pattern)| {
                         let pattern = gix_path::try_into_bstr(PathBuf::from(pattern)).ok()?;
-                        gix_glob::parse(pattern.as_ref()).map(|p| pattern::Mapping {
-                            pattern: p,
-                            value: (),
-                            sequence_number: seq_id,
-                        })
+                        crate::parse(pattern.as_ref())
+                            .next()
+                            .map(|(p, _seq_id, kind)| pattern::Mapping {
+                                pattern: p,
+                                value: kind,
+                                sequence_number: seq_id + 1,
+                            })
                     })
                     .collect(),
                 source: None,
@@ -112,7 +116,7 @@ pub fn pattern_matching_relative_path<'a>(
     list.patterns.iter().rev().find_map(
         |pattern::Mapping {
              pattern,
-             value: (),
+             value: kind,
              sequence_number,
          }| {
             pattern
@@ -125,6 +129,7 @@ pub fn pattern_matching_relative_path<'a>(
                 )
                 .then_some(Match {
                     pattern,
+                    kind: *kind,
                     source: list.source.as_deref(),
                     sequence_number: *sequence_number,
                 })

gix-ignore/src/search.rs

@@ -0,0 +1,12 @@
+$.config
+\$starts-with-dollar
+# html files are now precious and won't be discarded
+$*.html
+
+!foo.html
+
+# this isn't allowed and ignored
+!$foo.html
+
+# but this is a literal !/* that is precious
+$!/*

gix-ignore/tests/fixtures/ignore/precious.txt

@@ -2,6 +2,22 @@ use bstr::BString;
 use gix_glob::{pattern::Mode, Pattern};
 use gix_testtools::fixture_bytes;
 
+#[test]
+fn precious() {
+    let input = fixture_bytes("ignore/precious.txt");
+    let actual: Vec<_> = gix_ignore::parse(&input).map(flat_map).collect();
+    assert_eq!(
+        actual,
+        vec![
+            pat_precious(".config", Mode::NO_SUB_DIR, 1),
+            pat("$starts-with-dollar", Mode::NO_SUB_DIR, 2),
+            pat_precious("*.html", Mode::NO_SUB_DIR | Mode::ENDS_WITH, 4),
+            pat("foo.html", Mode::NO_SUB_DIR | Mode::NEGATIVE, 6),
+            pat_precious("!/*", Mode::empty(), 12),
+        ]
+    );
+}
+
 #[test]
 fn byte_order_marks_are_no_patterns() {
     assert_eq!(
@@ -58,7 +74,7 @@ fn backslashes_before_hashes_are_no_comments() {
 
 #[test]
 fn trailing_spaces_can_be_escaped_to_be_literal() {
-    fn parse_one(input: &str) -> (BString, Mode, usize) {
+    fn parse_one(input: &str) -> (BString, Mode, usize, gix_ignore::Kind) {
         let actual: Vec<_> = gix_ignore::parse(input.as_bytes()).map(flat_map).collect();
         assert_eq!(actual.len(), 1, "{input:?} should match");
         actual.into_iter().next().expect("present")
@@ -101,14 +117,20 @@ fn trailing_spaces_can_be_escaped_to_be_literal() {
     );
 }
 
-fn flatten(input: Option<(Pattern, usize)>) -> Option<(BString, gix_glob::pattern::Mode, usize)> {
+fn flatten(
+    input: Option<(Pattern, usize, gix_ignore::Kind)>,
+) -> Option<(BString, gix_glob::pattern::Mode, usize, gix_ignore::Kind)> {
     input.map(flat_map)
 }
 
-fn flat_map(input: (Pattern, usize)) -> (BString, gix_glob::pattern::Mode, usize) {
-    (input.0.text, input.0.mode, input.1)
+fn flat_map(input: (Pattern, usize, gix_ignore::Kind)) -> (BString, gix_glob::pattern::Mode, usize, gix_ignore::Kind) {
+    (input.0.text, input.0.mode, input.1, input.2)
+}
+
+fn pat(pattern: &str, mode: Mode, pos: usize) -> (BString, Mode, usize, gix_ignore::Kind) {
+    (pattern.into(), mode, pos, gix_ignore::Kind::Expendable)
 }
 
-fn pat(pattern: &str, mode: Mode, pos: usize) -> (BString, Mode, usize) {
-    (pattern.into(), mode, pos)
+fn pat_precious(pattern: &str, mode: Mode, pos: usize) -> (BString, Mode, usize, gix_ignore::Kind) {
+    (pattern.into(), mode, pos, gix_ignore::Kind::Precious)
 }