gix-status-improvements

crate-status.md

@@ -456,7 +456,9 @@ Make it the best-performing implementation and the most convenient one.
 
 ### gix-status
 * [x] differences between index and worktree to turn index into worktree
-* [ ] differences between tree and index to turn tree into index
+    - [ ] rename tracking
+* [ ] differences between index and index to learn what changed
+    - [ ] rename tracking
 * [ ] untracked files
 * [ ] fast answer to 'is it dirty'.
 * 

gitoxide-core/src/repository/index/entries.rs

@@ -404,9 +404,10 @@ pub(crate) mod function {
             out,
             "{} {}{:?} {} {}{}{}",
             match entry.flags.stage() {
-                0 => "BASE   ",
-                1 => "OURS   ",
-                2 => "THEIRS ",
+                0 => "       ",
+                1 => "BASE   ",
+                2 => "OURS   ",
+                3 => "THEIRS ",
                 _ => "UNKNOWN",
             },
             if entry.flags.is_empty() {

gitoxide-core/src/repository/status.rs

@@ -4,8 +4,8 @@ use gix::bstr::{BStr, BString};
 use gix::index::Entry;
 use gix::prelude::FindExt;
 use gix::Progress;
-use gix_status::index_as_worktree::content::FastEq;
-use gix_status::index_as_worktree::Change;
+use gix_status::index_as_worktree::traits::FastEq;
+use gix_status::index_as_worktree::{Change, Conflict, EntryStatus};
 
 pub enum Submodules {
     /// display all information about submodules, including ref changes, modifications and untracked files.
@@ -20,6 +20,8 @@ pub struct Options {
     pub format: OutputFormat,
     pub submodules: Submodules,
     pub thread_limit: Option<usize>,
+    pub statistics: bool,
+    pub allow_write: bool,
 }
 
 pub fn show(
@@ -33,12 +35,14 @@ pub fn show(
         // TODO: implement this
         submodules: _,
         thread_limit,
+        allow_write,
+        statistics,
     }: Options,
 ) -> anyhow::Result<()> {
     if format != OutputFormat::Human {
         bail!("Only human format is supported right now");
     }
-    let mut index = repo.index()?;
+    let mut index = repo.index_or_empty()?;
     let index = gix::threading::make_mut(&mut index);
     let pathspec = repo.pathspec(
         pathspecs,
@@ -48,73 +52,174 @@ pub fn show(
     )?;
     let mut progress = progress.add_child("traverse index");
     let start = std::time::Instant::now();
-    gix_status::index_as_worktree(
+    let options = gix_status::index_as_worktree::Options {
+        fs: repo.filesystem_options()?,
+        thread_limit,
+        stat: repo.stat_options()?,
+        attributes: match repo
+            .attributes_only(
+                index,
+                gix::worktree::stack::state::attributes::Source::WorktreeThenIdMapping,
+            )?
+            .detach()
+            .state_mut()
+        {
+            gix::worktree::stack::State::AttributesStack(attrs) => std::mem::take(attrs),
+            // TODO: this should be nicer by creating attributes directly, but it's a private API
+            _ => unreachable!("state must be attributes stack only"),
+        },
+    };
+    let mut printer = Printer {
+        out,
+        changes: Vec::new(),
+    };
+    let outcome = gix_status::index_as_worktree(
         index,
         repo.work_dir()
             .context("This operation cannot be run on a bare repository")?,
-        &mut Printer(out),
+        &mut printer,
         FastEq,
+        Submodule,
         {
             let odb = repo.objects.clone().into_arc()?;
             move |id, buf| odb.find_blob(id, buf)
         },
         &mut progress,
         pathspec.detach()?,
-        gix_status::index_as_worktree::Options {
-            fs: repo.filesystem_options()?,
-            thread_limit,
-            stat: repo.stat_options()?,
-        },
+        repo.filter_pipeline(Some(gix::hash::ObjectId::empty_tree(repo.object_hash())))?
+            .0
+            .into_parts()
+            .0,
+        &gix::interrupt::IS_INTERRUPTED,
+        options,
     )?;
 
+    if outcome.entries_to_update != 0 && allow_write {
+        {
+            let entries = index.entries_mut();
+            for (entry_index, change) in printer.changes {
+                let entry = &mut entries[entry_index];
+                match change {
+                    ApplyChange::SetSizeToZero => {
+                        entry.stat.size = 0;
+                    }
+                    ApplyChange::NewStat(new_stat) => {
+                        entry.stat = new_stat;
+                    }
+                }
+            }
+        }
+        index.write(gix::index::write::Options {
+            extensions: Default::default(),
+            skip_hash: false, // TODO: make this based on configuration
+        })?;
+    }
+
+    if statistics {
+        writeln!(err, "{outcome:#?}").ok();
+    }
+
     writeln!(err, "\nhead -> index and untracked files aren't implemented yet")?;
     progress.show_throughput(start);
     Ok(())
 }
 
-struct Printer<W>(W);
+#[derive(Clone)]
+struct Submodule;
+
+impl gix_status::index_as_worktree::traits::SubmoduleStatus for Submodule {
+    type Output = ();
+    type Error = std::convert::Infallible;
+
+    fn status(&mut self, _entry: &Entry, _rela_path: &BStr) -> Result<Option<Self::Output>, Self::Error> {
+        Ok(None)
+    }
+}
+
+struct Printer<W> {
+    out: W,
+    changes: Vec<(usize, ApplyChange)>,
+}
+
+enum ApplyChange {
+    SetSizeToZero,
+    NewStat(gix::index::entry::Stat),
+}
 
 impl<'index, W> gix_status::index_as_worktree::VisitEntry<'index> for Printer<W>
 where
     W: std::io::Write,
 {
     type ContentChange = ();
+    type SubmoduleStatus = ();
 
     fn visit_entry(
         &mut self,
-        entry: &'index Entry,
+        _entries: &'index [Entry],
+        _entry: &'index Entry,
+        entry_index: usize,
         rela_path: &'index BStr,
-        change: Option<Change<Self::ContentChange>>,
-        conflict: bool,
+        status: EntryStatus<Self::ContentChange>,
     ) {
-        self.visit_inner(entry, rela_path, change, conflict).ok();
+        self.visit_inner(entry_index, rela_path, status).ok();
     }
 }
 
 impl<W: std::io::Write> Printer<W> {
-    fn visit_inner(
-        &mut self,
-        _entry: &Entry,
-        rela_path: &BStr,
-        change: Option<Change<()>>,
-        conflict: bool,
-    ) -> anyhow::Result<()> {
-        if let Some(change) = conflict
-            .then_some('U')
-            .or_else(|| change.as_ref().and_then(change_to_char))
-        {
-            writeln!(&mut self.0, "{change} {rela_path}")?;
-        }
-        Ok(())
+    fn visit_inner(&mut self, entry_index: usize, rela_path: &BStr, status: EntryStatus<()>) -> std::io::Result<()> {
+        let char_storage;
+        let status = match status {
+            EntryStatus::Conflict(conflict) => as_str(conflict),
+            EntryStatus::Change(change) => {
+                if matches!(
+                    change,
+                    Change::Modification {
+                        set_entry_stat_size_zero: true,
+                        ..
+                    }
+                ) {
+                    self.changes.push((entry_index, ApplyChange::SetSizeToZero))
+                }
+                char_storage = change_to_char(&change);
+                std::str::from_utf8(std::slice::from_ref(&char_storage)).expect("valid ASCII")
+            }
+            EntryStatus::NeedsUpdate(stat) => {
+                self.changes.push((entry_index, ApplyChange::NewStat(stat)));
+                return Ok(());
+            }
+            EntryStatus::IntentToAdd => "A",
+        };
+
+        writeln!(&mut self.out, "{status: >3} {rela_path}")
+    }
+}
+
+fn as_str(c: Conflict) -> &'static str {
+    match c {
+        Conflict::BothDeleted => "DD",
+        Conflict::AddedByUs => "AU",
+        Conflict::DeletedByThem => "UD",
+        Conflict::AddedByThem => "UA",
+        Conflict::DeletedByUs => "DU",
+        Conflict::BothAdded => "AA",
+        Conflict::BothModified => "UU",
     }
 }
 
-fn change_to_char(change: &Change<()>) -> Option<char> {
+fn change_to_char(change: &Change<()>) -> u8 {
     // Known status letters: https://github.com/git/git/blob/6807fcfedab84bc8cd0fbf721bc13c4e68cda9ae/diff.h#L613
-    Some(match change {
-        Change::Removed => 'D',
-        Change::Type => 'T',
-        Change::Modification { .. } => 'M',
-        Change::IntentToAdd => return None,
-    })
+    match change {
+        Change::Removed => b'D',
+        Change::Type => b'T',
+        Change::SubmoduleModification(_) => b'M',
+        Change::Modification {
+            executable_bit_changed, ..
+        } => {
+            if *executable_bit_changed {
+                b'X'
+            } else {
+                b'M'
+            }
+        }
+    }
 }

gix-command/tests/command.rs

@@ -82,32 +82,26 @@ mod spawn {
 
         #[test]
         fn sh_shell_specific_script_code_with_single_extra_arg() -> crate::Result {
-            let out = gix_command::prepare("echo")
+            let out = gix_command::prepare("printf")
                 .with_shell()
                 .arg("1")
                 .spawn()?
                 .wait_with_output()?;
             assert!(out.status.success());
-            #[cfg(not(windows))]
-            assert_eq!(out.stdout.as_bstr(), "1\n");
-            #[cfg(windows)]
-            assert_eq!(out.stdout.as_bstr(), "1\r\n");
+            assert_eq!(out.stdout.as_bstr(), "1");
             Ok(())
         }
 
         #[test]
         fn sh_shell_specific_script_code_with_multiple_extra_args() -> crate::Result {
-            let out = gix_command::prepare("echo")
+            let out = gix_command::prepare("printf")
                 .with_shell()
-                .arg("1")
-                .arg("2")
+                .arg("%s")
+                .arg("arg")
                 .spawn()?
                 .wait_with_output()?;
             assert!(out.status.success());
-            #[cfg(not(windows))]
-            assert_eq!(out.stdout.as_bstr(), "1 2\n");
-            #[cfg(windows)]
-            assert_eq!(out.stdout.as_bstr(), "1 2\r\n");
+            assert_eq!(out.stdout.as_bstr(), "arg");
             Ok(())
         }
     }

gix-features/src/hash.rs

@@ -96,7 +96,7 @@ pub fn hasher(kind: gix_hash::Kind) -> Sha1 {
 #[cfg(all(feature = "progress", any(feature = "rustsha1", feature = "fast-sha1")))]
 pub fn bytes_of_file(
     path: &std::path::Path,
-    num_bytes_from_start: usize,
+    num_bytes_from_start: u64,
     kind: gix_hash::Kind,
     progress: &mut dyn crate::progress::Progress,
     should_interrupt: &std::sync::atomic::AtomicBool,
@@ -110,28 +110,42 @@ pub fn bytes_of_file(
     )
 }
 
-/// Similar to [`bytes_of_file`], but operates on an already open file.
+/// Similar to [`bytes_of_file`], but operates on a stream of bytes.
 #[cfg(all(feature = "progress", any(feature = "rustsha1", feature = "fast-sha1")))]
 pub fn bytes(
     read: &mut dyn std::io::Read,
-    num_bytes_from_start: usize,
+    num_bytes_from_start: u64,
     kind: gix_hash::Kind,
     progress: &mut dyn crate::progress::Progress,
     should_interrupt: &std::sync::atomic::AtomicBool,
 ) -> std::io::Result<gix_hash::ObjectId> {
-    let mut hasher = hasher(kind);
+    bytes_with_hasher(read, num_bytes_from_start, hasher(kind), progress, should_interrupt)
+}
+
+/// Similar to [`bytes()`], but takes a `hasher` instead of a hash kind.
+#[cfg(all(feature = "progress", any(feature = "rustsha1", feature = "fast-sha1")))]
+pub fn bytes_with_hasher(
+    read: &mut dyn std::io::Read,
+    num_bytes_from_start: u64,
+    mut hasher: Sha1,
+    progress: &mut dyn crate::progress::Progress,
+    should_interrupt: &std::sync::atomic::AtomicBool,
+) -> std::io::Result<gix_hash::ObjectId> {
     let start = std::time::Instant::now();
     // init progress before the possibility for failure, as convenience in case people want to recover
-    progress.init(Some(num_bytes_from_start), crate::progress::bytes());
+    progress.init(
+        Some(num_bytes_from_start as prodash::progress::Step),
+        crate::progress::bytes(),
+    );
 
     const BUF_SIZE: usize = u16::MAX as usize;
     let mut buf = [0u8; BUF_SIZE];
     let mut bytes_left = num_bytes_from_start;
 
     while bytes_left > 0 {
-        let out = &mut buf[..BUF_SIZE.min(bytes_left)];
+        let out = &mut buf[..BUF_SIZE.min(bytes_left as usize)];
         read.read_exact(out)?;
-        bytes_left -= out.len();
+        bytes_left -= out.len() as u64;
         progress.inc_by(out.len());
         hasher.update(out);
         if should_interrupt.load(std::sync::atomic::Ordering::SeqCst) {