Add bindings to git_indexer

kim · kim · commit d042e124d69c · 2023-01-12T17:05:49.000+01:00
The indexer API is a lower-level interface for storing and indexing pack
files, which, unlike `git_odb_write_pack`, allows the ouput to be
written to an arbitrary directory. This can be useful when working with
unusual validation requirements or non-standard repository layouts.
diff --git a/libgit2-sys/lib.rs b/libgit2-sys/lib.rs
@@ -25,6 +25,7 @@ pub const GIT_REFDB_BACKEND_VERSION: c_uint = 1;
 pub const GIT_CHERRYPICK_OPTIONS_VERSION: c_uint = 1;
 pub const GIT_APPLY_OPTIONS_VERSION: c_uint = 1;
 pub const GIT_REVERT_OPTIONS_VERSION: c_uint = 1;
+pub const GIT_INDEXER_OPTIONS_VERSION: c_uint = 1;
 
 macro_rules! git_enum {
     (pub enum $name:ident { $($variants:tt)* }) => {
@@ -91,6 +92,7 @@ pub enum git_odb_object {}
 pub enum git_worktree {}
 pub enum git_transaction {}
 pub enum git_mailmap {}
+pub enum git_indexer {}
 
 #[repr(C)]
 pub struct git_revspec {
@@ -354,6 +356,14 @@ pub type git_indexer_progress_cb =
 )]
 pub type git_transfer_progress = git_indexer_progress;
 
+#[repr(C)]
+pub struct git_indexer_options {
+    pub version: c_uint,
+    pub progress_cb: git_indexer_progress_cb,
+    pub progress_cb_payload: *mut c_void,
+    pub verify: bool,
+}
+
 pub type git_remote_ready_cb = Option<extern "C" fn(*mut git_remote, c_int, *mut c_void) -> c_int>;
 
 #[repr(C)]
@@ -3801,6 +3811,27 @@ extern "C" {
     ) -> c_int;
     pub fn git_packbuilder_free(pb: *mut git_packbuilder);
 
+    // indexer
+    pub fn git_indexer_new(
+        out: *mut *mut git_indexer,
+        path: *const c_char,
+        mode: c_uint,
+        odb: *mut git_odb,
+        opts: *mut git_indexer_options,
+    ) -> c_int;
+    pub fn git_indexer_append(
+        idx: *mut git_indexer,
+        data: *mut c_void,
+        size: size_t,
+        stats: *mut git_indexer_progress,
+    ) -> c_int;
+    pub fn git_indexer_commit(idx: *mut git_indexer, stats: *mut git_indexer_progress) -> c_int;
+    pub fn git_indexer_hash(idx: *mut git_indexer) -> *const git_oid;
+    pub fn git_indexer_name(idx: *mut git_indexer) -> *const c_char;
+    pub fn git_indexer_free(idx: *mut git_indexer);
+
+    pub fn git_indexer_options_init(opts: *mut git_indexer_options, version: c_uint) -> c_int;
+
     // odb
     pub fn git_repository_odb(out: *mut *mut git_odb, repo: *mut git_repository) -> c_int;
     pub fn git_odb_new(db: *mut *mut git_odb) -> c_int;
diff --git a/src/lib.rs b/src/lib.rs
@@ -109,7 +109,7 @@ pub use crate::message::{
 };
 pub use crate::note::{Note, Notes};
 pub use crate::object::Object;
-pub use crate::odb::{Odb, OdbObject, OdbPackwriter, OdbReader, OdbWriter};
+pub use crate::odb::{Indexer, Odb, OdbObject, OdbPackwriter, OdbReader, OdbWriter};
 pub use crate::oid::Oid;
 pub use crate::packbuilder::{PackBuilder, PackBuilderStage};
 pub use crate::patch::Patch;
diff --git a/src/odb.rs b/src/odb.rs
@@ -1,6 +1,9 @@
+use std::ffi::CStr;
 use std::io;
 use std::marker;
+use std::mem;
 use std::mem::MaybeUninit;
+use std::path::Path;
 use std::ptr;
 use std::slice;
 
@@ -10,6 +13,7 @@ use libc::{c_char, c_int, c_uint, c_void, size_t};
 
 use crate::panic;
 use crate::util::Binding;
+use crate::IntoCString;
 use crate::{
     raw, Error, IndexerProgress, Mempack, Object, ObjectType, OdbLookupFlags, Oid, Progress,
 };
@@ -183,6 +187,41 @@ impl<'repo> Odb<'repo> {
         })
     }
 
+    /// Create a stream for writing a pack file to an arbitrary path
+    ///
+    /// This [`Odb`] is used to resolve objects if the written pack is "thin", i.e. depends on
+    /// already-known objects.
+    ///
+    /// `mode` are the file permissions to use on the output.
+    pub fn indexer(&self, path: &Path, mode: u32) -> Result<Indexer<'_>, Error> {
+        let path = path.into_c_string()?;
+        let mut out = ptr::null_mut();
+        let progress = MaybeUninit::uninit();
+        let progress_cb: raw::git_indexer_progress_cb = Some(write_pack_progress_cb);
+        let progress_payload = Box::new(OdbPackwriterCb { cb: None });
+        let progress_payload_ptr = Box::into_raw(progress_payload);
+
+        unsafe {
+            let mut opts = mem::zeroed();
+            try_call!(raw::git_indexer_options_init(
+                &mut opts,
+                raw::GIT_INDEXER_OPTIONS_VERSION
+            ));
+            opts.progress_cb = progress_cb;
+            opts.progress_cb_payload = progress_payload_ptr as *mut c_void;
+
+            try_call!(raw::git_indexer_new(
+                &mut out, path, mode, self.raw, &mut opts
+            ));
+        }
+
+        Ok(Indexer {
+            raw: out,
+            progress,
+            progress_payload_ptr,
+        })
+    }
+
     /// Checks if the object database has an object.
     pub fn exists(&self, oid: Oid) -> bool {
         unsafe { raw::git_odb_exists(self.raw, oid.raw()) != 0 }
@@ -519,6 +558,78 @@ impl<'repo> Drop for OdbPackwriter<'repo> {
     }
 }
 
+/// A stream to write and index a packfile
+///
+/// This is a lower-level interface than [`OdbPackwriter`] which allows to write the pack data and
+/// index to an arbitrary path, but is otherwise equivalent.
+pub struct Indexer<'odb> {
+    raw: *mut raw::git_indexer,
+    progress: MaybeUninit<raw::git_indexer_progress>,
+    progress_payload_ptr: *mut OdbPackwriterCb<'odb>,
+}
+
+impl<'a> Indexer<'a> {
+    /// Finalize the pack and index
+    ///
+    /// Resolves any pending deltas and writes out the index file. The returned string is the
+    /// hexadecimal checksum of the packfile, which is also used to name the pack and index files
+    /// (`pack-<checksum>.pack` and `pack-<checksum>.idx` respectively).
+    pub fn commit(mut self) -> Result<String, Error> {
+        unsafe {
+            try_call!(raw::git_indexer_commit(
+                self.raw,
+                self.progress.as_mut_ptr()
+            ));
+
+            let name = CStr::from_ptr(raw::git_indexer_name(self.raw));
+            Ok(name.to_str().expect("pack name not utf8").to_owned())
+        }
+    }
+
+    /// The callback through which progress is monitored. Be aware that this is
+    /// called inline, so performance may be affected.
+    pub fn progress<F>(&mut self, cb: F) -> &mut Self
+    where
+        F: FnMut(Progress<'_>) -> bool + 'a,
+    {
+        let progress_payload =
+            unsafe { &mut *(self.progress_payload_ptr as *mut OdbPackwriterCb<'_>) };
+        progress_payload.cb = Some(Box::new(cb) as Box<IndexerProgress<'a>>);
+
+        self
+    }
+}
+
+impl io::Write for Indexer<'_> {
+    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
+        unsafe {
+            let ptr = buf.as_ptr() as *mut c_void;
+            let len = buf.len();
+
+            let res = raw::git_indexer_append(self.raw, ptr, len, self.progress.as_mut_ptr());
+
+            if res < 0 {
+                Err(io::Error::new(io::ErrorKind::Other, "Write error"))
+            } else {
+                Ok(buf.len())
+            }
+        }
+    }
+
+    fn flush(&mut self) -> io::Result<()> {
+        Ok(())
+    }
+}
+
+impl Drop for Indexer<'_> {
+    fn drop(&mut self) {
+        unsafe {
+            raw::git_indexer_free(self.raw);
+            drop(Box::from_raw(self.progress_payload_ptr))
+        }
+    }
+}
+
 pub type ForeachCb<'a> = dyn FnMut(&Oid) -> bool + 'a;
 
 struct ForeachCbData<'a> {
@@ -728,4 +839,39 @@ mod tests {
         t!(repo.reset(commit1.as_object(), ResetType::Hard, None));
         assert!(foo_file.exists());
     }
+
+    #[test]
+    fn indexer() {
+        let (_td, repo_source) = crate::test::repo_init();
+        let (_td, repo_target) = crate::test::repo_init();
+
+        let mut progress_called = false;
+
+        // Create an in-memory packfile
+        let mut builder = t!(repo_source.packbuilder());
+        let mut buf = Buf::new();
+        let (commit_source_id, _tree) = crate::test::commit(&repo_source);
+        t!(builder.insert_object(commit_source_id, None));
+        t!(builder.write_buf(&mut buf));
+
+        // Write it to the standard location in the target repo, but via indexer
+        let odb = repo_source.odb().unwrap();
+        let mut indexer = odb
+            .indexer(
+                repo_target.path().join("objects").join("pack").as_path(),
+                0o644,
+            )
+            .unwrap();
+        indexer.progress(|_| {
+            progress_called = true;
+            true
+        });
+        indexer.write(&buf).unwrap();
+        indexer.commit().unwrap();
+
+        // Assert that target repo picks it up as valid
+        let commit_target = repo_target.find_commit(commit_source_id).unwrap();
+        assert_eq!(commit_target.id(), commit_source_id);
+        assert!(progress_called);
+    }
 }
