| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
* Add test that fails for incomplete reads from an object
* Fix imcomplete reads
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* New Object interface is distinct from core.Object
* New Object interface is used in places where returned object could be of any type
* Object is implemented by Commit, Tree, Blob, File and Tag
* Added Repository.Object function for retrieving objects of any type
* Tag.Object now returns Object instead of core.Object
* Tag target hash is now publicly accessible
* Renamed Tag.Type field to Tag.TargetType, making it distinct from Tag.Type function
* Fixed infinite recursive loop in TagIter.Close
* TreeWalker.Next now returns Object instead of core.Object
* Removed some duplicate test setup code
|
|
|
|
| |
* Old name accidentally shadowed builtin close function
|
|
|
|
|
|
|
|
|
|
|
|
| |
* New function signatures provide the necessary interface to stream data from disk when using filesystem-based storage in the future
* New function signatures provide proper error handling
* ObjectReader and ObjectWriter interfaces added to avoid future refactoring, currently are type aliases for io.ReadCloser and io.WriteCloser respectively
* Object.Reader now returns (ObjectReader, error)
* Object.Writer now returns (ObjectWriter, error)
* File.Contents now returns (string, error)
* File.Lines now returns ([]string, error)
* Blob.Reader now returns (core.ObjectReader, error)
* Added internal close helper function for deferred calls to Close that need to check the return value
|
| |
|
|
|
|
| |
Tree's mapping of names to entries has been made internal, and will only be built when necessary with the first call to Tree.File().
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of returning a channel of files, Tree.Files() now returns a
FileIter with these qualities:
* It returns files in the original order of the repository (relying on a
* new Tree.OrderedNames property)
* It can return errors encountered when retrieving files and trees from
* underlying storage
* It can be Closed without having to drain the entire channel
* It defers the heavy lifting to a new TreeWalker type
* Its behavior is a little more consistent with other Iter types
* It's a little less prone to memory leaks
This update includes a new TreeWalker type that will iterate through all
of the entries of a tree and its descendant subtrees. It does the dirty
work that Tree.walkEntries() used to do, but with a public API.
A new TreeIter type is also included that just walks through subtrees.
This could be useful for performing a directory search while ignoring
files/blobs altogether.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Previously go-git used filepath.Join when walking tree structures and constructing paths, but its results are platform-dependent as it will return different results on different systems. For example, it will use backslashes as a path separator on Windows. As a result the SuiteTree.TestFiles test was failing on Windows because the returned paths didn't match what was expected.
filepath.Join has been changed to path.Join, which will return consistent results for all platforms. This change makes it so that go-git will always return paths with forward-slash delimiters.
Users of the library should convert the resulting file paths into platform-compatible paths when necessary.
|
| |
|
| |
|
|\
| |
| | |
Performance increase for big repos
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit.File() was leaking a goroutine because it was looping over
an iterator without closing its channel.
Now commit.File() calls the new Tree.File() method that searches
the file in the repository by trasversing the dir tree instead
of using the tree.Files() iterator.
This not only prevent the goroutine leak, but also speeds up
file searching.
|
|
|
|
| |
empty dir test
|
| |
|
|
|
|
|
|
|
|
| |
This also includes a diff package and revlist package (needed by
blame)
Some extra packfiles (<1MB) are also included, to be used as fixtures in
the tests.
|
| |
|
|
|