Documentation
¶
Index ¶
- Variables
- func AtomicWriteFile(filename string, data []byte, perm os.FileMode) error
- func NewAtomicFileWriter(filename string, perm os.FileMode) (io.WriteCloser, error)
- func NewCancelReadCloser(ctx context.Context, in io.ReadCloser) io.ReadCloser
- func NewReadCloserWrapper(r io.Reader, closer func() error) io.ReadCloser
- func NewReaderErrWrapper(r io.Reader, closer func()) io.Reader
- func NewWriteCloserWrapper(r io.Writer, closer func() error) io.WriteCloser
- func NopWriteCloser(w io.Writer) io.WriteCloser
- type AtomicWriteSet
- func (ws *AtomicWriteSet) Cancel() error
- func (ws *AtomicWriteSet) Commit(target string) error
- func (ws *AtomicWriteSet) FileWriter(name string, flag int, perm os.FileMode) (io.WriteCloser, error)
- func (ws *AtomicWriteSet) String() string
- func (ws *AtomicWriteSet) WriteFile(filename string, data []byte, perm os.FileMode) error
- type BytesPipe
- type NopFlusher
- type NopWriter
- type OnEOFReader
- type ReadCloserWrapper
- type WriteCounter
- type WriteFlusher
- Bugs
Constants ¶
This section is empty.
Variables ¶
var ( // ErrClosed is returned when Write is called on a closed BytesPipe. ErrClosed = errors.New("write to closed BytesPipe") )
var TempDir = longpath.MkdirTemp
TempDir is the equivalent of os.MkdirTemp, except that on Windows the result is in Windows longpath format. On Unix systems it is equivalent to os.MkdirTemp.
Deprecated: use longpath.MkdirTemp.
Functions ¶
func AtomicWriteFile ¶
AtomicWriteFile atomically writes data to a file named by filename.
func NewAtomicFileWriter ¶
NewAtomicFileWriter returns WriteCloser so that writing to it writes to a temporary file and closing it atomically changes the temporary file to destination path. Writing and closing concurrently is not allowed.
func NewCancelReadCloser ¶
func NewCancelReadCloser(ctx context.Context, in io.ReadCloser) io.ReadCloser
NewCancelReadCloser creates a wrapper that closes the ReadCloser when the context is cancelled. The returned io.ReadCloser must be closed when it is no longer needed.
func NewReadCloserWrapper ¶
func NewReadCloserWrapper(r io.Reader, closer func() error) io.ReadCloser
NewReadCloserWrapper returns a new io.ReadCloser.
func NewReaderErrWrapper ¶
NewReaderErrWrapper returns a new io.Reader.
func NewWriteCloserWrapper ¶
func NewWriteCloserWrapper(r io.Writer, closer func() error) io.WriteCloser
NewWriteCloserWrapper returns a new io.WriteCloser.
func NopWriteCloser ¶
func NopWriteCloser(w io.Writer) io.WriteCloser
NopWriteCloser returns a nopWriteCloser.
Types ¶
type AtomicWriteSet ¶
type AtomicWriteSet struct {
// contains filtered or unexported fields
}
AtomicWriteSet is used to atomically write a set of files and ensure they are visible at the same time. Must be committed to a new directory.
func NewAtomicWriteSet ¶
func NewAtomicWriteSet(tmpDir string) (*AtomicWriteSet, error)
NewAtomicWriteSet creates a new atomic write set to atomically create a set of files. The given directory is used as the base directory for storing files before commit. If no temporary directory is given the system default is used.
func (*AtomicWriteSet) Cancel ¶
func (ws *AtomicWriteSet) Cancel() error
Cancel cancels the set and removes all temporary data created in the set.
func (*AtomicWriteSet) Commit ¶
func (ws *AtomicWriteSet) Commit(target string) error
Commit moves all created files to the target directory. The target directory must not exist and the parent of the target directory must exist.
func (*AtomicWriteSet) FileWriter ¶
func (ws *AtomicWriteSet) FileWriter(name string, flag int, perm os.FileMode) (io.WriteCloser, error)
FileWriter opens a file writer inside the set. The file should be synced and closed before calling commit.
func (*AtomicWriteSet) String ¶
func (ws *AtomicWriteSet) String() string
String returns the location the set is writing to.
type BytesPipe ¶
type BytesPipe struct {
// contains filtered or unexported fields
}
BytesPipe is io.ReadWriteCloser which works similarly to pipe(queue). All written data may be read at most once. Also, BytesPipe allocates and releases new byte slices to adjust to current needs, so the buffer won't be overgrown after peak loads.
func NewBytesPipe ¶
func NewBytesPipe() *BytesPipe
NewBytesPipe creates new BytesPipe, initialized by specified slice. If buf is nil, then it will be initialized with slice which cap is 64. buf will be adjusted in a way that len(buf) == 0, cap(buf) == cap(buf).
func (*BytesPipe) CloseWithError ¶
CloseWithError causes further reads from a BytesPipe to return immediately.
type NopFlusher ¶
type NopFlusher struct{}
NopFlusher represents a type which flush operation is nop.
type OnEOFReader ¶
type OnEOFReader struct {
Rc io.ReadCloser
Fn func()
}
OnEOFReader wraps an io.ReadCloser and a function the function will run at the end of file or close the file.
func (*OnEOFReader) Close ¶
func (r *OnEOFReader) Close() error
Close closes the file and run the function.
type ReadCloserWrapper ¶
ReadCloserWrapper wraps an io.Reader, and implements an io.ReadCloser It calls the given callback function when closed. It should be constructed with NewReadCloserWrapper
func (*ReadCloserWrapper) Close ¶
func (r *ReadCloserWrapper) Close() error
Close calls back the passed closer function
type WriteCounter ¶
WriteCounter wraps a concrete io.Writer and hold a count of the number of bytes written to the writer during a "session". This can be convenient when write return is masked (e.g., json.Encoder.Encode())
func NewWriteCounter ¶
func NewWriteCounter(w io.Writer) *WriteCounter
NewWriteCounter returns a new WriteCounter.
type WriteFlusher ¶
type WriteFlusher struct {
// contains filtered or unexported fields
}
WriteFlusher wraps the Write and Flush operation ensuring that every write is a flush. In addition, the Close method can be called to intercept Read/Write calls if the targets lifecycle has already ended.
func NewWriteFlusher ¶
func NewWriteFlusher(w io.Writer) *WriteFlusher
NewWriteFlusher returns a new WriteFlusher.
func (*WriteFlusher) Close ¶
func (wf *WriteFlusher) Close() error
Close closes the write flusher, disallowing any further writes to the target. After the flusher is closed, all calls to write or flush will result in an error.
func (*WriteFlusher) Flushed ¶
func (wf *WriteFlusher) Flushed() bool
Flushed returns the state of flushed. If it's flushed, return true, or else it return false.
Notes ¶
Bugs ¶
Remove this method. Its use is inherently racy. Seems to be used to detect whether or a response code has been issued or not. Another hook should be used instead.
Source Files