README
¶
rubyobj
Provides an encoder and a decoder for Ruby ObjectSpace heap dumps.
Docs
Usage
go get github.com/aybabtme/robjspace
Then in your Go code (this shows the slow decoder):
r := yourFavoriteReader() // say a file, or stdin
rubyObj := rubyobj.RubyObject{}
var err error
for dec := rubyobj.NewDecoder(r); err == nil; err = dec.Decode(&rubyObj) {
fmt.Printf("%v\n", &rubyObj)
}
if err != io.EOF {
perror(err)
}
Performance
Using the fast ParallelDecode (using NumCPU cores and goroutines, and
fatherhood):
$ go run loadall.go parallel --filename ../testdata/huge.json
loading 549MB from 'huge.json'
2489364 heap objects in 4.699698607s
In comparison, the NewDecoder (using encoding/json):
$ go run loadall.go trivial --filename ../testdata/huge.json
loading 549MB from 'huge.json'
2489364 heap objects in 2m44.960795009s
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func ParallelDecode ¶
func ParallelDecode(r io.Reader, para uint) (<-chan RubyObject, <-chan error)
ParallelDecode will use many goroutines to decode io.Reader. io.Reader MUST present JSON objects seperated by \n characters.
Decoding will use para + 1 goroutines:
1 x goroutines to read all the lines in the io.Reader para x goroutines to decode the lines
The decoding will continue until it reaches EOF in the io.Reader, and return all the errors it encountered on the error channel, including Read errors, unmarshalling errors and loading errors.
Types ¶
type Decoder ¶
type Decoder struct {
// contains filtered or unexported fields
}
Decoder decodes RubyObjects from an io.Reader. It wraps a json.Decoder and is pretty slow, but simple to use for those acustomed to the json.Decoder from the stdlib.
func NewDecoder ¶
NewDecoder returns a trivial decoder wrapping a json.Decoder of the stdlib. It is pretty slow but simple to use.
For performance, prefer ParallelDecode.
func (*Decoder) Decode ¶
func (d *Decoder) Decode(rObj *RubyObject) (err error)
Decode decodes a ruby object from the underlying io.Reader.
type Encoder ¶
type Encoder struct {
// contains filtered or unexported fields
}
Encoder encodes RubyObjects to an io.Writer. It wraps a json.Encoder and is pretty slow, but simple to use for those acustomed to the json.Encoder from the stdlib.
func NewEncoder ¶
NewEncoder returns a trivial encoder wrapping a json.Encoder of the stdlib. It is pretty slow but simple to use.
func (*Encoder) Encode ¶
func (e *Encoder) Encode(rObj *RubyObject) (err error)
Encode encodes the object onto the underlying io.Writer.
type RubyObject ¶
type RubyObject struct {
Type RubyType
Value interface{}
Name string
NodeType string
Address uint64
Class uint64
References []uint64
Default uint64
Generation uint64
Bytesize uint64
Fd int
File string
Encoding string
Method string
Ivars uint64
Length uint64
Line uint64
Memsize uint64
Capacity uint64
Size uint64
Struct string
// contains filtered or unexported fields
}
RubyObject is the deserialized form of an object in an ObjectSpace dump.
func (RubyObject) Broken ¶
func (ro RubyObject) Broken() bool
func (RubyObject) Embedded ¶
func (ro RubyObject) Embedded() bool
func (RubyObject) Frozen ¶
func (ro RubyObject) Frozen() bool
func (RubyObject) Fstring ¶
func (ro RubyObject) Fstring() bool
func (RubyObject) GcMarked ¶
func (ro RubyObject) GcMarked() bool
func (RubyObject) GcOld ¶
func (ro RubyObject) GcOld() bool
func (RubyObject) GcWbProtected ¶
func (ro RubyObject) GcWbProtected() bool
func (RubyObject) Shared ¶
func (ro RubyObject) Shared() bool
Source Files
Directories