lang.UnmarshalData()
Converts a murex data-type into structured memory
data, err := lang.UnmarshalData(p, dataType)
func exampleCommand(p *lang.Process) error {
data := string `{ "foo": "hello foo", "bar": "hello bar" }`
dataType := "json"
v, err := lang.UnmarshalData(p, dataType)
if err != nil {
return err
}
s := fmt.Sprint(v)
_, err := p.Stdout.Write([]byte(s))
return err
}
Go source file:
package lang
import (
"errors"
)
// UnmarshalData is a global unmarshaller which should be called from within
// murex builtin commands (etc).
// See docs/apis/marshaldata.html for more details
func UnmarshalData(p *Process, dataType string) (v interface{}, err error) {
// This is one of the very few maps in Murex which isn't hidden behind a sync
// lock of one description or other. The rational is that even mutexes can
// add a noticeable overhead on the performance of tight loops and I expect
// this function to be called _a lot_ while also only needing to be written
// to via code residing in within builtin types init() function (ie while
// murex is effectively single threaded). So there shouldn't be any data-
// races -- PROVIDING developers strictly follow the pattern of only writing
// to this map within init() func's.
if Unmarshallers[dataType] == nil {
return nil, errors.New("I don't know how to unmarshal `" + dataType + "`.")
}
v, err = Unmarshallers[dataType](p)
if err != nil {
return nil, errors.New("[" + dataType + " unmarshaller] " + err.Error())
}
return v, nil
}
*lang.Process
: Process's runtime state. Typically expressed as the variable p
string
: murex data typeMarshal()
: Converts structured memory into a structured file format (eg for stdio)Unmarshal()
: Converts a structured file format into structured memorylang.MarshalData()
: Converts structured memory into a murex data-type (eg for stdio)