Allow org.apache names in Connect JSON schema (#123)

connectjson.Decoder deliberately panics if it encounters a schema
element whose name begins with org.apache.  Remove that behavior so
org.apache names are allowed.

(The panic is left over from development of the connectjson package,
when it wasn't clear whether to interpret values according to schema
element names during decoding.  For example, a value whose corresponding
schema element has type int64 and name
org.apache.kafka.connect.data.Timestamp could be interpreted either as a
simple int64 or, per the name, as a time in milliseconds since the Unix
epoch.  The connectjson package takes the former approach, mostly
because the current decoding process offers no easy way to perform the
necessary scaling.)

Author: nwt

connectjson/connectjson.go

@@ -9,7 +9,6 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"strings"
 
 	"github.com/brimdata/zed"
 	"github.com/brimdata/zed/runtime/sam/expr"
@@ -229,9 +228,6 @@ func (c *Decoder) Decode(b []byte) (zed.Value, error) {
 }
 
 func (c *Decoder) decodeSchema(s *connectSchema) (string, zed.Type, error) {
-	if strings.HasPrefix(s.Name, "org.apache") {
-		panic("XXX " + s.Name)
-	}
 	var typ zed.Type
 	var err error
 	switch s.Type {