Refactored decoder responsibilities (#1402)

- improved comment handling
- yaml decoder now responsible for leading content work around

pkg/yqlib/decoder_yaml.go

@@ -1,23 +1,114 @@
 package yqlib
 
 import (
+	"bufio"
+	"errors"
 	"io"
+	"regexp"
+	"strings"
 
 	yaml "gopkg.in/yaml.v3"
 )
 
 type yamlDecoder struct {
 	decoder yaml.Decoder
+	// work around of various parsing issues by yaml.v3 with document headers
+	prefs          YamlPreferences
+	leadingContent string
+	readAnything   bool
+	firstFile      bool
 }
 
-func NewYamlDecoder() Decoder {
-	return &yamlDecoder{}
+func NewYamlDecoder(prefs YamlPreferences) Decoder {
+	return &yamlDecoder{prefs: prefs, firstFile: true}
 }
 
-func (dec *yamlDecoder) Init(reader io.Reader) {
-	dec.decoder = *yaml.NewDecoder(reader)
+func (dec *yamlDecoder) processReadStream(reader *bufio.Reader) (io.Reader, string, error) {
+	var commentLineRegEx = regexp.MustCompile(`^\s*#`)
+	var sb strings.Builder
+	for {
+		peekBytes, err := reader.Peek(3)
+		if errors.Is(err, io.EOF) {
+			// EOF are handled else where..
+			return reader, sb.String(), nil
+		} else if err != nil {
+			return reader, sb.String(), err
+		} else if string(peekBytes) == "---" {
+			_, err := reader.ReadString('\n')
+			sb.WriteString("$yqDocSeperator$\n")
+			if errors.Is(err, io.EOF) {
+				return reader, sb.String(), nil
+			} else if err != nil {
+				return reader, sb.String(), err
+			}
+		} else if commentLineRegEx.MatchString(string(peekBytes)) {
+			line, err := reader.ReadString('\n')
+			sb.WriteString(line)
+			if errors.Is(err, io.EOF) {
+				return reader, sb.String(), nil
+			} else if err != nil {
+				return reader, sb.String(), err
+			}
+		} else {
+			return reader, sb.String(), nil
+		}
+	}
 }
 
-func (dec *yamlDecoder) Decode(rootYamlNode *yaml.Node) error {
-	return dec.decoder.Decode(rootYamlNode)
+func (dec *yamlDecoder) Init(reader io.Reader) error {
+	readerToUse := reader
+	leadingContent := ""
+	var err error
+	// if we 'evaluating together' - we only process the leading content
+	// of the first file - this ensures comments from subsequent files are
+	// merged together correctly.
+	if dec.prefs.LeadingContentPreProcessing && (!dec.prefs.EvaluateTogether || dec.firstFile) {
+		readerToUse, leadingContent, err = dec.processReadStream(bufio.NewReader(reader))
+		if err != nil {
+			return err
+		}
+	}
+	dec.leadingContent = leadingContent
+	dec.readAnything = false
+	dec.decoder = *yaml.NewDecoder(readerToUse)
+	dec.firstFile = false
+	return nil
+}
+
+func (dec *yamlDecoder) Decode() (*CandidateNode, error) {
+	var dataBucket yaml.Node
+
+	err := dec.decoder.Decode(&dataBucket)
+	if errors.Is(err, io.EOF) && dec.leadingContent != "" && !dec.readAnything {
+		// force returning an empty node with a comment.
+		dec.readAnything = true
+		return dec.blankNodeWithComment(), nil
+
+	} else if err != nil {
+		return nil, err
+	}
+
+	candidateNode := &CandidateNode{
+		Node: &dataBucket,
+	}
+
+	if dec.leadingContent != "" {
+		candidateNode.LeadingContent = dec.leadingContent
+		dec.leadingContent = ""
+	}
+	// move document comments into candidate node
+	// otherwise unwrap drops them.
+	candidateNode.TrailingContent = dataBucket.FootComment
+	dataBucket.FootComment = ""
+	return candidateNode, nil
+}
+
+func (dec *yamlDecoder) blankNodeWithComment() *CandidateNode {
+	return &CandidateNode{
+		Document:       0,
+		Filename:       "",
+		Node:           &yaml.Node{Kind: yaml.DocumentNode, Content: []*yaml.Node{{Tag: "!!null", Kind: yaml.ScalarNode}}},
+		FileIndex:      0,
+		LeadingContent: dec.leadingContent,
+	}
 }