p1: /v2/scrape, /v1/fetch, auth, CLI, contract tests

Single Fetcher.Fetch on the Guard that holds DenyPrefixes. skipTlsVerification
is ignored. No 502. keys add writes the secret only to --out. Release binary
has no -test-allow-prefix.

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_backslash.go

@@ -0,0 +1,9 @@
+package escape
+
+func IsBackslash(chars []byte, index int) int {
+	if chars[index] != '\\' {
+		return -1
+	}
+
+	return 1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_code.go

@@ -0,0 +1,45 @@
+package escape
+
+func IsFencedCode(chars []byte, index int) int {
+	if chars[index] != '`' && chars[index] != '~' {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == ' ' || chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == '\n' {
+			break
+		}
+
+		return -1
+	}
+
+	count := 1
+	i := index + 1
+	for ; i < len(chars); i++ {
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == '`' || chars[i] == '~' {
+			count++
+			continue
+		}
+
+		break
+	}
+	if count < 3 {
+		return -1
+	}
+
+	return i - index
+}
+
+func IsInlineCode(chars []byte, index int) int {
+	if chars[index] != '`' {
+		return -1
+	}
+
+	return 1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_divider.go

@@ -0,0 +1,47 @@
+package escape
+
+func IsDivider(chars []byte, index int) int {
+	if chars[index] != '-' && chars[index] != '_' && chars[index] != '*' {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == '\n' {
+			break
+		}
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+
+		return -1
+	}
+
+	count := 1
+	lastChar := len(chars)
+	for i := index + 1; i < len(chars); i++ {
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+		if chars[i] == chars[index] {
+			count++
+			continue
+		}
+		if chars[i] == '\n' {
+			lastChar = i
+			break
+		}
+
+		return -1
+	}
+
+	if count >= 3 {
+		return lastChar - index
+	}
+	return -1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_header.go

@@ -0,0 +1,77 @@
+package escape
+
+func IsAtxHeader(chars []byte, index int) int {
+	if chars[index] != '#' {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == '\n' {
+			break
+		}
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+		return -1
+	}
+
+	nPoundSigns := 1
+	for i := index + 1; i < len(chars); i++ {
+		if chars[i] == '#' {
+			nPoundSigns++
+
+			if nPoundSigns > 6 {
+				return -1
+			}
+			continue
+		}
+
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == ' ' || chars[i] == '\t' || chars[i] == '\n' || chars[i] == '\r' {
+			// TODO: fix calculation with placeholder (maybe own for loop construct?)
+			// Returns the count of # that we encountered
+			return i - index
+		}
+
+		return -1
+	}
+	return 1
+}
+
+func IsSetextHeader(chars []byte, index int) int {
+	if chars[index] != '=' && chars[index] != '-' {
+		return -1
+	}
+
+	var newlineCount int
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == placeholderByte || chars[i] == ' ' {
+			continue
+		}
+
+		if chars[i] == '\n' {
+			newlineCount++
+			continue
+		}
+
+		if newlineCount == 0 {
+			// Without any newlines, this character is on the same line
+			// as the delimiter. So the delimiter is inside a normal text...
+			return -1
+		} else if newlineCount == 1 {
+			// The heading content is on the line above the delimiter
+			// which qualifies for a setext heading...
+			return 1
+		} else {
+			return -1
+		}
+
+	}
+
+	return -1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_image.go

@@ -0,0 +1,36 @@
+package escape
+
+func IsImageOrLink(chars []byte, index int) int {
+	if chars[index] == '!' {
+		return isImageOrLinkStartExclamation(chars, index)
+	}
+	if chars[index] == '[' {
+		return isImageOrLinkStartBracket(chars, index)
+	}
+
+	return -1
+}
+
+func isImageOrLinkStartExclamation(chars []byte, index int) int {
+	nextIndex := index + 1
+	if nextIndex < len(chars) && chars[nextIndex] == '[' {
+		// It could be the start of an image
+		return 1
+	}
+
+	return -1
+}
+
+func isImageOrLinkStartBracket(chars []byte, index int) int {
+	for i := index + 1; i < len(chars); i++ {
+		if chars[i] == '\n' {
+			return -1
+		}
+
+		if chars[i] == ']' {
+			return 1
+		}
+	}
+
+	return -1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_italic_bold.go

@@ -0,0 +1,21 @@
+package escape
+
+import (
+	"unicode"
+)
+
+func IsItalicOrBold(chars []byte, index int) int {
+	if chars[index] != '*' && chars[index] != '_' {
+		return -1
+	}
+
+	next := getNextAsRune(chars, index)
+
+	nextIsWhitespace := unicode.IsSpace(next) || next == 0
+	if nextIsWhitespace {
+		// "not followed by Unicode whitespace"
+		return -1
+	}
+
+	return 1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_list.go

@@ -0,0 +1,67 @@
+package escape
+
+import (
+	"unicode"
+)
+
+func IsUnorderedList(chars []byte, index int) int {
+	if chars[index] != '-' && chars[index] != '*' && chars[index] != '+' {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == '\n' {
+			break
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+		if chars[i] == placeholderByte {
+			continue
+		}
+		return -1
+	}
+
+	next := getNext(chars, index)
+	if IsSpace(next) || next == 0 {
+		return 1
+	}
+
+	return -1
+}
+
+func IsOrderedList(chars []byte, index int) int {
+	if chars[index] != '.' && chars[index] != ')' {
+		return -1
+	}
+
+	// Directly before the dot needs to be a digit
+	prev := getPrevAsRune(chars, index)
+	if !unicode.IsDigit(prev) {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == '\n' {
+			break
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if IsDigit(chars[i]) {
+			continue
+		}
+
+		return -1
+	}
+
+	next := getNext(chars, index)
+	if IsSpace(next) || next == 0 {
+		return 1
+	}
+
+	return -1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/elem_quote.go

@@ -0,0 +1,23 @@
+package escape
+
+func IsBlockQuote(chars []byte, index int) int {
+	if chars[index] != '>' {
+		return -1
+	}
+
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == '\n' {
+			break
+		}
+		if chars[i] == placeholderByte {
+			continue
+		}
+		if chars[i] == ' ' {
+			continue
+		}
+
+		return -1
+	}
+
+	return 1
+}

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/replacer.go

@@ -0,0 +1,8 @@
+package escape
+
+import "github.com/JohannesKaufmann/html-to-markdown/v2/marker"
+
+var placeholderRune rune = marker.MarkerEscaping
+
+// IMPORTANT: Only internally we assume it is only byte
+var placeholderByte byte = marker.BytesMarkerEscaping[0]

vendor/github.com/JohannesKaufmann/html-to-markdown/v2/internal/escape/util.go

@@ -0,0 +1,70 @@
+package escape
+
+import "unicode/utf8"
+
+// TODO: move to markers package?
+
+func IsSpace(b byte) bool {
+	switch b {
+	case '\t', '\n', '\v', '\f', '\r', ' ', 0x85, 0xA0:
+		return true
+	}
+	return false
+}
+
+func IsDigit(b byte) bool {
+	switch b {
+	case '0', '1', '2', '3', '4', '5', '6', '7', '8', '9':
+		return true
+	}
+	return false
+}
+
+func getPrev(chars []byte, index int) byte {
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == placeholderByte {
+			continue
+		}
+		return chars[i]
+	}
+	return 0
+}
+
+func getNext(chars []byte, index int) byte {
+	for i := index + 1; i < len(chars); i++ {
+		if chars[i] == placeholderByte {
+			continue
+		}
+		return chars[i]
+	}
+	return 0
+}
+
+func getPrevAsRune(chars []byte, index int) rune {
+	for i := index - 1; i >= 0; i-- {
+		if chars[i] == placeholderByte {
+			continue
+		}
+
+		r, _ := utf8.DecodeLastRune(chars[:i+1])
+
+		return r
+	}
+	return 0
+}
+func getNextAsRune(source []byte, index int) rune {
+	for i := index + 1; i < len(source); i++ {
+		if source[i] == placeholderByte {
+			continue
+		}
+
+		r, _ := utf8.DecodeRune(source[i:])
+		return r
+	}
+	return 0
+}
+
+// TODO: make public?
+func GetNextAsRune(source []byte, index int) rune {
+	return getNextAsRune(source, index)
+}