Add support for removing K-9 Mail signatures from HTML message parts (2fe28987) · Commits · e / os / Mail

app/core/src/main/java/com/fsck/k9/helper/jsoup/AdvancedNodeTraversor.java

+1 −1

Original line number	Diff line number	Diff line
		@@ -125,7 +125,7 @@ public class AdvancedNodeTraversor {
		Node prev = node;
		node = node.nextSibling();

		if (headResult == HeadFilterDecision.REMOVE) {
		if (headResult == HeadFilterDecision.REMOVE \|\| tailResult == TailFilterDecision.REMOVE) {
		prev.remove();
		}

app/core/src/main/java/com/fsck/k9/message/signature/HtmlSignatureRemover.kt

+78 −27

Original line number	Diff line number	Diff line
		@@ -4,6 +4,7 @@ import com.fsck.k9.helper.jsoup.AdvancedNodeTraversor
		import com.fsck.k9.helper.jsoup.NodeFilter
		import com.fsck.k9.helper.jsoup.NodeFilter.HeadFilterDecision
		import com.fsck.k9.helper.jsoup.NodeFilter.TailFilterDecision
		import java.util.Stack
		import java.util.regex.Pattern
		import org.jsoup.Jsoup
		import org.jsoup.nodes.Document
		@@ -32,56 +33,106 @@ class HtmlSignatureRemover {

		private class StripSignatureFilter : NodeFilter {
		private var signatureFound = false
		private var lastElementCausedLineBreak = false
		private var brElementPrecedingDashes: Element? = null
		private var signatureParentNode: Node? = null

		override fun head(node: Node, depth: Int): HeadFilterDecision {
		if (signatureFound) return HeadFilterDecision.REMOVE

		if (node is Element) {
		lastElementCausedLineBreak = false
		if (node.tag() == BLOCKQUOTE) {
		if (node.isBlockquote()) {
		return HeadFilterDecision.SKIP_ENTIRELY
		}
		} else if (node is TextNode) {
		if (lastElementCausedLineBreak && DASH_SIGNATURE_HTML.matcher(node.wholeText).matches()) {
		val nextNode = node.nextSibling()
		if (nextNode is Element && nextNode.tag() == BR) {
		} else if (node.isSignatureDelimiter()) {
		val precedingLineBreak = node.findPrecedingLineBreak()
		if (precedingLineBreak != null && node.isFollowedByLineBreak()) {
		signatureFound = true
		brElementPrecedingDashes?.remove()
		brElementPrecedingDashes = null
		signatureParentNode = node.parent()
		precedingLineBreak.takeIf { it.isBR() }?.remove()

		return HeadFilterDecision.REMOVE
		}
		}
		}

		return HeadFilterDecision.CONTINUE
		}

		override fun tail(node: Node, depth: Int): TailFilterDecision {
		if (signatureFound) return TailFilterDecision.CONTINUE

		if (node is Element) {
		val elementIsBr = node.tag() == BR
		if (elementIsBr \|\| node.tag() == P) {
		lastElementCausedLineBreak = true
		brElementPrecedingDashes = if (elementIsBr) node else null
		if (signatureFound) {
		val signatureParentNode = this.signatureParentNode
		if (node == signatureParentNode) {
		return if (signatureParentNode.isEmpty()) {
		this.signatureParentNode = signatureParentNode.parent()
		TailFilterDecision.REMOVE
		} else {
		TailFilterDecision.STOP
		}
		}
		}

		return TailFilterDecision.CONTINUE
		}

		private fun Node.isBlockquote(): Boolean {
		return this is Element && tag() == BLOCKQUOTE
		}

		lastElementCausedLineBreak = false
		return TailFilterDecision.CONTINUE
		private fun Node.isSignatureDelimiter(): Boolean {
		return this is TextNode && DASH_SIGNATURE_HTML.matcher(wholeText).matches()
		}

		private fun Node.findPrecedingLineBreak(): Node? {
		val stack = Stack<Node>()
		stack.push(this)

		while (stack.isNotEmpty()) {
		val node = stack.pop()
		val previousSibling = node.previousSibling()
		if (previousSibling == null) {
		val parent = node.parent()
		if (parent is Element && parent.isBlock) {
		return parent
		} else {
		stack.push(parent)
		}
		} else if (previousSibling.isLineBreak()) {
		return previousSibling
		}
		}

		return null
		}

		private fun Node.isFollowedByLineBreak(): Boolean {
		val stack = Stack<Node>()
		stack.push(this)

		while (stack.isNotEmpty()) {
		val node = stack.pop()
		val nextSibling = node.nextSibling()
		if (nextSibling == null) {
		val parent = node.parent()
		if (parent is Element && parent.isBlock) {
		return true
		} else {
		stack.push(parent)
		}
		} else if (nextSibling.isLineBreak()) {
		return true
		}
		}

		return false
		}

		private fun Node?.isBR() = this is Element && tag() == BR

		private fun Node?.isLineBreak() = isBR() \|\| (this is Element && this.isBlock)

		private fun Node.isEmpty(): Boolean = childNodeSize() == 0
		}

		companion object {
		private val DASH_SIGNATURE_HTML = Pattern.compile("\\s-- \\s", Pattern.CASE_INSENSITIVE)
		private val BLOCKQUOTE = Tag.valueOf("blockquote")
		private val BR = Tag.valueOf("br")
		private val P = Tag.valueOf("p")

		@JvmStatic
		fun stripSignature(content: String): String {

app/core/src/test/java/com/fsck/k9/message/signature/HtmlSignatureRemoverTest.kt

+33 −0

Original line number	Diff line number	Diff line
		@@ -148,4 +148,37 @@ class HtmlSignatureRemoverTest {
		""".trimIndent().removeNewlines()
		)
		}

		@Test
		fun `K-9 Mail signature format`() {
		val html =
		"""
		<!DOCTYPE html>
		<html>
		<body>
		This is the body text.<br>
		<br>
		<div class='k9mail-signature'>
		-- <br>
		And this is the signature text.
		</div>
		</body>
		</html>
		""".trimIndent().removeNewlines()

		val withoutSignature = stripSignature(html)

		assertThat(withoutSignature).isEqualTo(
		"""
		<!doctype html>
		<html>
		<head></head>
		<body>
		This is the body text.<br>
		<br>
		</body>
		</html>
		""".trimIndent().removeNewlines()
		)
		}
		}