V2 of refactor LiveText to use diff-match-patch

.gitmodules

@@ -40,6 +40,9 @@
 [submodule "include/javaAccessBridge32"]
 	path = include/javaAccessBridge32
 	url = https://github.com/nvaccess/javaAccessBridge32-bin.git
+[submodule "include/nvda_dmp"]
+	path = include/nvda_dmp
+	url = https://github.com/codeofdusk/nvda_dmp
 [submodule "include/w3c-aria-practices"]
 	path = include/w3c-aria-practices
 	url = https://github.com/w3c/aria-practices

source/NVDAObjects/IAccessible/winConsole.py

@@ -1,22 +1,45 @@
-#NVDAObjects/IAccessible/WinConsole.py
-#A part of NonVisual Desktop Access (NVDA)
-#This file is covered by the GNU General Public License.
-#See the file COPYING for more details.
-#Copyright (C) 2007-2019 NV Access Limited, Bill Dengler
+# A part of NonVisual Desktop Access (NVDA)
+# This file is covered by the GNU General Public License.
+# See the file COPYING for more details.
+# Copyright (C) 2007-2020 NV Access Limited, Bill Dengler
 
 import config
 
+from NVDAObjects.behaviors import KeyboardHandlerBasedTypedCharSupport
 from winVersion import isWin10
 
 from . import IAccessible
 from ..window import winConsole
 
-class WinConsole(winConsole.WinConsole, IAccessible):
-	"The legacy console implementation for situations where UIA isn't supported."
-	pass
+
+class EnhancedLegacyWinConsole(KeyboardHandlerBasedTypedCharSupport, winConsole.WinConsole, IAccessible):
+	"""
+		A hybrid approach to console access, using legacy APIs to read output
+		and KeyboardHandlerBasedTypedCharSupport for input.
+	"""
+	#: Legacy consoles take quite a while to send textChange events.
+	#: This significantly impacts typing performance, so don't queue chars.
+	_supportsTextChange = False
+
+
+class LegacyWinConsole(winConsole.WinConsole, IAccessible):
+	"""
+		NVDA's original console support, used by default on Windows versions
+		before 1607.
+	"""
+
+	def _get_diffAlgo(self):
+		# Non-enhanced legacy consoles use caret proximity to detect
+		# typed/deleted text.
+		# Single-character changes are not reported as
+		# they are confused for typed characters.
+		# Force difflib to keep meaningful edit reporting in these consoles.
+		from diffHandler import get_difflib_algo
+		return get_difflib_algo()
+
 
 def findExtraOverlayClasses(obj, clsList):
 	if isWin10(1607) and config.conf['terminals']['keyboardSupportInLegacy']:
-		from NVDAObjects.behaviors import KeyboardHandlerBasedTypedCharSupport
-		clsList.append(KeyboardHandlerBasedTypedCharSupport)
-	clsList.append(WinConsole)
+		clsList.append(EnhancedLegacyWinConsole)
+	else:
+		clsList.append(LegacyWinConsole)

source/NVDAObjects/UIA/winConsoleUIA.py

@@ -1,4 +1,3 @@
-# NVDAObjects/UIA/winConsoleUIA.py
 # A part of NonVisual Desktop Access (NVDA)
 # This file is covered by the GNU General Public License.
 # See the file COPYING for more details.
@@ -363,20 +362,6 @@ def _get_TextInfo(self):
 		movement."""
 		return consoleUIATextInfo if self.is21H1Plus else consoleUIATextInfoPre21H1
 
-	def _getTextLines(self):
-		if self.is21H1Plus:
-			# #11760: the 21H1 UIA console wraps across lines.
-			# When text wraps, NVDA starts reading from the beginning of the visible text for every new line of output.
-			# Use the superclass _getTextLines instead.
-			return super()._getTextLines()
-		# This override of _getTextLines takes advantage of the fact that
-		# the console text contains linefeeds for every line
-		# Thus a simple string splitlines is much faster than splitting by unit line.
-		ti = self.makeTextInfo(textInfos.POSITION_ALL)
-		text = ti.text or ""
-		return text.splitlines()
-
-
 	def detectPossibleSelectionChange(self):
 		try:
 			return super().detectPossibleSelectionChange()

source/NVDAObjects/behaviors.py

@@ -1,5 +1,4 @@
 # -*- coding: UTF-8 -*-
-# NVDAObjects/behaviors.py
 # A part of NonVisual Desktop Access (NVDA)
 # This file is covered by the GNU General Public License.
 # See the file COPYING for more details.
@@ -12,7 +11,6 @@
 import os
 import time
 import threading
-import difflib
 import tones
 import queueHandler
 import eventHandler
@@ -30,6 +28,8 @@
 import braille
 import nvwave
 import globalVars
+from typing import List
+import diffHandler
 
 
 class ProgressBar(NVDAObject):
@@ -265,15 +265,31 @@ def event_textChange(self):
 		"""
 		self._event.set()
 
-	def _getTextLines(self):
-		"""Retrieve the text of this object in lines.
+	def _get_diffAlgo(self):
+		"""
+			This property controls which diffing algorithm should be used by
+			this object. Most subclasses should simply use the base
+			implementation, which returns DMP (character-based diffing).
+			
+			@Note: DMP is experimental, and can be disallowed via user
+			preference. In this case, the prior stable implementation, Difflib
+			(line-based diffing), will be used.
+		"""
+		return diffHandler.get_dmp_algo()
+
+	def _get_devInfo(self):
+		info = super().devInfo
+		info.append(f"diffing algorithm: {self.diffAlgo}")
+		return info
+
+	def _getText(self) -> str:
+		"""Retrieve the text of this object.
 		This will be used to determine the new text to speak.
 		The base implementation uses the L{TextInfo}.
 		However, subclasses should override this if there is a better way to retrieve the text.
-		@return: The current lines of text.
-		@rtype: list of str
 		"""
-		return list(self.makeTextInfo(textInfos.POSITION_ALL).getTextInChunks(textInfos.UNIT_LINE))
+		ti = self.makeTextInfo(textInfos.POSITION_ALL)
+		return self.diffAlgo._getText(ti)
 
 	def _reportNewLines(self, lines):
 		"""
@@ -291,10 +307,10 @@ def _reportNewText(self, line):
 
 	def _monitor(self):
 		try:
-			oldLines = self._getTextLines()
+			oldText = self._getText()
 		except:
-			log.exception("Error getting initial lines")
-			oldLines = []
+			log.exception("Error getting initial text")
+			oldText = ""
 
 		while self._keepMonitoring:
 			self._event.wait()
@@ -309,71 +325,23 @@ def _monitor(self):
 			self._event.clear()
 
 			try:
-				newLines = self._getTextLines()
+				newText = self._getText()
 				if config.conf["presentation"]["reportDynamicContentChanges"]:
-					outLines = self._calculateNewText(newLines, oldLines)
+					outLines = self._calculateNewText(newText, oldText)
 					if len(outLines) == 1 and len(outLines[0].strip()) == 1:
 						# This is only a single character,
 						# which probably means it is just a typed character,
 						# so ignore it.
 						del outLines[0]
 					if outLines:
 						queueHandler.queueFunction(queueHandler.eventQueue, self._reportNewLines, outLines)
-				oldLines = newLines
+				oldText = newText
 			except:
-				log.exception("Error getting lines or calculating new text")
-
-	def _calculateNewText(self, newLines, oldLines):
-		outLines = []
-
-		prevLine = None
-		for line in difflib.ndiff(oldLines, newLines):
-			if line[0] == "?":
-				# We're never interested in these.
-				continue
-			if line[0] != "+":
-				# We're only interested in new lines.
-				prevLine = line
-				continue
-			text = line[2:]
-			if not text or text.isspace():
-				prevLine = line
-				continue
-
-			if prevLine and prevLine[0] == "-" and len(prevLine) > 2:
-				# It's possible that only a few characters have changed in this line.
-				# If so, we want to speak just the changed section, rather than the entire line.
-				prevText = prevLine[2:]
-				textLen = len(text)
-				prevTextLen = len(prevText)
-				# Find the first character that differs between the two lines.
-				for pos in range(min(textLen, prevTextLen)):
-					if text[pos] != prevText[pos]:
-						start = pos
-						break
-				else:
-					# We haven't found a differing character so far and we've hit the end of one of the lines.
-					# This means that the differing text starts here.
-					start = pos + 1
-				# Find the end of the differing text.
-				if textLen != prevTextLen:
-					# The lines are different lengths, so assume the rest of the line changed.
-					end = textLen
-				else:
-					for pos in range(textLen - 1, start - 1, -1):
-						if text[pos] != prevText[pos]:
-							end = pos + 1
-							break
-
-				if end - start < 15:
-					# Less than 15 characters have changed, so only speak the changed chunk.
-					text = text[start:end]
+				log.exception("Error getting or calculating new text")
 
-			if text and not text.isspace():
-				outLines.append(text)
-			prevLine = line
+	def _calculateNewText(self, newText: str, oldText: str) -> List[str]:
+		return self.diffAlgo.diff(newText, oldText)
 
-		return outLines
 
 class Terminal(LiveText, EditableText):
 	"""An object which both accepts text input and outputs text which should be reported automatically.

source/NVDAObjects/window/winConsole.py

@@ -1,8 +1,7 @@
-#NVDAObjects/WinConsole.py
-#A part of NonVisual Desktop Access (NVDA)
-#This file is covered by the GNU General Public License.
-#See the file COPYING for more details.
-#Copyright (C) 2007-2019 NV Access Limited, Bill Dengler
+# A part of NonVisual Desktop Access (NVDA)
+# This file is covered by the GNU General Public License.
+# See the file COPYING for more details.
+# Copyright (C) 2007-2020 NV Access Limited, Bill Dengler
 
 import winConsoleHandler
 from . import Window
@@ -14,18 +13,12 @@
 
 class WinConsole(Terminal, EditableTextWithoutAutoSelectDetection, Window):
 	"""
-		NVDA's legacy Windows Console support.
+		Base class for NVDA's legacy Windows Console support.
 		This is used in situations where UIA isn't available.
 		Please consider using NVDAObjects.UIA.winConsoleUIA instead.
 	"""
 	STABILIZE_DELAY = 0.03
 
-	def initOverlayClass(self):
-		# Legacy consoles take quite a while to send textChange events.
-		# This significantly impacts typing performance, so don't queue chars.
-		if isinstance(self, KeyboardHandlerBasedTypedCharSupport):
-			self._supportsTextChange = False
-
 	def _get_windowThreadID(self):
 		# #10113: Windows forces the thread of console windows to match the thread of the first attached process.
 		# However, To correctly handle speaking of typed characters,
@@ -69,8 +62,8 @@ def event_loseFocus(self):
 	def event_nameChange(self):
 		pass
 
-	def _getTextLines(self):
-		return winConsoleHandler.getConsoleVisibleLines()
+	def _getText(self):
+		return '\n'.join(winConsoleHandler.getConsoleVisibleLines())
 
 	def script_caret_backspaceCharacter(self, gesture):
 		super(WinConsole, self).script_caret_backspaceCharacter(gesture)

source/config/configSpec.py

@@ -225,6 +225,7 @@
 [terminals]
 	speakPasswords = boolean(default=false)
 	keyboardSupportInLegacy = boolean(default=True)
+	diffAlgo = option("auto", "dmp", "difflib", default="auto")
 
 [update]
 	autoCheck = boolean(default=true)

source/core.py

@@ -605,6 +605,14 @@ def run(self):
 	_terminate(speech)
 	_terminate(addonHandler)
 	_terminate(garbageHandler)
+	# DMP is only started if needed.
+	# Terminate manually (and let it write to the log if necessary)
+	# as core._terminate always writes an entry.
+	try:
+		import diffHandler
+		diffHandler._dmp._terminate()
+	except Exception:
+		log.exception("Exception while terminating DMP")
 
 	if not globalVars.appArgs.minimal and config.conf["general"]["playStartAndExitSounds"]:
 		try: