python3-psautohint 1.1.0-1 / usr / lib / python3 / dist-packages / psautohint / ufoFont.py

from __future__ import print_function, absolute_import

"""
ufoFont.py v1.30 May 2 2017

This module supports using the Adobe FDK tools which operate on 'bez'
files with UFO fonts. It provides low level utilities to manipulate UFO
data without fully parsing and instantiating UFO objects, and without
requiring that the AFDKO contain the robofab libraries.

Modified in Nov 2014, when AFDKO added the robofab libraries. It can now
be used with UFO fonts only to support the hash mechanism.

Developed in order to support checkOutlines and autohint, the code
supports two main functions:
- convert between UFO GLIF and bez formats
- keep a history of processing in a hash map, so that the (lengthy)
processing by autohint and checkOutlines can be avoided if the glyph has
already been processed, and the source data has not changed.

The basic model is:
 - read GLIF file
 - transform GLIF XML element to bez file
 - call FDK tool on bez file
 - transform new bez file to GLIF XML element with new data, and save in list

After all glyphs are done save all the new GLIF XML elements to GLIF
files, and update the hash map.

A complication in the Adobe UFO workflow comes from the fact we want to
make sure that checkOutlines and autohint have been run on each glyph in
a UFO font, when building an OTF font from the UFO font. We need to run
checkOutlines, because we no longer remove overlaps from source UFO font
data, because this can make revising a glyph much easier. We need to run
autohint, because the glyphs must be hinted after checkOutlines is run,
and in any case we want all glyphs to have been autohinted. The problem
with this is that it can take a minute or two to run autohint or
checkOutlines on a 2K-glyph font. The way we avoid this is to make and
keep a hash of the source glyph drawing operators for each tool. When
the tool is run, it calculates a hash of the source glyph, and compares
it to the saved hash. If these are the same, the tool can skip the
glyph. This saves a lot of time: if checkOutlines and autohint are run
on all glyphs in a font, then a second pass is under 2 seconds.

Another issue is that since we no longer remove overlaps from the source
glyph files, checkOutlines must write any edited glyph data to a
different layer in order to not destroy the source data. The ufoFont
defines an Adobe-specific glyph layer for processed glyphs, named
"glyphs.com.adobe.type.processedGlyphs".
checkOutlines writes new glyph files to the processed glyphs layer only
when it makes a change to the glyph data.

When the autohint program is run, the ufoFont must be able to tell
whether checkOutlines has been run and has altered a glyph: if so, the
input file needs to be from the processed glyph layer, else it needs to
be from the default glyph layer.

The way the hashmap works is that we keep an entry for every glyph that
has been processed, identified by a hash of its marking path data. Each
entry contains:
- a hash of the glyph point coordinates, from the default layer.
This is set after a program has been run.
- a history list: a list of the names of each program that has been run, in order.
- an editStatus flag.
Altered GLIF data is always written to the Adobe processed glyph layer. The
program may or may not have altered the outline data. For example, autohint adds
private hint data, and adds names to points, but does not change any paths.

If the stored hash for the glyph does not exist, the ufoFont lib will save the
new hash in the hash map entry and will set the history list to contain just the
current program. The program will read the glyph from the default layer.

If the stored hash matches the hash for the current glyph file in the default
layer, and the current program name is in the history list,then ufoFont
will return "skip=1", and the calling program may skip the glyph.

If the stored hash matches the hash for the current glyph file in the default
layer, and the current program name is not in the history list, then the
ufoFont will return "skip=0". If the font object field 'usedProcessedLayer' is
set True, the program will read the glyph from the from the Adobe processed
layer, if it exists, else it will always read from the default layer.

If the hash differs between the hash map entry and the current glyph in the
default layer, and usedProcessedLayer is False, then ufoFont will return
"skip=0". If usedProcessedLayer is True, then the program will consult the list
of required programs. If any of these are in the history list, then the program
will report an error and return skip =1, else it will return skip=1. The program
will then save the new hash in the hash map entry and reset the history list to
contain just the current program. If the old and new hash match, but the program
name is not in the history list, then the ufoFont will not skip the glyph, and
will add the program name to the history list.


The only tools using this are, at the moment, checkOutlines, checkOutlinesUFO
and autohint. checkOutlines and checkOutlinesUFO use the hash map to skip
processing only when being used to edit glyphs, not when reporting them.
checkOutlines necessarily flattens any components in the source glyph file to
actual outlines. autohint adds point names, and saves the hint data as a private
data in the new GLIF file.

autohint saves the hint data in the GLIF private data area, /lib/dict,
as a key and data pair. See below for the format.

autohint started with _hintFormat1_, a reasonably compact XML representation of
the data. In Sep 2105, autohhint switched to _hintFormat2_ in order to be plist
compatible. This was necessary in order to be compatible with the UFO spec, by
was driven more immediately by the fact the the UFO font file normalization
tools stripped out the _hintFormat1_ hint data as invalid elements.


"""

__copyright__ = """Copyright 2014 Adobe Systems Incorporated (http://www.adobe.com/). All Rights Reserved.
"""

_hintFormat1_ = """

Deprecated. See _hintFormat2_ below.

A <hintset> element identifies a specific point by its name, and
describes a new set of stem hints which should be applied before the
specific point.

A <flex> element identifies a specific point by its name. The point is
the first point of a curve. The presence of the <flex> element is a
processing suggestion, that the curve and its successor curve should
be converted to a flex operator.

One challenge in applying the hintset and flex elements is that in the
GLIF format, there is no explicit start and end operator: the first path
operator is both the end and the start of the path. I have chosen to
convert this to T1 by taking the first path operator, and making it a
move-to. I then also use it as the last path operator. An exception is a
line-to; in T1, this is omitted, as it is implied by the need to close
the path. Hence, if a hintset references the first operator, there is a
potential ambiguity: should it be applied before the T1 move-to, or
before the final T1 path operator? The logic here applies it before the
move-to only.

 <glyph>
...
	<lib>
		<dict>
			<key><com.adobe.type.autohint><key>
			<data>
				<hintSetList>
					<hintset pointTag="point name">
						(<hstem pos="<decimal value>" width="decimal value" />)*
						(<vstem pos="<decimal value>" width="decimal value" />)*
						<hstem3 stem3List="n0,n1,n2,n3,n4,n5" />* # where n1-5 are decimal values
						<vstem3 stem3List="n0,n1,n2,n3,n4,n5" />* # where n1-5 are decimal values
					</hintset>*
					(<hintSetList>*
						(<hintset pointIndex="positive integer">
							(<stemindex>positive integer</stemindex>)+
						</hintset>)+
					</hintSetList>)*
					<flexList>
						<flex pointTag="point Name" />
					</flexList>*
				</hintSetList>
			</data>
		</dict>
	</lib>
</glyph>

Example from "B" in SourceCodePro-Regular
<key><com.adobe.type.autohint><key>
<data>
<hintSetList id="64bf4987f05ced2a50195f971cd924984047eb1d79c8c43e6a0054f59cc85dea23a49deb20946a4ea84840534363f7a13cca31a81b1e7e33c832185173369086">
	<hintset pointTag="hintSet0000">
		<hstem pos="0" width="28" />
		<hstem pos="338" width="28" />
		<hstem pos="632" width="28" />
		<vstem pos="100" width="32" />
		<vstem pos="496" width="32" />
	</hintset>
	<hintset pointTag="hintSet0005">
		<hstem pos="0" width="28" />
		<hstem pos="338" width="28" />
		<hstem pos="632" width="28" />
		<vstem pos="100" width="32" />
		<vstem pos="454" width="32" />
		<vstem pos="496" width="32" />
	</hintset>
	<hintset pointTag="hintSet0016">
		<hstem pos="0" width="28" />
		<hstem pos="338" width="28" />
		<hstem pos="632" width="28" />
		<vstem pos="100" width="32" />
		<vstem pos="496" width="32" />
	</hintset>
</hintSetList>
</data>

"""

_hintFormat2_ = """

A <dict> element in the hintSetList array identifies a specific point by its
name, and describes a new set of stem hints which should be applied before the
specific point.

A <string> element in the flexList identifies a specific point by its name. The
point is the first point of a curve. The presence of the element is a processing
suggestion, that the curve and its successor curve should be converted to a flex
operator.

One challenge in applying the hintSetList and flexList elements is that in the
GLIF format, there is no explicit start and end operator: the first path
operator is both the end and the start of the path. I have chosen to convert
this to T1 by taking the first path operator, and making it a move-to. I then
also use it as the last path operator. An exception is a line-to; in T1, this is
omitted, as it is implied by the need to close the path. Hence, if a hintset
references the first operator, there is a potential ambiguity: should it be
applied before the T1 move-to, or before the final T1 path operator? The logic
here applies it before the move-to only.
 <glyph>
...
	<lib>
		<dict>
			<key><com.adobe.type.autohint></key>
			<dict>
				<key>id</key>
				<string> <fingerprint for glyph> </string>
				<key>hintSetList</key>
				<array>
					<dict>
						<key>pointTag</key>
						<string> <point name> </string>
						<key>stems</key>
						<array>
							<string>hstem <position value> <width value></string>*
							<string>vstem <position value> <width value></string>*
							<string>hstem3 <position value 0>...<position value 5></string>*
							<string>vstem3 <position value 0>...<position value 5></string>*
						</array>
					</dict>*
				</array>

				<key>flexList</key>*
				<array>
					<string><point name></string>+
				</array>
			</dict>
		</dict>
	</lib>
</glyph>

Example from "B" in SourceCodePro-Regular
<key><com.adobe.type.autohint><key>
<dict>
	<key>id</key>
	<string>64bf4987f05ced2a50195f971cd924984047eb1d79c8c43e6a0054f59cc85dea23a49deb20946a4ea84840534363f7a13cca31a81b1e7e33c832185173369086</string>
	<key>hintSetList</key>
	<array>
		<dict>
			<key>pointTag</key>
			<string>hintSet0000</string>
			<key>stems</key>
			<array>
				<string>hstem 338 28</string>
				<string>hstem 632 28</string>
				<string>hstem 100 32</string>
				<string>hstem 496 32</string>
			</array>
		</dict>
		<dict>
			<key>pointTag</key>
			<string>hintSet0005</string>
			<key>stems</key>
			<array>
				<string>hstem 0 28</string>
				<string>hstem 338 28</string>
				<string>hstem 632 28</string>
				<string>hstem 100 32</string>
				<string>hstem 454 32</string>
				<string>hstem 496 32</string>
			</array>
		</dict>
		<dict>
			<key>pointTag</key>
			<string>hintSet0016</string>
			<key>stems</key>
			<array>
				<string>hstem 0 28</string>
				<string>hstem 338 28</string>
				<string>hstem 632 28</string>
				<string>hstem 100 32</string>
				<string>hstem 496 32</string>
			</array>
		</dict>
	</array>
<dict>

"""

import re
import os
import plistlib
import hashlib

try:
	import xml.etree.cElementTree as ET
except ImportError:
	import xml.etree.ElementTree as ET

from psautohint import fdTools

XML = ET.XML
XMLElement = ET.Element
xmlToString = ET.tostring
debug = False

def debugMsg(*args):
	if debug:
		print(args)

#UFO names
kDefaultGlyphsLayerName = "public.default"
kDefaultGlyphsLayer = "glyphs"
kProcessedGlyphsLayerName = "AFDKO ProcessedGlyphs"
kProcessedGlyphsLayer = "glyphs.com.adobe.type.processedGlyphs"
kFontInfoName = "fontinfo.plist"
kContentsName = "contents.plist"
kLibName = "lib.plist"
kPublicGlyphOrderKey = "public.glyphOrder"

kAdobeDomainPrefix = "com.adobe.type"
kAdobHashMapName = "%s.processedHashMap" % (kAdobeDomainPrefix)
kAdobHashMapVersionName = "hashMapVersion"
kAdobHashMapVersion = (1,0) # If major version differs, do not use.
kAutohintName = "autohint"
kCheckOutlineName = "checkOutlines"
kCheckOutlineNameUFO = "checkOutlines"

kOutlinePattern = re.compile(r"<outline.+?outline>", re.DOTALL)

kStemHintsName = "stemhints"
kStemListName = "stemList"
kStemPosName = "pos"
kStemWidthName = "width"
kHStemName = "hstem"
kVStemName = "vstem"
kHStem3Name = "hstem3"
kVStem3Name = "vstem3"
kStem3Pos = "stem3List"
kHintSetListName = "hintSetList"
kFlexListName = "hintSetList"
kHintSetName = "hintset"
kBaseFlexName = "flexCurve"
kPointTag = "pointTag"
kStemIndexName = "stemindex"
kFlexIndexListName = "flexList"
kHintDomainName1 = "com.adobe.type.autohint"
kHintDomainName2 = "com.adobe.type.autohint.v2"
kPointName = "name"
# Hint stuff
kStackLimit = 46
kStemLimit = 96

kDefaultFMNDBPath = "FontMenuNameDB"
kDefaultGOADBPath = "GlyphOrderAndAliasDB"


class UFOParseError(ValueError):
	pass

class BezParseError(ValueError):
	pass

class UFOFontData:
	def __init__(self, parentPath, useHashMap, programName):
		self.parentPath = parentPath
		self.glyphMap = {}
		self.processedLayerGlyphMap = {}
		self.newGlyphMap = {}
		self.glyphList = []
		self.fontInfo = None
		self.useHashMap = useHashMap # if False, will skip reading has map and testing to see if glyph can be skipped.
		# should be used only when calling program is running in report mdoe only, and not changing any glyph data.
		self.hashMap = {} # Used to skip getting glyph data when glyph hash matches hash of current glyph data.
		self.fontDict = None
		self.programName = programName
		self.curSrcDir = None
		self.hashMapChanged = False
		self.glyphDefaultDir = os.path.join(parentPath, "glyphs")
		self.glyphLayerDir = os.path.join(parentPath, kProcessedGlyphsLayer)
		self.glyphWriteDir = self.glyphLayerDir
		self.historyList = []
		self.requiredHistory = [] # See documentation above.
		self.useProcessedLayer = False # If False, then read data only from the default layer; else read glyphs from processed layer, if it exists.
		self.writeToDefaultLayer = False # If True, then write data to the default layer
		self.doAll = False # if True, then do not skip any glyphs.
		self.deletedGlyph = False # track whether checkSkipGLyph has deleted a out of date glyph from the processed glyph layer
		self.allowDecimalCoords = False # if true, do NOT round x,y values when processing.

	def getUnitsPerEm(self):
		unitsPerEm = "1000"
		if self.fontInfo == None:
			self.loadFontInfo()
		if self.fontInfo:
			unitsPerEm = int(self.fontInfo["unitsPerEm"])
		unitsPerEm = int(unitsPerEm)
		return unitsPerEm

	def getPSName(self):
		psName = "PSName-Undefined"
		if self.fontInfo == None:
			self.loadFontInfo()
		if self.fontInfo:
			psName = self.fontInfo.get("postscriptFontName", psName)
		return psName

	def isCID(self):
		return 0

	def checkForHints(self, glyphName):
		hasHints = False
		glyphPath = self.getGlyphProcessedPath(glyphName)
		if glyphPath and os.path.exists(glyphPath):
			with open(glyphPath, "rt") as fp:
				data = fp.read()
			if "hintSetList" in data:
				hasHints = True
		return hasHints

	def convertToBez(self, glyphName, beVerbose, doAll=False):
		# convertGLIFToBez does not yet support hints - no need for removeHints arg.
		bezString, width = convertGLIFToBez(self, glyphName, beVerbose, doAll)
		hasHints = self.checkForHints(glyphName)
		return bezString, width, hasHints

	def updateFromBez(self, bezData, glyphName, width, beVerbose):
		# For UFO font, we don't use the width parameter: it is carried over from the input glif file.
		glifXML = convertBezToGLIF(self, glyphName, bezData )
		self.newGlyphMap[glyphName] = glifXML

	def saveChanges(self):
		if self.hashMapChanged:
			self.writeHashMap()
		self.hashMapChanged = False

		if not os.path.exists(self.glyphWriteDir):
			os.makedirs(self.glyphWriteDir)

		for glyphName, glifXML in self.newGlyphMap.items():
			glyphPath = self.getWriteGlyphPath(glyphName)
			#print("Saving file", glyphPath)
			et = ET.ElementTree(glifXML)
			with open(glyphPath, "wb") as fp:
				et.write(fp, encoding="UTF-8", xml_declaration=True)

		# Update the layer contents.plist file
		layerContentsFilePath = os.path.join(self.parentPath, "layercontents.plist")
		self.updateLayerContents(layerContentsFilePath)
		glyphContentsFilePath = os.path.join(self.glyphWriteDir, "contents.plist")
		self.updateLayerGlyphContents(glyphContentsFilePath, self.newGlyphMap)

	def getWriteGlyphPath(self, glyphName):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()

		glyphFileName = self.glyphMap[glyphName]
		if not self.writeToDefaultLayer:
			if self.processedLayerGlyphMap:
				try:
					glyphFileName = self.processedLayerGlyphMap[glyphName]
				except KeyError:
					pass

		glifFilePath = os.path.join(self.glyphWriteDir, glyphFileName)
		return glifFilePath

	def getGlyphMap(self):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		return self.glyphMap

	def readHashMap(self):
		hashPath = os.path.join(self.parentPath, "data", kAdobHashMapName)
		if os.path.exists(hashPath):
			with open(hashPath, "rt") as fp:
				data = fp.read()
			newMap = eval(data)
		else:
			newMap = {kAdobHashMapVersionName:kAdobHashMapVersion}

		try:
			version = newMap[kAdobHashMapVersionName]
			if version[0] > kAdobHashMapVersion[0]:
				raise UFOParseError("Hash map version is newer than program. Please update the FDK")
			elif version[0] < kAdobHashMapVersion[0]:
				print("Updating hash map: was older version")
				newMap = {kAdobHashMapVersionName:kAdobHashMapVersion}
		except KeyError:
			print("Updating hash map: was older version")
			newMap = {kAdobHashMapVersionName:kAdobHashMapVersion}
		self.hashMap = newMap
		return

	def writeHashMap(self):
		hashMap = self.hashMap
		if len(hashMap) == 0:
			return # no glyphs were processed.

		hashDir = os.path.join(self.parentPath, "data")
		if not os.path.exists(hashDir):
			os.makedirs(hashDir)
		hashPath = os.path.join(hashDir, kAdobHashMapName)

		hasMapKeys = hashMap.keys()
		hasMapKeys = sorted(hasMapKeys)
		data = ["{"]
		for gName in hasMapKeys:
			data.append("'%s': %s," % (gName, hashMap[gName]))
		data.append("}")
		data.append("")
		data = "\n".join(data)
		with open(hashPath, "wt") as fp:
			fp.write(data)
		return

	def getCurGlyphPath(self, glyphName):
		if self.curSrcDir == None:
			self.curSrcDir = self.glyphDefaultDir

		# Get the glyph file name.
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		glyphFileName = self.glyphMap[glyphName]
		path = os.path.join(self.curSrcDir, glyphFileName)
		return path

	def getGlyphSrcPath(self, glyphName):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		glyphFileName = self.glyphMap[glyphName]

		if self.useProcessedLayer and self.processedLayerGlyphMap: # Try for processed layer first.
			try:
				glyphFileName = self.processedLayerGlyphMap[glyphName]
				self.curSrcDir = self.glyphLayerDir
				glyphPath = os.path.join(self.glyphLayerDir, glyphFileName)
				if os.path.exists(glyphPath):
					return glyphPath
			except KeyError:
				pass

		self.curSrcDir = self.glyphDefaultDir
		glyphPath = os.path.join(self.curSrcDir, glyphFileName)

		return glyphPath

	def getGlyphDefaultPath(self, glyphName):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		glyphFileName = self.glyphMap[glyphName]
		glyphPath = os.path.join(self.glyphDefaultDir, glyphFileName)
		return glyphPath

	def getGlyphProcessedPath(self, glyphName):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		if not self.processedLayerGlyphMap:
			return None

		try:
			glyphFileName = self.processedLayerGlyphMap[glyphName]
			glyphPath = os.path.join(self.glyphLayerDir, glyphFileName)
		except KeyError:
			glyphPath = None
		return glyphPath

	def updateHashEntry(self, glyphName, changed):
		# srcHarsh has already been set: we are fixing the history list.
		if not self.useHashMap:
			return
		# Get hash entry for glyph
		try:
			hashEntry = self.hashMap[glyphName]
			srcHash, historyList = hashEntry
		except KeyError:
			hashEntry = None

		self.hashMapChanged = True
		# If the program always reads data from the default layer, and we have just created a new glyph in the processed layer, then reset the history.
		if (not self.useProcessedLayer) and changed:
			self.hashMap[glyphName] = [srcHash, [self.programName] ]
			return
		else:
			try:
				programHistoryIndex = historyList.index(self.programName)
			except ValueError:
				#If the program is not in the history list, add it.
				historyList.append(self.programName)


	def checkSkipGlyph(self, glyphName, newSrcHash, doAll):
		skip = False
		if not self.useHashMap:
			return skip

		if len(self.hashMap) == 0:
			# Hash maps have not yet been read in. Get them.
			self.readHashMap()

		hashEntry = srcHash = None
		historyList = []
		programHistoryIndex = -1 # not found in historyList

		# Get hash entry for glyph
		try:
			hashEntry = self.hashMap[glyphName]
			srcHash, historyList = hashEntry
			try:
				programHistoryIndex = historyList.index(self.programName)
			except ValueError:
				pass
		except KeyError:
			# Glyph is as yet untouched by any program.
			pass

		if (srcHash == newSrcHash):
			if (programHistoryIndex >= 0):
				# The glyph has already been processed by this program, and there have been no changes since.
				skip = True and (not doAll)
			if not skip:
				if not self.useProcessedLayer: # case for Checkoutlines
					self.hashMapChanged = True
					self.hashMap[glyphName] = [newSrcHash, [self.programName] ]
					glyphPath = self.getGlyphProcessedPath(glyphName)
					if glyphPath and os.path.exists(glyphPath):
						os.remove(glyphPath)
				else:
					if (programHistoryIndex < 0):
						historyList.append(self.programName)
		else:
			if self.useProcessedLayer: # case for autohint
				# default layer glyph and stored glyph hash differ,and useProcessedLayer is True
				# If any of the programs in requiredHistory in are in the historyList, we need to complain and skip.
				foundMatch = False
				if len(historyList) > 0:
					for programName in self.requiredHistory:
						if programName in historyList:
							foundMatch = True
				if foundMatch:
					print("Error. Glyph '%s' has been edited. You must first run '%s' before running '%s'. Skipping." % (glyphName, self.requiredHistory, self.programName))
					skip = True

			# If the source hash has changed, we need to delete the processed layer glyph.
			self.hashMapChanged = True
			self.hashMap[glyphName] = [newSrcHash, [self.programName] ]
			glyphPath = self.getGlyphProcessedPath(glyphName)
			if glyphPath and os.path.exists(glyphPath):
				os.remove(glyphPath)
				self.deletedGlyph = True

		return skip

	def getGlyphXML(self, glyphDir, glyphFileName):
		glyphPath = os.path.join(glyphDir, glyphFileName) # default
		etRoot = ET.ElementTree()
		glifXML = etRoot.parse(glyphPath)
		outlineXML = glifXML.find("outline")
		try:
			widthXML = glifXML.find("advance")
			if widthXML != None:
				width = int(eval(widthXML.get("width")))
			else:
				width = 1000
		except UFOParseError as e:
			print("Error. skipping glyph '%s' because of parse error: %s" % (glyphName, e.message))
			return None, None, None
		return width, glifXML, outlineXML

	def getOrSkipGlyphXML(self, glyphName, doAll=False):
		# Get default glyph layer data, so we can check if the glyph has been edited since this program was last run.
		# If the program name is in the history list, and the srcHash matches the default glyph layer data, we can skip.
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		glyphFileName = self.glyphMap[glyphName]
		width, glifXML, outlineXML = self.getGlyphXML(self.glyphDefaultDir, glyphFileName)
		if glifXML == None:
			skip = True
			return None, None, skip

		useDefaultGlyphDir = True # Hash is always from the default glyph layer.
		newHash, dataList = self.buildGlyphHashValue(width, outlineXML, glyphName, useDefaultGlyphDir)
		skip = self.checkSkipGlyph(glyphName, newHash, doAll)

		# If self.useProcessedLayer and there is a glyph in the processed layer, get the outline from that.
		if self.useProcessedLayer and self.processedLayerGlyphMap:
			try:
				glyphFileName = self.processedLayerGlyphMap[glyphName]
			except KeyError:
				pass
			glyphPath = os.path.join(self.glyphLayerDir, glyphFileName)
			if os.path.exists(glyphPath):
				width, glifXML, outlineXML = self.getGlyphXML(self.glyphLayerDir, glyphFileName)
				if glifXML == None:
					skip = True
					return None, None, skip

		return width, outlineXML, skip


	def getGlyphList(self):
		if len(self.glyphMap) == 0:
			self.loadGlyphMap()
		return self.glyphList

	def loadGlyphMap(self):
		# Need to both get the list of glyphs from contents.plist, and also the glyph order.
		# the latter is take from the public.glyphOrder key in lib.py, if it exists,
		# else it is taken from the contents.plist file. Any glyphs in contents.plist
		# which are not named in the public.glyphOrder are sorted after all glyphs which are named
		# in the public.glyphOrder,, in the order that they occured in contents.plist.
		contentsPath = os.path.join(self.parentPath, "glyphs", kContentsName)
		self.glyphMap, self.glyphList = parsePList(contentsPath)
		orderPath = os.path.join(self.parentPath, kLibName)
		self.orderMap = parseGlyphOrder(orderPath)
		if self.orderMap != None:
			orderIndex = len(self.orderMap)
			orderList = []

			# If there are glyphs in the font which are not named in the public.glyphOrder entry, add then in the order of the contents.plist file.
			for glyphName in self.glyphList:
				try:
					entry = [self.orderMap[glyphName], glyphName]
				except KeyError:
					entry = [orderIndex, glyphName]
					self.orderMap[glyphName] = orderIndex
					orderIndex += 1
				orderList.append(entry)
			orderList.sort()
			self.glyphList = []
			for entry in orderList:
				self.glyphList.append(entry[1])
		else:
			self.orderMap = {}
			numGlyphs = len(self.glyphList)
			for i in range(numGlyphs):
				self.orderMap[self.glyphList[i]] = i

		# I also need to get the glyph map for the processed layer,
		# and use this when the glyph is read from the processed layer.
		# Because checkOutliensUFO used the defcon library, it can write glyph file names
		# that differ from what is in the default glyph layer.
		contentsPath = os.path.join(self.glyphLayerDir, kContentsName)
		if os.path.exists(contentsPath):
			self.processedLayerGlyphMap, self.processedLayerGlyphList = parsePList(contentsPath)



	def loadFontInfo(self):
		fontInfoPath = os.path.join(self.parentPath, "fontinfo.plist")
		if not os.path.exists(fontInfoPath):
			return
		self.fontInfo, tempList = parsePList(fontInfoPath)

	def updateLayerContents(self, contentsFilePath):
		if os.path.exists(contentsFilePath):
			contentsList= plistlib.readPlist(contentsFilePath)
			# If the layer name already exists, don't add a new one, or change the path
			seenPublicDefault = False
			seenProcessedGlyph = False
			for layerName, layerPath in contentsList:
				if (layerPath == kProcessedGlyphsLayer):
					seenProcessedGlyph = True
				if (layerPath == kDefaultGlyphsLayer):
					seenPublicDefault = True
			update = False
			if not seenPublicDefault:
				update = True
				contentsList = [[kDefaultGlyphsLayerName, kDefaultGlyphsLayer]] + contentsList
			if not seenProcessedGlyph:
				update = True
				contentsList.append([kProcessedGlyphsLayerName, kProcessedGlyphsLayer])
			if update:
				plistlib.writePlist(contentsList, contentsFilePath)
		else:
			contentsList = [[kDefaultGlyphsLayerName, kDefaultGlyphsLayer]]
			contentsList.append([kProcessedGlyphsLayerName, kProcessedGlyphsLayer])
		plistlib.writePlist(contentsList, contentsFilePath)

	def updateLayerGlyphContents(self, contentsFilePath, newGlyphData):
		if os.path.exists(contentsFilePath):
			contentsDict = plistlib.readPlist(contentsFilePath)
		else:
			contentsDict = {}
		for glyphName in newGlyphData.keys():
			if self.useProcessedLayer and self.processedLayerGlyphMap: # Try for processed layer first.
				try:
					contentsDict[glyphName] = self.processedLayerGlyphMap[glyphName]
				except KeyError:
					contentsDict[glyphName] = self.glyphMap[glyphName]
			else:
				contentsDict[glyphName] = self.glyphMap[glyphName]
		plistlib.writePlist(contentsDict, contentsFilePath)

	def getFontInfo(self, fontPSName, inputPath, allow_no_blues, noFlex, vCounterGlyphs, hCounterGlyphs, fdIndex=0):
		if self.fontDict != None:
			return self.fontDict

		if self.fontInfo == None:
			self.loadFontInfo()

		fdDict = fdTools.FDDict()
		fdDict.LanguageGroup = self.fontInfo.get("languagegroup", "0") # should be 1 if the glyphs are ideographic, else 0.
		fdDict.OrigEmSqUnits = self.getUnitsPerEm()
		fdDict.FontName = self.getPSName()
		upm = self.getUnitsPerEm()
		low = min( -upm*0.25, float(self.fontInfo.get("openTypeOS2WinDescent", "0")) - 200)
		high = max ( upm*1.25, float(self.fontInfo.get("openTypeOS2WinAscent", "0")) + 200)
		# Make a set of inactive alignment zones: zones outside of the font bbox so as not to affect hinting.
		# Used when src font has no BlueValues or has invalid BlueValues. Some fonts have bad BBox values, so
		# I don't let this be smaller than -upm*0.25, upm*1.25.
		inactiveAlignmentValues = [low, low, high, high]
		blueValues = self.fontInfo.get("postscriptBlueValues", [])
		numBlueValues = len(blueValues)
		if numBlueValues < 4:
			if allow_no_blues:
				blueValues = inactiveAlignmentValues
				numBlueValues = len(blueValues)
			else:
				raise	UFOParseError("ERROR: Font must have at least four values in its BlueValues array for AC to work!")
		blueValues.sort()
		# The first pair only is a bottom zone, where the first value is the overshoot position;
		# the rest are top zones, and second value of the pair is the overshoot position.
		blueValues[0] = blueValues[0] - blueValues[1]
		for i in range(3, numBlueValues,2):
			blueValues[i] = blueValues[i] - blueValues[i-1]

		blueValues = [str(v) for v in blueValues]
		numBlueValues = min(numBlueValues, len(fdTools.kBlueValueKeys))
		for i in range(numBlueValues):
			key = fdTools.kBlueValueKeys[i]
			value = blueValues[i]
			exec("fdDict.%s = %s" % (key, value))

		otherBlues = self.fontInfo.get("postscriptOtherBlues", [])
		numBlueValues = len(otherBlues)

		if len(otherBlues) > 0:
			i = 0
			numBlueValues = len(otherBlues)
			otherBlues.sort()
			for i in range(0, numBlueValues,2):
				otherBlues[i] = otherBlues[i] - otherBlues[i+1]
			otherBlues = [str(v) for v in otherBlues]
			numBlueValues = min(numBlueValues, len(fdTools.kOtherBlueValueKeys))
			for i in range(numBlueValues):
				key = fdTools.kOtherBlueValueKeys[i]
				value = otherBlues[i]
				exec("fdDict.%s = %s" % (key, value))

		vstems = self.fontInfo.get("postscriptStemSnapV", [])
		if len(vstems) == 0:
			if allow_no_blues:
				vstems = [fdDict.OrigEmSqUnits] # dummy value. Needs to be larger than any hint will likely be,
				# as the autohint program strips out any hint wider than twice the largest global stem width.
			else:
				raise	UFOParseError("ERROR: Font does not have postscriptStemSnapV!")

		vstems.sort()
		if (len(vstems) == 0) or ((len(vstems) == 1) and (vstems[0] < 1) ):
			vstems = [fdDict.OrigEmSqUnits] # dummy value that will allow PyAC to run
			logMsg("WARNING: There is no value or 0 value for DominantV.")
		fdDict.DominantV = "[" + " ".join([str(v) for v in vstems]) + "]"

		hstems = self.fontInfo.get("postscriptStemSnapH", [])
		if len(hstems) == 0:
			if allow_no_blues:
				hstems = [fdDict.OrigEmSqUnits] # dummy value. Needs to be larger than any hint will likely be,
				# as the autohint program strips out any hint wider than twice the largest global stem width.
			else:
				raise	UFOParseError("ERROR: Font does not have postscriptStemSnapH!")

		hstems.sort()
		if (len(hstems) == 0) or ((len(hstems) == 1) and (hstems[0] < 1) ):
			hstems = [fdDict.OrigEmSqUnits] # dummy value that will allow PyAC to run
			logMsg("WARNING: There is no value or 0 value for DominantH.")
		fdDict.DominantH = "[" + " ".join([str(v) for v in hstems]) + "]"

		if noFlex:
			fdDict.FlexOK = "false"
		else:
			fdDict.FlexOK = "true"

		# Add candidate lists for counter hints, if any.
		if vCounterGlyphs:
			temp = " ".join(vCounterGlyphs)
			fdDict.VCounterChars = "( %s )" % (temp)
		if hCounterGlyphs:
			temp = " ".join(hCounterGlyphs)
			fdDict.HCounterChars = "( %s )" % (temp)

		fdDict.BlueFuzz = self.fontInfo.get("postscriptBlueFuzz", 1)
		# postscriptBlueShift
		# postscriptBlueScale
		self.fontDict = fdDict
		return fdDict

	def getfdIndex(self, gid):
		fdIndex = 0
		return fdIndex

	def getfdInfo(self, psName, inputPath, allow_no_blues, noFlex, vCounterGlyphs, hCounterGlyphs, glyphList, fdIndex=0):
		fontDictList = []
		fdGlyphDict = None
		fdDict = self.getFontInfo(psName, inputPath, allow_no_blues, noFlex, vCounterGlyphs, hCounterGlyphs, fdIndex)
		fontDictList.append(fdDict)

		# Check the fontinfo file, and add any other font dicts
		srcFontInfo = os.path.dirname(inputPath)
		srcFontInfo = os.path.join(srcFontInfo, "fontinfo")
		maxX = self.getUnitsPerEm()*2
		maxY = maxX
		minY = -self.getUnitsPerEm()
		if os.path.exists(srcFontInfo):
			with open(srcFontInfo, "rU") as fi:
				fontInfoData = fi.read()
			fontInfoData = re.sub(r"#[^\r\n]+", "", fontInfoData)

			if "FDDict" in fontInfoData:

				blueFuzz = fdDict.BlueFuzz
				fdGlyphDict, fontDictList, finalFDict = fdTools.parseFontInfoFile(fontDictList, fontInfoData, glyphList, maxY, minY, psName, blueFuzz)
				if finalFDict == None:
					# If a font dict was not explicitly specified for the output font, use the first user-specified font dict.
					fdTools.mergeFDDicts( fontDictList[1:], self.fontDict )
				else:
					fdTools.mergeFDDicts( [finalFDict], topDict )

		return fdGlyphDict, fontDictList

	def getGlyphID(self, glyphName):
		try:
			gid = self.orderMap[glyphName]
		except IndexError:
			raise UFOParseError("Could not find glyph name '%s' in UFO font contents plist. '%s'. " % (glyphName, self.parentPath))
		return gid

	def buildGlyphHashValue(self, width, outlineXML, glyphName, useDefaultGlyphDir, level = 0):
		"""
		glyphData must be the official <outline> XML from a GLIF.
		We skip contours with only one point.
		"""
		dataList = ["w%s" % (str(width))]
		if level > 10:
			raise UFOParseError("In parsing component, exceeded 10 levels of reference. '%s'. " % (glyphName))
		# <outline> tag is optional per spec., e.g. space glyph does not necessarily have it.
		if outlineXML:
			for childContour in outlineXML:
				if childContour.tag == "contour":
					if len(childContour) < 2:
						continue
					else:
						for child in childContour:
							if child.tag == "point":
								try:
									pointType = child.attrib["type"][0]
								except KeyError:
									pointType = ""
								dataList.append("%s%s%s" % (pointType, child.attrib["x"], child.attrib["y"]))
								#print(dataList[-3:])
				elif childContour.tag == "component":
					# append the component hash.
					try:
						compGlyphName = childContour.attrib["base"]
						dataList.append("%s%s" % ("base:", compGlyphName))
					except KeyError:
						raise UFOParseError("'%s' is missing the 'base' attribute in a component. glyph '%s'." % (glyphName))

					if useDefaultGlyphDir:
						try:
							componentPath = self.getGlyphDefaultPath(compGlyphName)
						except KeyError:
							raise UFOParseError("'%s' component glyph is missing from contents.plist." % (compGlyphName))
					else:
						# If we are not necessarily using the default layer for the main glyph, then a missing component
						# may not have been processed, and may just be in the default layer. We need to look for component
						# glyphs in the src list first, then in the defualt layer.
						try:
							componentPath = self.getGlyphSrcPath(compGlyphName)
							if not os.path.exists(componentPath):
								componentPath = self.getGlyphDefaultPath(compGlyphName)
						except KeyError:
							try:
								componentPath = self.getGlyphDefaultPath(compGlyphName)
							except KeyError:
								raise UFOParseError("'%s' component glyph is missing from contents.plist." % (compGlyphName))

					if not os.path.exists(componentPath):
						raise UFOParseError("'%s' component file is missing: '%s'." % (compGlyphName, componentPath))

					etRoot = ET.ElementTree()

					# Collect transformm fields, if any.
					for transformTag in ["xScale", "xyScale", "yxScale", "yScale", "xOffset", "yOffset"]:
						try:
							value = childContour.attrib[transformTag]
							rval = eval(value)
							if int(rval) == rval:
								value = str(int(rval))
							dataList.append(value)
						except KeyError:
							pass
					componentXML = etRoot.parse(componentPath)
					componentOutlineXML = componentXML.find("outline")
					componentHash, componentDataList = self.buildGlyphHashValue(width, componentOutlineXML, glyphName, useDefaultGlyphDir, level+1)
					dataList.extend(componentDataList)
		data = "".join(dataList)
		if len(data) < 128:
			hash = data
		else:
			hash = hashlib.sha512(data.encode("utf-8")).hexdigest()
		return hash, dataList

	def getComponentOutline(self, componentItem):
		try:
			compGlyphName = componentItem.attrib["base"]
		except KeyError:
			raise UFOParseError("'%s' attribute missing from component '%s'." % ("base", xmlToString(componentXML)))

		if not self.useProcessedLayer:
			try:
				compGlyphFilePath = self.getGlyphDefaultPath(compGlyphName)
			except KeyError:
				raise UFOParseError("'%s' component glyph is missing from contents.plist." % (compGlyphName))
		else:
			# If we are not necessarily using the default layer for the main glyph, then a missing component
			# may not have been processed, and may just be in the default layer. We need to look for component
			# glyphs in the src list first, then in the defualt layer.
			try:
				compGlyphFilePath = self.getGlyphSrcPath(compGlyphName)
				if not os.path.exists(compGlyphFilePath):
					compGlyphFilePath = self.getGlyphDefaultPath(compGlyphName)
			except KeyError:
				try:
					compGlyphFilePath = self.getGlyphDefaultPath(compGlyphName)
				except KeyError:
					raise UFOParseError("'%s' component glyph is missing from contents.plist." % (compGlyphName))

		if not os.path.exists(compGlyphFilePath):
			raise UFOParseError("'%s' component file is missing: '%s'." % (compGlyphName, compGlyphFilePath))

		etRoot = ET.ElementTree()
		glifXML = etRoot.parse(compGlyphFilePath)
		outlineXML = glifXML.find("outline")
		return outlineXML

	def copyTo(self, dstPath):
		""" Copy UFO font to target path"""
		return

	def close(self):
		if self.hashMapChanged:
			self.writeHashMap()
			self.hashMapChanged = False
		return

	def clearHashMap(self):
		self.hashMap = {kAdobHashMapVersionName:kAdobHashMapVersion}
		hashDir = os.path.join(self.parentPath, "data")
		if not os.path.exists(hashDir):
			return

		hashPath = os.path.join(hashDir, kAdobHashMapName)
		if os.path.exists(hashPath):
			os.remove(hashPath)

	def setWriteToDefault(self):
		self.useProcessedLayer = False
		self.writeToDefaultLayer = True
		self.glyphWriteDir = self.glyphDefaultDir

def parseGlyphOrder(filePath):
	orderMap = None
	if os.path.exists(filePath):
		publicOrderDict, temp = parsePList(filePath, kPublicGlyphOrderKey)
		if publicOrderDict != None:
			orderMap = {}
			glyphList = publicOrderDict[kPublicGlyphOrderKey]
			numGlyphs = len(glyphList)
			for i in range(numGlyphs):
				orderMap[glyphList[i]] = i
	return orderMap

def	parsePList(filePath, dictKey = None):
	# if dictKey is defined, parse and return only the data for that key.
	# This helps avoid needing to parse all plist types. I need to support only data for known keys amd lists.
	plistDict = {}
	plistKeys = []

	# I uses this rather than the plistlib in order to get a list that allows preserving key order.
	with open(filePath, "r") as fp:
		data = fp.read()
	contents = XML(data)
	dict = contents.find("dict")
	if dict == None:
		raise UFOParseError("In '%s', failed to find dict. '%s'." % (filePath))
	lastTag = "string"
	for child in dict:
		if child.tag == "key":
			if lastTag == "key":
				raise UFOParseError("In contents.plist, key name '%s' followed another key name '%s'." % (xmlToString(child), lastTag) )
			skipKeyData = False
			lastName = child.text
			lastTag = "key"
			if dictKey != None:
				if lastName != dictKey:
					skipKeyData = True
		elif child.tag != "key":
			if lastTag != "key":
				raise UFOParseError("In contents.plist, key value '%s' followed a non-key tag '%s'." % (xmlToString(child), lastTag) )
			lastTag = child.tag

			if skipKeyData:
				continue

			if lastName in plistDict:
				raise UFOParseError("Encountered duplicate key name '%s' in '%s'." % (lastName, filePath) )
			if child.tag == "array":
				list = []
				for listChild in child:
					val = listChild.text
					if listChild.tag == "integer":
						val= int(eval(val))
					elif listChild.tag == "real":
						val = float(eval(val))
					elif listChild.tag == "string":
						pass
					else:
						raise UFOParseError("In plist file, encountered unhandled key type '%s' in '%s' for parent key %s. %s." % (listChild.tag, child.tag, lastName, filePath))
					list.append(val)
				plistDict[lastName] = list
			elif child.tag == "integer":
				plistDict[lastName] = int(eval(child.text))
			elif child.tag == "real":
				plistDict[lastName] = float(eval(child.text))
			elif child.tag == "false":
				plistDict[lastName] = 0
			elif child.tag == "true":
				plistDict[lastName] = 1
			elif child.tag == "string":
				plistDict[lastName] = child.text
			else:
				raise UFOParseError("In plist file, encountered unhandled key type '%s' for key %s. %s." % (child.tag, lastName, filePath))
			plistKeys.append(lastName)
		else:
			raise UFOParseError("In plist file, encountered unexpected element '%s'. %s." % (xmlToString(child), filePath ))
	if len(plistDict) == 0:
		plistDict = None
	return plistDict, plistKeys




class UFOTransform:
	kTransformTagList = ["xScale", "xyScale", "yxScale", "yScale", "xOffset", "yOffset"]
	def __init__(self, componentXML):
		self.transformFactors = []
		self.isDefault = True
		self.isOffsetOnly = True
		hasScale = False
		hasOffset = False
		for tag in self.kTransformTagList:
			val = componentXML.attrib.get(tag, None)
			if tag in ["xScale", "yScale"]:
				if val == None:
					val = 1.0
				else:
					val = float(val)
					if val != 1.0:
						hasScale = True
			else:
				if val == None:
					val = 0
				else:
					val = float(val)
					if val != 0:
						self.isDefault = False
						if tag in ["xyScale", "yxScale"]:
							hasScale = True
						elif tag in ["xOffset", "yOffset"]:
							hasOffset = True
						else:
							raise UFOParseError("Unknown tag '%s' in component '%s'." % (tag, xmlToString(componentXML)))

			self.transformFactors.append(val)
		if hasScale or hasOffset:
			self.isDefault = False
		if hasScale:
			self.isOffsetOnly = False

	def concat(self, transform):
		if transform == None:
			return
		if transform.isDefault:
			return
		tfCur = self.transformFactors
		tfPrev = transform.transformFactors
		if transform.isOffsetOnly:
			tfCur[4] += tfPrev[4]
			tfCur[5] += tfPrev[5]
			self.isDefault = False
		else:
			tfNew = [0.0]*6
			tfNew[0] = tfCur[0]*tfPrev[0] + tfCur[1]*tfPrev[2];
			tfNew[1] = tfCur[0]*tfPrev[1] + tfCur[1]*tfPrev[3];
			tfNew[2] = tfCur[2]*tfPrev[0] + tfCur[3]*tfPrev[2];
			tfNew[3] = tfCur[2]*tfPrev[1] + tfCur[3]*tfPrev[3];
			tfNew[4] = tfCur[4]*tfPrev[0] + tfCur[5]*tfPrev[2] + tfPrev[4];
			tfNew[5] = tfCur[4]*tfPrev[1] + tfCur[5]*tfPrev[3] + tfPrev[5];
			self.transformFactors = tfNew
			self.isOffsetOnly = self.isOffsetOnly and transform.isOffsetOnly
			self.isDefault = False

	def apply(self, x, y):
		tfCur = self.transformFactors
		if self.isOffsetOnly:
			x += tfCur[4]
			y += tfCur[5]
		else:
			x, y = x*tfCur[0] + y*tfCur[2] + tfCur[4], \
				   x*tfCur[1] + y*tfCur[3] + tfCur[5]
		return x,y




def convertGlyphOutlineToBezString(outlineXML, ufoFontData, transform = None, level = 0):
	"""convert XML outline element containing contours and components to a bez string.
	Since xml.etree.CElementTree is compiled code, this
	will run faster than tokenizing and parsing in regular Python.

	glyphMap is a dict mapping glyph names to component file names.
	If it has a length of 1, it needs to be filled from the contents.plist file.
	It must have a key/value [UFO_FONTPATH] = path to parent UFO font.

	transformList is None, or a list of floats in the order:
	["xScale", "xyScale", "yxScale", "yScale", "xOffset", "yOffset"]

	Build a list of outlines and components.
	For each item:
		if is outline:
			get list of points
			for each point:
				if transform matrix:
					apply transform to coordinate
				if off line
					push coord on stack
				if online:
					add operator
			if is component:
				if any scale, skew, or offset factors:
					set transform
					call convertGlyphOutlineToBezString with transform
					add to bez string
	I don't bother adding in any hinting info, as this is used only for making
	temp bez files as input to checkOutlines or autohint, which invalidate hinting data
	by changing the outlines data, or at best ignore hinting data.

	bez ops output: ["rrmt", "dt", "rct", "cp" , "ed"]
	"""

	if level > 10:
		raise UFOParseError("In parsing component, exceeded 10 levels of reference. '%s'. " % (outlineXML))

	allowDecimals = ufoFontData.allowDecimalCoords

	bezStringList = []
	if outlineXML == None:
		bezstring = ""
	else:
		for outlineItem in outlineXML:

			if outlineItem.tag == "component":
				newTransform = UFOTransform(outlineItem)
				if newTransform.isDefault:
					if transform != None:
						newTransform.concat(transform)
					else:
						newTransform = None
				else:
					if transform != None:
						newTransform.concat(transform)
				componentOutline = ufoFontData.getComponentOutline(outlineItem)
				if componentOutline:
					outlineBezString = convertGlyphOutlineToBezString(componentOutline, ufoFontData, newTransform, level+1)
					bezStringList.append(outlineBezString)
				continue

			if outlineItem.tag != "contour":
				continue

			# May be an empty contour.
			if len(outlineItem) == 0:
				continue

			# We have a regular contour. Iterate over points.
			argStack = []
			# Deal with setting up move-to.
			lastItem = outlineItem[0]
			try:
				type = lastItem.attrib["type"]
			except KeyError:
				type = "offcurve"
			if type in ["curve","line","qccurve"]:
				outlineItem = outlineItem[1:]
				if type != "line":
					outlineItem.append(lastItem) # I don't do this for line, as AC behaves differently when a final line-to is explicit.
				x = float(lastItem.attrib["x"])
				y = float(lastItem.attrib["y"])
				if transform != None:
					x,y = transform.apply(x,y)

				if (not allowDecimals):
					x = int(round(x))
					y = int(round(y))

				if (allowDecimals):
					op = "%.3f %.3f mt" % (x , y)
				else:
					op = "%s %s mt" % (x, y)
				bezStringList.append(op)
			elif type == "move":
				# first op is a move-to.
				if len(outlineItem) == 1:
					# this is a move, and is the only op in this outline. Don't pass it through.
					# this is most likely left over from a GLIF format 1 anchor.
					argStack = []
					continue
			elif type == "offcurve":
				# We should only see an off curve point as the first point when the first op is a curve and the last op is a line.
				# In this case, insert a move-to to the line's coords, then omit the line.
				# Breaking news! In rare cases, a first off-curve point can occur when the first AND last op is a curve.
				curLastItem = outlineItem[-1]
				# In this case, insert a move-to to the last op's end pos.
				try:
					lastType = curLastItem.attrib["type"]
					x = float(curLastItem.attrib["x"])
					y = float(curLastItem.attrib["y"])
					if transform != None:
						x,y = transform.apply(x,y)

					if (not allowDecimals):
						x = int(round(x))
						y = int(round(y))

					if lastType == "line":

						if (allowDecimals):
							op = "%.3f %.3f mt" % (x , y)
						else:
							op = "%s %s mt" % (x, y)

						bezStringList.append(op)
						outlineItem = outlineItem[:-1]
					elif lastType == "curve":

						if (allowDecimals):
							op = "%.3f %.3f mt" % (x , y)
						else:
							op = "%s %s mt" % (x, y)
						bezStringList.append(op)

				except KeyError:
					raise UFOParseError("Unhandled case for first and last points in outline '%s'." % (xmlToString(outlineItem)))
			else:
				raise UFOParseError("Unhandled case for first point in outline '%s'." % (xmlToString(outlineItem)))

			for contourItem in outlineItem:
				if contourItem.tag != "point":
					continue
				x = float(contourItem.attrib["x"])
				y = float(contourItem.attrib["y"])
				if transform != None:
					x,y = transform.apply(x,y)

				if (not allowDecimals):
					x = int(round(x))
					y = int(round(y))

				try:
					type = contourItem.attrib["type"]
				except KeyError:
					type = "offcurve"

				if type == "offcurve":
					argStack.append(x)
					argStack.append(y)
				elif type == "move":
					if (allowDecimals):
						op = "%.3f %.3f mt" % (x , y)
					else:
						op = "%s %s mt" % (x, y)
					bezStringList.append(op)
					argStack = []
				elif type == "line":
					if (allowDecimals):
						op = "%.3f %.3f dt" % (x, y)
					else:
						op = "%s %s dt" % (x, y)
					bezStringList.append(op)
					argStack = []
				elif type == "curve":
					if len(argStack) != 4:
						raise UFOParseError("Argument stack error seen for curve point '%s'." % (xmlToString(contourItem)))
					if (allowDecimals):
						op = "%.3f %.3f %.3f %.3f %.3f %.3f ct" % ( argStack[0], argStack[1], argStack[2], argStack[3], x, y)
					else:
						op = "%s %s %s %s %s %s ct" % ( argStack[0], argStack[1], argStack[2], argStack[3], x, y)
					argStack = []
					bezStringList.append(op)
				elif type == "qccurve":
					raise UFOParseError("Point type not supported '%s'." % (xmlToString(contourItem)))
				else:
					raise UFOParseError("Unknown Point type not supported '%s'." % (xmlToString(contourItem)))

			# we get here only if there was at least a move.
			bezStringList.append("cp")
		bezstring = "\n".join(bezStringList)
	return bezstring


def convertGLIFToBez(ufoFontData, glyphName, beVerbose, doAll=False):
	width, outlineXML, skip = ufoFontData.getOrSkipGlyphXML(glyphName, doAll)
	if skip:
		return None, width

	if outlineXML == None:
		return None, width

	bezString = convertGlyphOutlineToBezString(outlineXML, ufoFontData)
	bezString = "\n".join(["% " + glyphName, "sc", bezString, "ed", ""])
	return bezString, width


class HintMask:
	# class used to collect hints for the current hint mask when converting bez to T2.
	def __init__(self, listPos):
		self.listPos = listPos # The index into the pointList is kept so we can quickly find them later.
		self.hList = [] # These contain the actual hint values.
		self.vList = []
		self.hstem3List = []
		self.vstem3List = []
		self.pointName = "hintSet" + str(listPos).zfill(4)	# The name attribute of the point which follows the new hint set.

	def addHintSet(self, hintSetList):
		# Add the hint set to hintSetList
		newHintSetDict = XMLElement("dict")
		hintSetList.append(newHintSetDict)

		newHintSetKey = XMLElement("key")
		newHintSetKey.text = kPointTag
		newHintSetDict.append(newHintSetKey)

		newHintSetValue = XMLElement("string")
		newHintSetValue.text = self.pointName
		newHintSetDict.append(newHintSetValue)

		stemKey = XMLElement("key")
		stemKey.text = "stems"
		newHintSetDict.append(stemKey)

		newHintSetArray = XMLElement("array")
		newHintSetDict.append(newHintSetArray)

		newHintSet = []
		if (len(self.hList) > 0) or (len(self.vstem3List)):
			isH = True
			addHintList(self.hList, self.hstem3List, newHintSetArray, isH)

		if (len(self.vList) > 0) or (len(self.vstem3List)):
			isH = False
			addHintList(self.vList, self.vstem3List, newHintSetArray, isH)


def makeStemHintList(hintsStem3, stemList, isH):
	stem3Args = []
	lastPos = 0
	# In bez terms, the first coordinate in each pair is absolute, second is relative, and hence is the width.
	if isH:
		op = kHStem3Name
	else:
		op = kVStem3Name
	newStem = XMLElement("string")
	posList = [op]
	for stem3 in hintsStem3:
		for pos, width in stem3:
			if (type(pos) == type(0.0)) and (int(pos) == pos):
				pos = int(pos)
			if (type(width) == type(0.0)) and (int(width) == width):
				width = int(width)
			posList.append("%s %s" % (pos, width))
	posString = " ".join(posList)
	newStem.text = posString
	stemList.append(newStem)

def makeHintList(hints, newHintSetArray, isH):
	# Add the list of hint operators
	hintList = []
	lastPos = 0
	# In bez terms, the first coordinate in each pair is absolute, second is relative, and hence is the width.
	for hint in hints:
		if not hint:
			continue
		pos = hint[0]
		if (type(pos) == type(0.0)) and (int(pos) == pos):
			pos = int(pos)
		width = hint[1]
		if (type(width) == type(0.0)) and (int(width) == width):
			width = int(width)
		if isH:
			op = kHStemName
		else:
			op = kVStemName
		newStem = XMLElement("string")
		newStem.text = "%s %s %s" % (op, pos, width)
		newHintSetArray.append(newStem)

def addFlexHint(flexList, flexArray):
	for pointTag in flexList:
		newFlexTag = XMLElement("string")
		newFlexTag.text = pointTag
		flexArray.append(newFlexTag)

def fixStartPoint(outlineItem, opList):
	# For the GLIF format, the idea of first/last point is funky, because the format avoids identifying a
	# a start point. This means there is no implied close-path line-to. If the last implied or explicit path-close
	# operator is a line-to, then replace the "mt" with linto, and remove the last explicit path-closing line-to, if any.
	# If the last op is a curve, then leave the first two point args on the stack at the end of the point list,
	# and move the last curveto to the first op, replacing the move-to.

	firstOp, firstX, firstY = opList[0]
	lastOp, lastX, lastY = opList[-1]
	firstPointElement = outlineItem[0]
	if (firstX == lastX) and (firstY == lastY):
		del outlineItem[-1]
		firstPointElement.set("type", lastOp)
	else:
		# we have an implied final line to. All we need to do is convert the inital moveto to a lineto.
		firstPointElement.set("type", "line")


bezToUFOPoint = {
		"mt" : 'move',
		"rmt" : 'move',
		 "hmt" : 'move',
		"vmt" : 'move',
		"rdt" : 'line',
		"dt" : 'line',
		"hdt" : "line",
		"vdt" : "line",
		"rct" : 'curve',
		"ct" : 'curve',
		"rcv" : 'curve', # Morisawa's alternate name for 'rct'.
		"vhct": 'curve',
		"hvct": 'curve',
		}

def convertCoords(curX, curY):
	showX = int(curX)
	if showX != curX:
		showX = curX
	showY = int(curY)
	if showY != curY:
		showY = curY
	return showX, showY

def convertBezToOutline(ufoFontData, glyphName, bezString):
	""" Since the UFO outline element as no attributes to preserve, I can
	just make a new one.
	"""
	# convert bez data to a UFO glif XML representation
	#
	# Convert all bez ops to simplest UFO equivalent
	# Add all hints to vertical and horizontal hint lists as encountered; insert a HintMask class whenever a
	# new set of hints is encountered
	# after all operators have been processed, convert HintMask items into hintmask ops and hintmask bytes
	# add all hints as prefix
	# review operator list to optimize T2 operators.
	# if useStem3 == 1, then any counter hints must be processed as stem3 hints, else the opposite.
	# counter hints are used only in LanguageGroup 1 glyphs, aka ideographs

	bezString = re.sub(r"%.+?\n", "", bezString) # supress comments
	bezList = re.findall(r"(\S+)", bezString)
	if not bezList:
		return "", None, None
	flexList = []
	hintMask = HintMask(0) # Create an initial hint mask. We use this if there is no explicit initial hint sub.
	hintMaskList = [hintMask]
	vStem3Args = []
	hStem3Args = []
	vStem3List = []
	hStem3List = []
	argList = []
	opList = []
	newHintMaskName = None
	inPreFlex = False
	hintInfoDict = None
	opIndex = 0
	lastPathOp = None
	curX = 0
	curY = 0
	newOutline = XMLElement("outline")
	outlineItem = None
	seenHints = False

	for token in bezList:
		try:
			val = float(token)
			argList.append(val)
			continue
		except ValueError:
			pass
		if token == "newcolors":
			lastPathOp = token
			pass
		elif token in ["beginsubr", "endsubr"]:
			lastPathOp = token
			pass
		elif token in ["snc"]:
			lastPathOp = token
			hintMask = HintMask(opIndex)
			if opIndex == 0: # If the new colors precedes any marking operator, then we want throw away the initial hint mask we made, and use the new one as the first hint mask.
				hintMaskList = [hintMask]
			else:
				hintMaskList.append(hintMask)
			newHintMaskName = hintMask.pointName
		elif token in ["enc"]:
			lastPathOp = token
			pass
		elif token == "div":
			# I specifically do NOT set lastPathOp for this.
			value = argList[-2]/float(argList[-1])
			argList[-2:] =[value]
		elif token == "rb":
			if newHintMaskName == None:
				newHintMaskName = hintMask.pointName
			lastPathOp = token
			hintMask.hList.append(argList)
			argList = []
			seenHints = True
		elif token == "ry":
			if newHintMaskName == None:
				newHintMaskName = hintMask.pointName
			lastPathOp = token
			hintMask.vList.append(argList)
			argList = []
			seenHints = True
		elif token == "rm": # vstem3's are vhints
			if newHintMaskName == None:
				newHintMaskName = hintMask.pointName
			seenHints = True
			vStem3Args.append(argList)
			argList = []
			lastPathOp = token
			if len(vStem3Args) == 3:
				hintMask.vstem3List.append(vStem3Args)
				vStem3Args = []

		elif token == "rv": # hstem3's are hhints
			seenHints = True
			hStem3Args.append(argList)
			argList = []
			lastPathOp = token
			if len(hStem3Args) == 3:
				hintMask.hstem3List.append(hStem3Args)
				hStem3Args = []

		elif token == "preflx1":
			# the preflx1/preflx2 sequence provides the same i as the flex sequence; the difference is that the
			# preflx1/preflx2 sequence provides the argument values needed for building a Type1 string
			# while the flex sequence is simply the 6 rcurveto points. Both sequences are always provided.
			lastPathOp = token
			argList = []
			inPreFlex = True # need to skip all move-tos until we see the "flex" operator.
		elif token == "preflx2a":
			lastPathOp = token
			argList = []
		elif token == "preflx2":
			lastPathOp = token
			argList = []
		elif token == "flxa": # flex with absolute coords.
			inPreFlex = False
			flexPointName = kBaseFlexName + str(opIndex).zfill(4)
			flexList.append(flexPointName)
			curveCnt = 2
			i = 0
			# The first 12 args are the 6 args for each of the two curves that make up the flex feature.
			while i < curveCnt:
				curX = argList[0]
				curY = argList[1]
				showX, showY = convertCoords(curX, curY)
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
				outlineItem.append(newPoint)
				curX = argList[2]
				curY = argList[3]
				showX, showY = convertCoords(curX, curY)
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
				outlineItem.append(newPoint)
				curX = argList[4]
				curY = argList[5]
				showX, showY = convertCoords(curX, curY)
				opName = 'curve'
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": opName } )
				outlineItem.append(newPoint)
				opList.append([opName,curX, curY])
				opIndex += 1
				if i == 0:
					argList = argList[6:12]
				i += 1
			outlineItem[-6].set(kPointName, flexPointName) # attach the point name to the first point of the first curve.
			if (newHintMaskName != None):
				# We have a hint mask that we want to attach to the first point of the flex op. However, there is already
				# a flex name in that attribute. What we do is set the flex point name into the hint mask.
				hintMask.pointName = flexPointName
				newHintMaskName = None
			lastPathOp = token
			argList = []
		elif token == "flx":
			inPreFlex = False
			flexPointName = kBaseFlexName + str(opIndex).zfill(4)
			flexList.append(flexPointName)
			curveCnt = 2
			i = 0
			# The first 12 args are the 6 args for each of the two curves that make up the flex feature.
			while i < curveCnt:
				curX += argList[0]
				curY += argList[1]
				showX, showY = convertCoords(curX, curY)
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
				outlineItem.append(newPoint)
				curX += argList[2]
				curY += argList[3]
				showX, showY = convertCoords(curX, curY)
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
				outlineItem.append(newPoint)
				curX += argList[4]
				curY += argList[5]
				showX, showY = convertCoords(curX, curY)
				opName = 'curve'
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": opName } )
				outlineItem.append(newPoint)
				opList.append([opName,curX, curY])
				opIndex += 1
				if i == 0:
					argList = argList[6:12]
				i += 1
			outlineItem[-6].set(kPointName, flexPointName) # attach the point name to the first point of the first curve.
			if (newHintMaskName != None):
				# We have a hint mask that we want to attach to the first point of the flex op. However, there is already
				# a flex name in that attribute. What we do is set the flex point name into the hint mask.
				hintMask.pointName = flexPointName
				newHintMaskName = None
			lastPathOp = token
			argList = []
		elif token == "sc":
			lastPathOp = token
			pass
		elif token == "cp":
			pass
		elif token == "ed":
		 	pass
		else:
			if inPreFlex and (token[-2:] == "mt"):
				continue

			if token[-2:] in ["mt", "dt", "ct", "cv"]:
				lastPathOp = token
				opIndex += 1
			else:
				print("Unhandled operation", argList, token)
				raise BezParseError("Unhandled operation: '%s' '%s'.", argList, token)
			dx = dy = 0
			opName = bezToUFOPoint[token]
			if token[-2:] in ["mt", "dt"]:
				if token in ["mt", "dt"]:
					curX = argList[0]
					curY = argList[1]
				else:
					if token in ["rmt", "rdt"]:
						dx = argList[0]
						dy = argList[1]
					elif token in ["hmt", "hdt"]:
						dx = argList[0]
					elif token in ["vmt", "vdt"]:
						dy = argList[0]
					curX += dx
					curY += dy
				showX, showY = convertCoords(curX, curY)
				newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": "%s" % (opName) } )

				if opName == "move":
					if outlineItem != None:
						if len(outlineItem) == 1:
							# Just in case we see two moves in a row, delete the previous outlineItem if it has only the move-to''
							print("Deleting moveto:", xmlToString(newOutline[-1]), "adding", xmlToString(outlineItem))
							del newOutline[-1]
						else:
							fixStartPoint(outlineItem, opList) # Fix the start/implied end path of the previous path.
					opList = []
					outlineItem = XMLElement('contour')
					newOutline.append(outlineItem)

				if (newHintMaskName != None):
					newPoint.set(kPointName, newHintMaskName)
					newHintMaskName = None
				outlineItem.append(newPoint)
				opList.append([opName,curX, curY])
			else:
				if token in ["ct", "cv"]:
					curX = argList[0]
					curY = argList[1]
					showX, showY = convertCoords(curX, curY)
					newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
					outlineItem.append(newPoint)
					curX = argList[2]
					curY = argList[3]
					showX, showY = convertCoords(curX, curY)
					newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
					outlineItem.append(newPoint)
					curX = argList[4]
					curY = argList[5]
					showX, showY = convertCoords(curX, curY)
					newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": "%s" % (opName) } )
					outlineItem.append(newPoint)
				else:
					if token in ["rct", "rcv"]:
						curX += argList[0]
						curY += argList[1]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curX += argList[2]
						curY += argList[3]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curX += argList[4]
						curY += argList[5]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": "%s" % (opName) } )
						outlineItem.append(newPoint)
					elif token == "vhct":
						curY += argList[0]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curX += argList[1]
						curY += argList[2]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curX += argList[3]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": "%s" % (opName) } )
						outlineItem.append(newPoint)
					elif token == "hvct":
						curX += argList[0]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curX += argList[1]
						curY += argList[2]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY) } )
						outlineItem.append(newPoint)
						curY += argList[3]
						showX, showY = convertCoords(curX, curY)
						newPoint = XMLElement("point", {"x": "%s" % (showX), "y": "%s" % (showY), "type": "%s" % (opName) } )
						outlineItem.append(newPoint)
				if (newHintMaskName != None):
					outlineItem[-3].set(kPointName, newHintMaskName) # attach the pointName to the first point of the curve.
					newHintMaskName = None
				opList.append([opName,curX, curY])
			argList = []
	if outlineItem != None:
		if len(outlineItem) == 1:
			# Just in case we see two moves in a row, delete the previous outlineItem if it has zero length.
			del newOutline[-1]
		else:
			fixStartPoint(outlineItem, opList)

	# add hints, if any
	# Must be done at the end of op processing to make sure we have seen all the
	# hints in the bez string.
	# Note that the hintmasks are identified in the opList by the point name.
	# We will follow the T1 spec: a glyph may have stem3 counter hints or regular hints, but not both.

	hintSetList = None
	if (seenHints) or (len(flexList) > 0):
		hintInfoDict = XMLElement("dict")

		idItem = XMLElement("key")
		idItem.text = "id"
		hintInfoDict.append(idItem)

		idString = XMLElement("string")
		idString.text = "id"
		hintInfoDict.append(idString)

		hintSetListItem = XMLElement("key")
		hintSetListItem.text = kHintSetListName
		hintInfoDict.append(hintSetListItem)

		hintSetListArray = XMLElement("array")
		hintInfoDict.append(hintSetListArray)
		# Convert the rest of the hint masks to a hintmask op and hintmask bytes.
		for hintMask in hintMaskList:
			hintMask.addHintSet(hintSetListArray)

		if len(flexList) > 0:
			hintSetListItem = XMLElement("key")
			hintSetListItem.text = kFlexIndexListName
			hintInfoDict.append(hintSetListItem)

			flexArray = XMLElement("array")
			hintInfoDict.append(flexArray)
			addFlexHint(flexList, flexArray)
	return newOutline, hintInfoDict

def addHintList(hints, hintsStem3, newHintSetArray, isH):
	# A charstring may have regular v stem hints or vstem3 hints, but not both.
	# Same for h stem hints and hstem3 hints.
	if len(hintsStem3) > 0:
		hintsStem3.sort()
		numHints = len(hintsStem3)
		hintLimit = (kStackLimit-2)/2
		if numHints >=hintLimit:
			hintsStem3 = hintsStem3[:hintLimit]
			numHints = hintLimit
		makeStemHintList(hintsStem3, newHintSetArray, isH)

	else:
		hints.sort()
		numHints = len(hints)
		hintLimit = (kStackLimit-2)/2
		if numHints >=hintLimit:
			hints = hints[:hintLimit]
			numHints = hintLimit
		makeHintList(hints, newHintSetArray, isH)


def addWhiteSpace(parent, level):
	child = None
	childIndent = "\n" + ("  "*(level +1))
	prentIndent = "\n" + ("  "*(level))
	#print("parent Tag", parent.tag, repr(parent.text), repr(parent.tail))
	for child in parent:
		child.tail = childIndent
		addWhiteSpace(child, level +1)
	if child != None:
		if parent.text == None:
			parent.text = childIndent
		child.tail = prentIndent
		#print("lastChild Tag", child.tag, repr(child.text), repr(child.tail), "parent Tag", parent.tag)


def convertBezToGLIF(ufoFontData, glyphName, bezString, hintsOnly = False):
	# I need to replace the contours with data from the bez string.
	glyphPath = ufoFontData.getGlyphSrcPath(glyphName)

	with open(glyphPath, "r") as fp:
		data = fp.read()

	glifXML = XML(data)

	outlineItem = None
	libIndex = outlineIndex = -1
	outlineIndex = outlineIndex = -1
	childIndex = 0
	for childElement in glifXML:
		if childElement.tag == "outline":
			outlineItem = childElement
			outlineIndex = childIndex
		if childElement.tag == "lib":
			libIndex = childIndex
		childIndex += 1


	newOutlineElement, hintInfoDict = convertBezToOutline(ufoFontData, glyphName, bezString )
	#print(xmlToString(stemHints))

	if not hintsOnly:
		if outlineItem == None:
			# need to add it. Add it before the lib item, if any.
			if libIndex > 0:
				glifXML.insert(libIndex, newOutlineElement)
			else:
				glifXML.append(newOutlineElement)
		else:
			# remove the old one and add the new one.
			glifXML.remove(outlineItem)
			glifXML.insert(outlineIndex, newOutlineElement)

	# convertBezToGLIF is called only if the GLIF has been edited by a tool. We need to update the edit status
	# in the has map entry.
	# I assume that convertGLIFToBez has ben run before, which will add an entry for this glyph.
	ufoFontData.updateHashEntry(glyphName, changed=True)
	# Add the stem hints.
	if (hintInfoDict != None):
		widthXML = glifXML.find("advance")
		if widthXML != None:
			width = int(eval(widthXML.get("width")))
		else:
			width = 1000
		useDefaultGlyphDir = False
		newGlyphHash, dataList = ufoFontData.buildGlyphHashValue(width, newOutlineElement, glyphName, useDefaultGlyphDir)
		# We add this hash to the T1 data, as it is the hash which matches the output outline data.
		# This is not necessarily the same as the the hash of the source data - autohint can be used to change outlines.
		if libIndex > 0:
			libItem = glifXML[libIndex]
		else:
			libItem = XMLElement("lib")
			glifXML.append(libItem)

		dictItem = libItem.find("dict")
		if dictItem == None:
			dictItem = XMLElement("dict")
			libItem.append(dictItem)

		# Remove any existing hint data.
		i = 0
		childList = list(dictItem)
		for childItem in childList:
			i += 1
			if (childItem.tag == "key") and ((childItem.text == kHintDomainName1) or (childItem.text == kHintDomainName2)):
				dictItem.remove(childItem) # remove key
				dictItem.remove(childList[i]) # remove data item.

		glyphDictItem = dictItem
		key = XMLElement("key")
		key.text = kHintDomainName2
		glyphDictItem.append(key)

		glyphDictItem.append(hintInfoDict)

		childList = list(hintInfoDict)
		idValue = childList[1]
		idValue.text = newGlyphHash

	addWhiteSpace(glifXML, 0)
	return glifXML