appypod-rattail/appy/pod/buffers.py

# ------------------------------------------------------------------------------
# This file is part of Appy, a framework for building applications in the Python
# language. Copyright (C) 2007 Gaetan Delannay

# Appy is free software; you can redistribute it and/or modify it under the
# terms of the GNU General Public License as published by the Free Software
# Foundation; either version 3 of the License, or (at your option) any later
# version.

# Appy is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
# A PARTICULAR PURPOSE. See the GNU General Public License for more details.

# You should have received a copy of the GNU General Public License along with
# Appy. If not, see <http://www.gnu.org/licenses/>.

# ------------------------------------------------------------------------------
import re
from xml.sax.saxutils import quoteattr
from appy.shared.xml_parser import xmlPrologue, escapeXml
from appy.pod import PodError
from appy.shared.utils import Traceback
from appy.pod.elements import *
from appy.pod.actions import IfAction, ElseAction, ForAction, VariablesAction, \
                             NullAction

# ------------------------------------------------------------------------------
class ParsingError(Exception): pass
class EvaluationError(Exception): pass

# ParsingError-related constants -----------------------------------------------
ELEMENT = 'identifies the part of the document that will be impacted ' \
          'by the command. It must be one of %s.' % str(PodElement.POD_ELEMS)
FOR_EXPRESSION = 'must be of the form: {name} in {expression}. {name} must be '\
                 'a Python variable name. It is the name of the iteration ' \
                 'variable. {expression} is a Python expression that, when ' \
                 'evaluated, produces a Python sequence (tuple, string, list, '\
                 'etc).'
POD_STATEMENT = 'A Pod statement has the ' \
                'form: do {element} [{command} {expression}]. {element} ' + \
                ELEMENT + ' Optional {command} can be "if" ' \
                '(conditional inclusion of the element) or "for" (multiple ' \
                'inclusion of the element). For an "if" command, {expression} '\
                'is any Python expression. For a "for" command, {expression} '+\
                FOR_EXPRESSION
FROM_CLAUSE = 'A "from" clause has the form: from {expression}, where ' \
              '{expression} is a Python expression that, when evaluated, ' \
              'produces a valid chunk of odt content that will be inserted ' \
              'instead of the element that is the target of the note.'
BAD_STATEMENT_GROUP = 'Syntax error while parsing a note whose content is ' \
                      '"%s". In a note, you may specify at most 2 lines: a ' \
                      'pod statement and a "from" clause. ' + POD_STATEMENT + \
                      ' ' + FROM_CLAUSE
BAD_STATEMENT = 'Syntax error for statement "%s". ' + POD_STATEMENT
BAD_ELEMENT = 'Bad element "%s". An element ' + ELEMENT
BAD_MINUS = "The '-' operator can't be used with element '%s'. It can only be "\
            "specified for elements among %s."
ELEMENT_NOT_FOUND = 'Action specified element "%s" but available elements ' \
                    'in this part of the document are %s.'
BAD_FROM_CLAUSE = 'Syntax error in "from" clause "%s". ' + FROM_CLAUSE
DUPLICATE_NAMED_IF = 'An "if" statement with the same name already exists.'
ELSE_WITHOUT_IF = 'No previous "if" statement could be found for this "else" ' \
                  'statement.'
ELSE_WITHOUT_NAMED_IF = 'I could not find an "if" statement named "%s".'
BAD_FOR_EXPRESSION = 'Bad "for" expression "%s". A "for" expression ' + \
                     FOR_EXPRESSION
BAD_VAR_EXPRESSION = 'Bad variable definition "%s". A variable definition ' \
    'must have the form {name} = {expression}. {name} must be a Python-' \
    'compliant variable name. {expression} is a Python expression. When ' \
    'encountering such a statement, pod will define, in the specified part ' \
    'of the document, a variable {name} whose value will be the evaluated ' \
    '{expression}.'
EVAL_EXPR_ERROR = 'Error while evaluating expression "%s". %s'
NULL_ACTION_ERROR = 'There was a problem with this action. Possible causes: ' \
                    '(1) you specified no action (ie "do text") while not ' \
                    'specifying any from clause; (2) you specified the from ' \
                    'clause on the same line as the action, which is not ' \
                    'allowed (ie "do text from ...").'
# ------------------------------------------------------------------------------
class BufferIterator:
    def __init__(self, buffer):
        self.buffer = buffer
        self.remainingSubBufferIndexes = list(self.buffer.subBuffers.keys())
        self.remainingElemIndexes = list(self.buffer.elements.keys())
        self.remainingSubBufferIndexes.sort()
        self.remainingElemIndexes.sort()

    def hasNext(self):
        return self.remainingSubBufferIndexes or self.remainingElemIndexes

    def __next__(self):
        nextSubBufferIndex = None
        if self.remainingSubBufferIndexes:
            nextSubBufferIndex = self.remainingSubBufferIndexes[0]
        nextExprIndex = None
        if self.remainingElemIndexes:
            nextExprIndex = self.remainingElemIndexes[0]
        # Compute min between nextSubBufferIndex and nextExprIndex
        if (nextSubBufferIndex != None) and (nextExprIndex != None):
            res = min(nextSubBufferIndex, nextExprIndex)
        elif (nextSubBufferIndex == None) and (nextExprIndex != None):
            res = nextExprIndex
        elif (nextSubBufferIndex != None) and (nextExprIndex == None):
            res = nextSubBufferIndex
        # Update "remaining" lists
        if res == nextSubBufferIndex:
            self.remainingSubBufferIndexes = self.remainingSubBufferIndexes[1:]
            resDict = self.buffer.subBuffers
        elif res == nextExprIndex:
            self.remainingElemIndexes = self.remainingElemIndexes[1:]
            resDict = self.buffer.elements
        return res, resDict[res]

# ------------------------------------------------------------------------------
class Buffer:
    '''Abstract class representing any buffer used during rendering.'''
    elementRex = re.compile('([\w-]+:[\w-]+)\s*(.*?)>', re.S)

    def __init__(self, env, parent):
        self.parent = parent
        self.subBuffers = {} # ~{i_bufferIndex: Buffer}~
        self.env = env
        # Are we computing for pod (True) or px (False)
        self.pod = env.__class__.__name__ != 'PxEnvironment'

    def addSubBuffer(self, subBuffer=None):
        if not subBuffer:
            subBuffer = MemoryBuffer(self.env, self)
        self.subBuffers[self.getLength()] = subBuffer
        subBuffer.parent = self
        return subBuffer

    def removeLastSubBuffer(self):
        subBufferIndexes = list(self.subBuffers.keys())
        subBufferIndexes.sort()
        lastIndex = subBufferIndexes.pop()
        del self.subBuffers[lastIndex]

    def write(self, something): pass # To be overridden

    def getLength(self): pass # To be overridden

    def patchTableElement(self, elem, attrs):
        '''Convert the name of a table to an expression allowing the user to
           define himself this name via variable "tableName".

           Convert attribute "number-columns-repeated" of every table column
           (or add it if it does not exist) to let the user define how he will
           repeat table columns via variable "columnsRepeated".'''
        if elem == self.env.tags['table']:
            attrs = attrs._attrs
            name = self.env.tags['table-name']
            attrs[name] = ':tableName|"%s"' % attrs[name]
        elif elem == self.env.tags['table-column']:
            attrs = attrs._attrs
            key = self.env.tags['number-columns-repeated']
            columnNumber = self.env.getTable().nbOfColumns -1
            nb = (key in attrs) and attrs[key] or '1'
            attrs[key] = ':columnsRepeated[%d]|%s' % (columnNumber, nb)

    def dumpStartElement(self, elem, attrs={}, ignoreAttrs=(), hook=False,
                         noEndTag=False, renamedAttrs=None):
        '''Inserts into this buffer the start tag p_elem, with its p_attrs,
           excepted those listed in p_ignoreAttrs. Attrs can be dumped with an
           alternative name if specified in dict p_renamedAttrs. If p_hook is
           not None (works only for MemoryBuffers), we will insert, at the end
           of the list of dumped attributes:
           * [pod] an Attributes instance, in order to be able, when evaluating
                   the buffer, to dump additional attributes, not known at this
                   dump time;
           * [px]  an Attribute instance, representing a special HTML attribute
                   like "checked" or "selected", that, if the tied expression
                   returns False, must not be dumped at all. In this case,
                   p_hook must be a tuple (s_attrName, s_expr).
        '''
        self.write('<%s' % elem)
        # Some table elements must be patched (pod only)
        if self.pod: self.patchTableElement(elem, attrs)
        for name, value in list(attrs.items()):
            if ignoreAttrs and (name in ignoreAttrs): continue
            if renamedAttrs and (name in renamedAttrs): name=renamedAttrs[name]
            # If the value begins with ':', it is a Python expression. Else,
            # it is a static value.
            if not value.startswith(':'):
                self.write(' %s=%s' % (name, quoteattr(value)))
            else:
                self.write(' %s="' % name)
                self.addExpression(value[1:])
                self.write('"')
        res = None
        if hook:
            if self.pod:
                res = self.addAttributes()
            else:
                self.addAttribute(*hook)
        # Close the tag
        self.write(noEndTag and '/>' or '>')
        return res

    def dumpEndElement(self, elem):
        self.write('</%s>' % elem)

    def dumpElement(self, elem, content=None, attrs={}):
        '''For dumping a whole element at once.'''
        self.dumpStartElement(elem, attrs)
        if content:
            self.dumpContent(content)
        self.dumpEndElement(elem)

    def dumpContent(self, content):
        '''Dumps string p_content into the buffer.'''
        if self.pod:
            # Take care of converting line breaks and tabs
            content = escapeXml(content, format='odf',
                                nsText=self.env.namespaces[self.env.NS_TEXT])
        else:
            content = escapeXml(content)
        self.write(content)

# ------------------------------------------------------------------------------
class FileBuffer(Buffer):
    def __init__(self, env, result):
        Buffer.__init__(self, env, None)
        self.result = result
        self.content = file(result, 'w')
        self.content.write(xmlPrologue)

    # getLength is used to manage insertions into sub-buffers. But in the case
    # of a FileBuffer, we will only have 1 sub-buffer at a time, and we don't
    # care about where it will be inserted into the FileBuffer.
    def getLength(self): return 0

    def write(self, something):
        try:
            self.content.write(something.encode('utf-8'))
        except UnicodeDecodeError:
            self.content.write(something)

    def addExpression(self, expression, tiedHook=None):
        # At 2013-02-06, this method was not called within the whole test suite.
        try:
            expr = Expression(expression, self.pod)
            if tiedHook: tiedHook.tiedExpression = expr
            res, escape = expr.evaluate(self.env.context)
            if escape: self.dumpContent(res)
            else: self.write(res)
        except Exception as e:
            if not self.env.raiseOnError:
                PodError.dump(self, EVAL_EXPR_ERROR % (expression, e),
                              dumpTb=False)
            else:
                raise Exception(EVAL_EXPR_ERROR % (expression, e))

    def addAttributes(self):
        # Into a FileBuffer, it is not possible to insert Attributes. Every
        # Attributes instance is tied to an Expression; because dumping
        # expressions directly into FileBuffer instances seems to be a rather
        # theorical case (see comment inside the previous method), it does not
        # seem to be a real problem.
        pass

    def pushSubBuffer(self, subBuffer): pass
    def getRootBuffer(self): return self

# ------------------------------------------------------------------------------
class MemoryBuffer(Buffer):
    actionRex = re.compile('(?:(\w+)\s*\:\s*)?do\s+(\w+)(-)?' \
                           '(?:\s+(for|if|else|with)\s*(.*))?')
    forRex = re.compile('\s*([\w\-_]+)\s+in\s+(.*)')
    varRex = re.compile('\s*(@?[\w\-_]+)\s*=\s*(.*)')

    def __init__(self, env, parent):
        Buffer.__init__(self, env, parent)
        self.content = ''
        self.elements = {}
        self.action = None

    def clone(self):
        '''Produces an empty buffer that is a clone of this one.'''
        return MemoryBuffer(self.env, self.parent)

    def addSubBuffer(self, subBuffer=None):
        sb = Buffer.addSubBuffer(self, subBuffer)
        self.content += ' ' # To avoid having several subbuffers referenced at
                            # the same place within this buffer.
        return sb

    def getRootBuffer(self):
        '''Returns the root buffer. For POD it is always a FileBuffer. For PX,
           it is a MemoryBuffer.'''
        if self.parent: return self.parent.getRootBuffer()
        return self

    def getLength(self): return len(self.content)

    def write(self, thing): self.content += thing

    def getIndex(self, podElemName):
        res = -1
        for index, podElem in self.elements.items():
            if podElem.__class__.__name__.lower() == podElemName:
                if index > res:
                    res = index
        return res

    def getMainElement(self):
        res = None
        if 0 in self.elements:
            res = self.elements[0]
        return res

    def isMainElement(self, elem):
        '''Is p_elem the main element within this buffer?'''
        mainElem = self.getMainElement()
        if not mainElem: return
        if hasattr(mainElem, 'OD'): mainElem = mainElem.OD.elem
        if elem != mainElem: return
        # elem is the same as the main elem. But is it really the main elem, or
        # the same elem, found deeper in the buffer?
        for index, iElem in self.elements.items():
            foundElem = None
            if hasattr(iElem, 'OD'):
                if iElem.OD:
                    foundElem = iElem.OD.elem
            else:
                foundElem = iElem
            if (foundElem == mainElem) and (index != 0):
                return
        return True

    def unreferenceElement(self, elem):
        # Find last occurrence of this element
        elemIndex = -1
        for index, iElem in self.elements.items():
            foundElem = None
            if hasattr(iElem, 'OD'):
                # A POD element
                if iElem.OD:
                    foundElem = iElem.OD.elem
            else:
                # A PX elem
                foundElem = iElem
            if (foundElem == elem) and (index > elemIndex):
                elemIndex = index
        del self.elements[elemIndex]

    def pushSubBuffer(self, subBuffer):
        '''Sets p_subBuffer at the very end of the buffer.'''
        subIndex = None
        for index, aSubBuffer in self.subBuffers.items():
            if aSubBuffer == subBuffer:
                subIndex = index
                break
        if subIndex != None:
            # Indeed, it is possible that this buffer is not referenced
            # in the parent (if it is a temp buffer generated from a cut)
            del self.subBuffers[subIndex]
            self.subBuffers[self.getLength()] = subBuffer
            self.content += ' '

    def transferAllContent(self):
        '''Transfer all content to parent.'''
        if isinstance(self.parent, FileBuffer):
            # First unreference all elements
            for index in self.getElementIndexes(expressions=False):
                del self.elements[index]
            self.evaluate(self.parent, self.env.context)
        else:
            # Transfer content in itself
            oldParentLength = self.parent.getLength()
            self.parent.write(self.content)
            # Transfer elements
            for index, podElem in self.elements.items():
                self.parent.elements[oldParentLength + index] = podElem
            # Transfer sub-buffers
            for index, buf in self.subBuffers.items():
                self.parent.subBuffers[oldParentLength + index] = buf
        # Empty the buffer
        MemoryBuffer.__init__(self, self.env, self.parent)
        # Change buffer position wrt parent
        self.parent.pushSubBuffer(self)

    def addElement(self, elem, elemType='pod'):
        if elemType == 'pod':
            elem = PodElement.create(elem)
        self.elements[self.getLength()] = elem
        if isinstance(elem, Cell) or isinstance(elem, Table):
            elem.tableInfo = self.env.getTable()
            if isinstance(elem, Cell):
                # Remember where this cell is in the table
                elem.colIndex = elem.tableInfo.curColIndex
        if elem == 'x':
            # See comment on similar statement in the method below.
            self.content += ' '

    def addExpression(self, expression, tiedHook=None):
        # Create the POD expression
        expr = Expression(expression, self.pod)
        if tiedHook: tiedHook.tiedExpression = expr
        self.elements[self.getLength()] = expr
        # To be sure that an expr and an elem can't be found at the same index
        # in the buffer.
        self.content += ' '

    def addAttributes(self):
        '''pod-only: adds an Attributes instance into this buffer.'''
        attrs = Attributes(self.env)
        self.elements[self.getLength()] = attrs
        self.content += ' '
        return attrs

    def addAttribute(self, name, expr):
        '''px-only: adds an Attribute instance into this buffer.'''
        attr = Attribute(name, expr)
        self.elements[self.getLength()] = attr
        self.content += ' '
        return attr

    def _getVariables(self, expr):
        '''Returns variable definitions in p_expr as a list
           ~[(s_varName, s_expr)]~.'''
        exprs = expr.strip().split(';')
        res = []
        for sub in exprs:
            varRes = MemoryBuffer.varRex.match(sub)
            if not varRes:
                raise ParsingError(BAD_VAR_EXPRESSION % sub)
            res.append(varRes.groups())
        return res

    def createAction(self, statementGroup):
        '''Tries to create an action based on p_statementGroup. If the statement
           is not correct, r_ is -1. Else, r_ is the index of the element within
           the buffer that is the object of the action.'''
        res = -1
        try:
            # Check the whole statement group
            if not statementGroup or (len(statementGroup) > 2):
                raise ParsingError(BAD_STATEMENT_GROUP % str(statementGroup))
            # Check the statement
            statement = statementGroup[0]
            aRes = self.actionRex.match(statement)
            if not aRes:
                raise ParsingError(BAD_STATEMENT % statement)
            statementName, podElem, minus, actionType, subExpr = aRes.groups()
            if not (podElem in PodElement.POD_ELEMS):
                raise ParsingError(BAD_ELEMENT % podElem)
            if minus and (not podElem in PodElement.MINUS_ELEMS):
                raise ParsingError(
                    BAD_MINUS % (podElem, PodElement.MINUS_ELEMS))
            indexPodElem = self.getIndex(podElem)
            if indexPodElem == -1:
                raise ParsingError(
                    ELEMENT_NOT_FOUND % (podElem, str([
                        e.__class__.__name__.lower() \
                        for e in list(self.elements.values())])))
            podElem = self.elements[indexPodElem]
            # Check the 'from' clause
            fromClause = None
            source = 'buffer'
            if len(statementGroup) > 1:
                fromClause = statementGroup[1]
                source = 'from'
                if not fromClause.startswith('from '):
                    raise ParsingError(BAD_FROM_CLAUSE % fromClause)
                fromClause = fromClause[5:]
            # Create the action
            if actionType == 'if':
                self.action = IfAction(statementName, self, subExpr, podElem,
                                       minus, source, fromClause)
                self.env.ifActions.append(self.action)
                if self.action.name:
                    # We must register this action as a named action
                    if self.action.name in self.env.namedIfActions:
                        raise ParsingError(DUPLICATE_NAMED_IF)
                    self.env.namedIfActions[self.action.name] = self.action
            elif actionType == 'else':
                if not self.env.ifActions:
                    raise ParsingError(ELSE_WITHOUT_IF)
                # Does the "else" action reference a named "if" action?
                ifReference = subExpr.strip()
                if ifReference:
                    if ifReference not in self.env.namedIfActions:
                        raise ParsingError(ELSE_WITHOUT_NAMED_IF % ifReference)
                    linkedIfAction = self.env.namedIfActions[ifReference]
                    # This "else" action "consumes" the "if" action: this way,
                    # it is not possible to define two "else" actions related to
                    # the same "if".
                    del self.env.namedIfActions[ifReference]
                    self.env.ifActions.remove(linkedIfAction)
                else:
                    linkedIfAction = self.env.ifActions.pop()
                self.action = ElseAction(statementName, self, None, podElem,
                                         minus, source, fromClause,
                                         linkedIfAction)
            elif actionType == 'for':
                forRes = MemoryBuffer.forRex.match(subExpr.strip())
                if not forRes:
                    raise ParsingError(BAD_FOR_EXPRESSION % subExpr)
                iter, subExpr = forRes.groups()
                self.action = ForAction(statementName, self, subExpr, podElem,
                                        minus, iter, source, fromClause)
            elif actionType == 'with':
                variables = self._getVariables(subExpr)
                self.action = VariablesAction(statementName, self, podElem,
                                           minus, variables, source, fromClause)
            else: # null action
                if not fromClause:
                    raise ParsingError(NULL_ACTION_ERROR)
                self.action = NullAction(statementName, self, None, podElem,
                                         None, source, fromClause)
            res = indexPodElem
        except ParsingError as ppe:
            PodError.dump(self, ppe, removeFirstLine=True)
        return res

    def createPxAction(self, elem, actionType, statement):
        '''Creates a PX action and link it to this buffer. If an action is
           already linked to this buffer (in self.action), this action is
           chained behind the last action via self.action.subAction.'''
        res = 0
        statement = statement.strip()
        if actionType == 'for':
            forRes = MemoryBuffer.forRex.match(statement)
            if not forRes:
                raise ParsingError(BAD_FOR_EXPRESSION % statement)
            iter, subExpr = forRes.groups()
            action = ForAction('for', self, subExpr, elem, False, iter,
                               'buffer', None)
        elif actionType == 'if':
            action= IfAction('if', self, statement, elem, False, 'buffer', None)
        elif actionType in ('var', 'var2'):
            variables = self._getVariables(statement)
            action = VariablesAction('var', self, elem, False, variables,
                                     'buffer', None)
        # Is it the first action for this buffer or not?
        if not self.action:
            self.action = action
        else:
            self.action.addSubAction(action)
        return res

    def cut(self, index, keepFirstPart):
        '''Cuts this buffer into 2 parts. Depending on p_keepFirstPart, the 1st
        (from 0 to index-1) or the second (from index to the end) part of the
        buffer is returned as a MemoryBuffer instance without parent; the other
        part is self.'''
        res = MemoryBuffer(self.env, None)
        # Manage buffer meta-info (elements, expressions, subbuffers)
        iter = BufferIterator(self)
        subBuffersToDelete = []
        elementsToDelete = []
        mustShift = False
        while iter.hasNext():
            itemIndex, item = next(iter)
            if keepFirstPart:
                if itemIndex >= index:
                    newIndex = itemIndex-index
                    if isinstance(item, MemoryBuffer):
                        res.subBuffers[newIndex] = item
                        subBuffersToDelete.append(itemIndex)
                    else:
                        res.elements[newIndex] = item
                        elementsToDelete.append(itemIndex)
            else:
                if itemIndex < index:
                    if isinstance(item, MemoryBuffer):
                        res.subBuffers[itemIndex] = item
                        subBuffersToDelete.append(itemIndex)
                    else:
                        res.elements[itemIndex] = item
                        elementsToDelete.append(itemIndex)
                else:
                    mustShift = True
        if elementsToDelete:
            for elemIndex in elementsToDelete:
                del self.elements[elemIndex]
        if subBuffersToDelete:
            for subIndex in subBuffersToDelete:
                del self.subBuffers[subIndex]
        if mustShift:
            elements = {}
            for elemIndex, elem in self.elements.items():
                elements[elemIndex-index] = elem
            self.elements = elements
            subBuffers = {}
            for subIndex, buf in self.subBuffers.items():
                subBuffers[subIndex-index] = buf
            self.subBuffers = subBuffers
        # Manage content
        if keepFirstPart:
            res.write(self.content[index:])
            self.content = self.content[:index]
        else:
            res.write(self.content[:index])
            self.content = self.content[index:]
        return res

    def getElementIndexes(self, expressions=True):
        res = []
        for index, elem in self.elements.items():
            condition = isinstance(elem, Expression) or \
                        isinstance(elem, Attributes)
            if not expressions:
                condition = not condition
            if condition:
                res.append(index)
        return res

    def transferActionIndependentContent(self, actionElemIndex):
        # Manage content to transfer to parent buffer
        if actionElemIndex != 0:
            actionIndependentBuffer = self.cut(actionElemIndex,
                                               keepFirstPart=False)
            actionIndependentBuffer.parent = self.parent
            actionIndependentBuffer.transferAllContent()
            self.parent.pushSubBuffer(self)
        # Manage content to transfer to a child buffer
        actionElemIndex = self.getIndex(
            self.action.elem.__class__.__name__.lower())
        # We recompute actionElemIndex because after cut it may have changed
        elemIndexes = self.getElementIndexes(expressions=False)
        elemIndexes.sort()
        if elemIndexes.index(actionElemIndex) != (len(elemIndexes)-1):
            # I must create a sub-buffer with the impactable elements after
            # the action-related element
            childBuffer = self.cut(elemIndexes[elemIndexes.index(
                actionElemIndex)+1], keepFirstPart=True)
            self.addSubBuffer(childBuffer)
            res = childBuffer
        else:
            res = self
        return res

    def getStartIndex(self, removeMainElems):
        '''When I must dump the buffer, sometimes (if p_removeMainElems is
        True), I must dump only a subset of it. This method returns the start
        index of the buffer part I must dump.'''
        if not removeMainElems: return 0
        # Find the start position of the deepest element to remove
        deepestElem = self.action.elem.DEEPEST_TO_REMOVE
        pos = self.content.find('<%s' % deepestElem.elem)
        pos = pos + len(deepestElem.elem)
        # Now we must find the position of the end of this start tag,
        # skipping potential attributes.
        inAttrValue = False # Are we parsing an attribute value ?
        endTagFound = False # Have we found the end of this tag ?
        while not endTagFound:
            pos += 1
            nextChar = self.content[pos]
            if (nextChar == '>') and not inAttrValue:
                # Yes we have it
                endTagFound = True
            elif nextChar == '"':
                inAttrValue = not inAttrValue
        return pos + 1

    def getStopIndex(self, removeMainElems):
        '''This method returns the stop index of the buffer part I must dump.'''
        if removeMainElems:
            ns = self.env.namespaces
            deepestElem = self.action.elem.DEEPEST_TO_REMOVE
            pos = self.content.rfind('</%s>' % deepestElem.getFullName(ns))
            res = pos
        else:
            res = self.getLength()
        return res

    def removeAutomaticExpressions(self):
        '''When a buffer has an action with minus=True, we must remove the
           "columnsRepeat" expressions automatically inserted by pod. Else, we
           will have problems when computing the index of the part to keep
           (m_getStartIndex).'''
        # Find the start position of the deepest element to remove
        deepestElem = self.action.elem.DEEPEST_TO_REMOVE
        pos = self.content.find('<%s' % deepestElem.elem)
        for index in self.elements.keys():
            if index < pos: del self.elements[index]

    reTagContent = re.compile('<(?P<p>[\w-]+):(?P<f>[\w-]+)(.*?)>.*</(?P=p):' \
                              '(?P=f)>', re.S)
    def evaluate(self, result, context, subElements=True,
                 removeMainElems=False):
        '''Evaluates this buffer given the current p_context and add the result
           into p_result. With pod, p_result is the root file buffer; with px
           it is a memory buffer.'''
        if not subElements:
            # Dump the root tag in this buffer, but not its content
            res = self.reTagContent.match(self.content.strip())
            if not res: result.write(self.content)
            else:
                g = res.group
                result.write('<%s:%s%s></%s:%s>' % (g(1),g(2),g(3),g(1),g(2)))
        else:
            if removeMainElems: self.removeAutomaticExpressions()
            iter = BufferIterator(self)
            currentIndex = self.getStartIndex(removeMainElems)
            while iter.hasNext():
                index, evalEntry = next(iter)
                result.write(self.content[currentIndex:index])
                currentIndex = index + 1
                if isinstance(evalEntry, Expression):
                    try:
                        res, escape = evalEntry.evaluate(context)
                        if escape: result.dumpContent(res)
                        else: result.write(res)
                    except EvaluationError as e:
                        # This exception has already been treated (see the 
                        # "except" block below). Simply re-raise it when needed.
                        if self.env.raiseOnError: raise e
                    except Exception as e:
                        if not self.env.raiseOnError:
                            PodError.dump(result, EVAL_EXPR_ERROR % (
                                          evalEntry.expr, e))
                        else:
                            raise EvaluationError(EVAL_EXPR_ERROR % \
                                        (evalEntry.expr, '\n'+Traceback.get(5)))
                elif isinstance(evalEntry, Attributes) or \
                     isinstance(evalEntry, Attribute):
                    result.write(evalEntry.evaluate(context))
                else: # It is a subBuffer
                    if evalEntry.action:
                        evalEntry.action.execute(result, context)
                    else:
                        result.write(evalEntry.content)
            stopIndex = self.getStopIndex(removeMainElems)
            if currentIndex < (stopIndex-1):
                result.write(self.content[currentIndex:stopIndex])

    def clean(self):
        '''Cleans the buffer content.'''
        self.content = ''
# ------------------------------------------------------------------------------