Linux/Ubuntu Eclipse CDT pretty printer, better view of STL objects during debugging

Eclipse CDT pretty printer is very useful for viewing STL objects in debug mode, just add the following text into any file (usually ~/.gdbinit or /etc/gdb/gdbinit) and add that file into "Debug configuration->Debugger->Main->GDB command file" will do (add the following text starting from the 1st word 'python', inclusive):

python

# Pretty-printers for libstc++.

# Copyright (C) 2008-2013 Free Software Foundation, Inc.

# This program 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.
#
# This program 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 this program.  If not, see <http://www.gnu.org/licenses/>.

import gdb
import itertools
import re

# Try to use the new-style pretty-printing if available.
_use_gdb_pp = True
try:
    import gdb.printing
except ImportError:
    _use_gdb_pp = False

# Try to install type-printers.
_use_type_printing = False
try:
    import gdb.types
    if hasattr(gdb.types, 'TypePrinter'):
        _use_type_printing = True
except ImportError:
    pass

# Starting with the type ORIG, search for the member type NAME.  This
# handles searching upward through superclasses.  This is needed to
# work around http://sourceware.org/bugzilla/show_bug.cgi?id=13615.
def find_type(orig, name):
    typ = orig.strip_typedefs()
    while True:
        search = str(typ) + '::' + name
        try:
            return gdb.lookup_type(search)
        except RuntimeError:
            pass
        # The type was not found, so try the superclass.  We only need
        # to check the first superclass, so we don't bother with
        # anything fancier here.
        field = typ.fields()[0]
        if not field.is_base_class:
            raise ValueError( "Cannot find type %s::%s" % (str(orig), name))
        typ = field.type

class SharedPointerPrinter:
    "Print a shared_ptr or weak_ptr"

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val

    def to_string (self):
        state = 'empty'
        refcounts = self.val['_M_refcount']['_M_pi']
        if refcounts != 0:
            usecount = refcounts['_M_use_count']
            weakcount = refcounts['_M_weak_count']
            if usecount == 0:
                state = 'expired, weak %d' % weakcount
            else:
                state = 'count %d, weak %d' % (usecount, weakcount - 1)
        return '%s (%s) %s' % (self.typename, state, self.val['_M_ptr'])

class UniquePointerPrinter:
    "Print a unique_ptr"

    def __init__ (self, typename, val):
        self.val = val

    def to_string (self):
        v = self.val['_M_t']['_M_head_impl']
        return ('std::unique_ptr<%s> containing %s' % (str(v.type.target()),
                                                       str(v)))

class StdListPrinter:
    "Print a std::list"

    class _iterator:
        def __init__(self, nodetype, head):
            self.nodetype = nodetype
            self.base = head['_M_next']
            self.head = head.address
            self.count = 0

        def __iter__(self):
            return self

        def next(self):
            if self.base == self.head:
                raise StopIteration
            elt = self.base.cast(self.nodetype).dereference()
            self.base = elt['_M_next']
            count = self.count
            self.count = self.count + 1
            return ('[%d]' % count, elt['_M_data'])

    def __init__(self, typename, val):
        self.typename = typename
        self.val = val

    def children(self):
        nodetype = find_type(self.val.type, '_Node')
        nodetype = nodetype.strip_typedefs().pointer()
        return self._iterator(nodetype, self.val['_M_impl']['_M_node'])

    def to_string(self):
        if self.val['_M_impl']['_M_node'].address == self.val['_M_impl']['_M_node']['_M_next']:
            return 'empty %s' % (self.typename)
        return '%s' % (self.typename)

class StdListIteratorPrinter:
    "Print std::list::iterator"

    def __init__(self, typename, val):
        self.val = val
        self.typename = typename

    def to_string(self):
        nodetype = find_type(self.val.type, '_Node')
        nodetype = nodetype.strip_typedefs().pointer()
        return self.val['_M_node'].cast(nodetype).dereference()['_M_data']

class StdSlistPrinter:
    "Print a __gnu_cxx::slist"

    class _iterator:
        def __init__(self, nodetype, head):
            self.nodetype = nodetype
            self.base = head['_M_head']['_M_next']
            self.count = 0

        def __iter__(self):
            return self

        def next(self):
            if self.base == 0:
                raise StopIteration
            elt = self.base.cast(self.nodetype).dereference()
            self.base = elt['_M_next']
            count = self.count
            self.count = self.count + 1
            return ('[%d]' % count, elt['_M_data'])

    def __init__(self, typename, val):
        self.val = val

    def children(self):
        nodetype = find_type(self.val.type, '_Node')
        nodetype = nodetype.strip_typedefs().pointer()
        return self._iterator(nodetype, self.val)

    def to_string(self):
        if self.val['_M_head']['_M_next'] == 0:
            return 'empty __gnu_cxx::slist'
        return '__gnu_cxx::slist'

class StdSlistIteratorPrinter:
    "Print __gnu_cxx::slist::iterator"

    def __init__(self, typename, val):
        self.val = val

    def to_string(self):
        nodetype = find_type(self.val.type, '_Node')
        nodetype = nodetype.strip_typedefs().pointer()
        return self.val['_M_node'].cast(nodetype).dereference()['_M_data']

class StdVectorPrinter:
    "Print a std::vector"

    class _iterator:
        def __init__ (self, start, finish, bitvec):
            self.bitvec = bitvec
            if bitvec:
                self.item   = start['_M_p']
                self.so     = start['_M_offset']
                self.finish = finish['_M_p']
                self.fo     = finish['_M_offset']
                itype = self.item.dereference().type
                self.isize = 8 * itype.sizeof
            else:
                self.item = start
                self.finish = finish
            self.count = 0

        def __iter__(self):
            return self

        def next(self):
            count = self.count
            self.count = self.count + 1
            if self.bitvec:
                if self.item == self.finish and self.so >= self.fo:
                    raise StopIteration
                elt = self.item.dereference()
                if elt & (1 << self.so):
                    obit = 1
                else:
                    obit = 0
                self.so = self.so + 1
                if self.so >= self.isize:
                    self.item = self.item + 1
                    self.so = 0
                return ('[%d]' % count, obit)
            else:
                if self.item == self.finish:
                    raise StopIteration
                elt = self.item.dereference()
                self.item = self.item + 1
                return ('[%d]' % count, elt)

    def __init__(self, typename, val):
        self.typename = typename
        self.val = val
        self.is_bool = val.type.template_argument(0).code  == gdb.TYPE_CODE_BOOL

    def children(self):
        return self._iterator(self.val['_M_impl']['_M_start'],
                              self.val['_M_impl']['_M_finish'],
                              self.is_bool)

    def to_string(self):
        start = self.val['_M_impl']['_M_start']
        finish = self.val['_M_impl']['_M_finish']
        end = self.val['_M_impl']['_M_end_of_storage']
        if self.is_bool:
            start = self.val['_M_impl']['_M_start']['_M_p']
            so    = self.val['_M_impl']['_M_start']['_M_offset']
            finish = self.val['_M_impl']['_M_finish']['_M_p']
            fo     = self.val['_M_impl']['_M_finish']['_M_offset']
            itype = start.dereference().type
            bl = 8 * itype.sizeof
            length   = (bl - so) + bl * ((finish - start) - 1) + fo
            capacity = bl * (end - start)
            return ('%s<bool> of length %d, capacity %d'
                    % (self.typename, int (length), int (capacity)))
        else:
            return ('%s of length %d, capacity %d'
                    % (self.typename, int (finish - start), int (end - start)))

    def display_hint(self):
        return 'array'

class StdVectorIteratorPrinter:
    "Print std::vector::iterator"

    def __init__(self, typename, val):
        self.val = val

    def to_string(self):
        return self.val['_M_current'].dereference()

class StdTuplePrinter:
    "Print a std::tuple"

    class _iterator:
        def __init__ (self, head):
            self.head = head

            # Set the base class as the initial head of the
            # tuple.
            nodes = self.head.type.fields ()
            if len (nodes) == 1:
                # Set the actual head to the first pair.
                self.head  = self.head.cast (nodes[0].type)
            elif len (nodes) != 0:
                raise ValueError( "Top of tuple tree does not consist of a single node.")
            self.count = 0

        def __iter__ (self):
            return self

        def next (self):
            nodes = self.head.type.fields ()
            # Check for further recursions in the inheritance tree.
            if len (nodes) == 0:
                raise StopIteration
            # Check that this iteration has an expected structure.
            if len (nodes) != 2:
                raise ValueError( "Cannot parse more than 2 nodes in a tuple tree.")

            # - Left node is the next recursion parent.
            # - Right node is the actual class contained in the tuple.

            # Process right node.
            impl = self.head.cast (nodes[1].type)

            # Process left node and set it as head.
            self.head  = self.head.cast (nodes[0].type)
            self.count = self.count + 1

            # Finally, check the implementation.  If it is
            # wrapped in _M_head_impl return that, otherwise return
            # the value "as is".
            fields = impl.type.fields ()
            if len (fields) < 1 or fields[0].name != "_M_head_impl":
                return ('[%d]' % self.count, impl)
            else:
                return ('[%d]' % self.count, impl['_M_head_impl'])

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val;

    def children (self):
        return self._iterator (self.val)

    def to_string (self):
        if len (self.val.type.fields ()) == 0:
            return 'empty %s' % (self.typename)
        return '%s containing' % (self.typename)

class StdStackOrQueuePrinter:
    "Print a std::stack or std::queue"

    def __init__ (self, typename, val):
        self.typename = typename
        self.visualizer = gdb.default_visualizer(val['c'])

    def children (self):
        return self.visualizer.children()

    def to_string (self):
        return '%s wrapping: %s' % (self.typename,
                                    self.visualizer.to_string())

    def display_hint (self):
        if hasattr (self.visualizer, 'display_hint'):
            return self.visualizer.display_hint ()
        return None

class RbtreeIterator:
    def __init__(self, rbtree):
        self.size = rbtree['_M_t']['_M_impl']['_M_node_count']
        self.node = rbtree['_M_t']['_M_impl']['_M_header']['_M_left']
        self.count = 0

    def __iter__(self):
        return self

    def __len__(self):
        return int (self.size)

    def next(self):
        if self.count == self.size:
            raise StopIteration
        result = self.node
        self.count = self.count + 1
        if self.count < self.size:
            # Compute the next node.
            node = self.node
            if node.dereference()['_M_right']:
                node = node.dereference()['_M_right']
                while node.dereference()['_M_left']:
                    node = node.dereference()['_M_left']
            else:
                parent = node.dereference()['_M_parent']
                while node == parent.dereference()['_M_right']:
                    node = parent
                    parent = parent.dereference()['_M_parent']
                if node.dereference()['_M_right'] != parent:
                    node = parent
            self.node = node
        return result

# This is a pretty printer for std::_Rb_tree_iterator (which is
# std::map::iterator), and has nothing to do with the RbtreeIterator
# class above.
class StdRbtreeIteratorPrinter:
    "Print std::map::iterator"

    def __init__ (self, typename, val):
        self.val = val

    def to_string (self):
        typename = str(self.val.type.strip_typedefs()) + '::_Link_type'
        nodetype = gdb.lookup_type(typename).strip_typedefs()
        return self.val.cast(nodetype).dereference()['_M_value_field']

class StdDebugIteratorPrinter:
    "Print a debug enabled version of an iterator"

    def __init__ (self, typename, val):
        self.val = val

    # Just strip away the encapsulating __gnu_debug::_Safe_iterator
    # and return the wrapped iterator value.
    def to_string (self):
        itype = self.val.type.template_argument(0)
        return self.val['_M_current'].cast(itype)

class StdMapPrinter:
    "Print a std::map or std::multimap"

    # Turn an RbtreeIterator into a pretty-print iterator.
    class _iter:
        def __init__(self, rbiter, type):
            self.rbiter = rbiter
            self.count = 0
            self.type = type

        def __iter__(self):
            return self

        def next(self):
            if self.count % 2 == 0:
                n = self.rbiter.next()
                n = n.cast(self.type).dereference()['_M_value_field']
                self.pair = n
                item = n['first']
            else:
                item = self.pair['second']
            result = ('[%d]' % self.count, item)
            self.count = self.count + 1
            return result

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val

    def to_string (self):
        return '%s with %d elements' % (self.typename,
                                        len (RbtreeIterator (self.val)))

    def children (self):
        rep_type = find_type(self.val.type, '_Rep_type')
        node = find_type(rep_type, '_Link_type')
        node = node.strip_typedefs()
        return self._iter (RbtreeIterator (self.val), node)

    def display_hint (self):
        return 'map'

class StdSetPrinter:
    "Print a std::set or std::multiset"

    # Turn an RbtreeIterator into a pretty-print iterator.
    class _iter:
        def __init__(self, rbiter, type):
            self.rbiter = rbiter
            self.count = 0
            self.type = type

        def __iter__(self):
            return self

        def next(self):
            item = self.rbiter.next()
            item = item.cast(self.type).dereference()['_M_value_field']
            # FIXME: this is weird ... what to do?
            # Maybe a 'set' display hint?
            result = ('[%d]' % self.count, item)
            self.count = self.count + 1
            return result

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val

    def to_string (self):
        return '%s with %d elements' % (self.typename,
                                        len (RbtreeIterator (self.val)))

    def children (self):
        rep_type = find_type(self.val.type, '_Rep_type')
        node = find_type(rep_type, '_Link_type')
        node = node.strip_typedefs()
        return self._iter (RbtreeIterator (self.val), node)

class StdBitsetPrinter:
    "Print a std::bitset"

    def __init__(self, typename, val):
        self.typename = typename
        self.val = val

    def to_string (self):
        # If template_argument handled values, we could print the
        # size.  Or we could use a regexp on the type.
        return '%s' % (self.typename)

    def children (self):
        words = self.val['_M_w']
        wtype = words.type

        # The _M_w member can be either an unsigned long, or an
        # array.  This depends on the template specialization used.
        # If it is a single long, convert to a single element list.
        if wtype.code == gdb.TYPE_CODE_ARRAY:
            tsize = wtype.target ().sizeof
        else:
            words = [words]
            tsize = wtype.sizeof

        nwords = wtype.sizeof / tsize
        result = []
        byte = 0
        while byte < nwords:
            w = words[byte]
            bit = 0
            while w != 0:
                if (w & 1) != 0:
                    # Another spot where we could use 'set'?
                    result.append(('[%d]' % (byte * tsize * 8 + bit), 1))
                bit = bit + 1
                w = w >> 1
            byte = byte + 1
        return result

class StdDequePrinter:
    "Print a std::deque"

    class _iter:
        def __init__(self, node, start, end, last, buffer_size):
            self.node = node
            self.p = start
            self.end = end
            self.last = last
            self.buffer_size = buffer_size
            self.count = 0

        def __iter__(self):
            return self

        def next(self):
            if self.p == self.last:
                raise StopIteration

            result = ('[%d]' % self.count, self.p.dereference())
            self.count = self.count + 1

            # Advance the 'cur' pointer.
            self.p = self.p + 1
            if self.p == self.end:
                # If we got to the end of this bucket, move to the
                # next bucket.
                self.node = self.node + 1
                self.p = self.node[0]
                self.end = self.p + self.buffer_size

            return result

    def __init__(self, typename, val):
        self.typename = typename
        self.val = val
        self.elttype = val.type.template_argument(0)
        size = self.elttype.sizeof
        if size < 512:
            self.buffer_size = int (512 / size)
        else:
            self.buffer_size = 1

    def to_string(self):
        start = self.val['_M_impl']['_M_start']
        end = self.val['_M_impl']['_M_finish']

        delta_n = end['_M_node'] - start['_M_node'] - 1
        delta_s = start['_M_last'] - start['_M_cur']
        delta_e = end['_M_cur'] - end['_M_first']

        size = self.buffer_size * delta_n + delta_s + delta_e

        return '%s with %d elements' % (self.typename, long (size))

    def children(self):
        start = self.val['_M_impl']['_M_start']
        end = self.val['_M_impl']['_M_finish']
        return self._iter(start['_M_node'], start['_M_cur'], start['_M_last'],
                          end['_M_cur'], self.buffer_size)

    def display_hint (self):
        return 'array'

class StdDequeIteratorPrinter:
    "Print std::deque::iterator"

    def __init__(self, typename, val):
        self.val = val

    def to_string(self):
        return self.val['_M_cur'].dereference()

class StdStringPrinter:
    "Print a std::basic_string of some kind"

    def __init__(self, typename, val):
        self.val = val

    def to_string(self):
        # Make sure &string works, too.
        type = self.val.type
        if type.code == gdb.TYPE_CODE_REF:
            type = type.target ()

        # Calculate the length of the string so that to_string returns
        # the string according to length, not according to first null
        # encountered.
        ptr = self.val ['_M_dataplus']['_M_p']
        realtype = type.unqualified ().strip_typedefs ()
        reptype = gdb.lookup_type (str (realtype) + '::_Rep').pointer ()
        header = ptr.cast(reptype) - 1
        len = header.dereference ()['_M_length']
        if hasattr(ptr, "lazy_string"):
            return ptr.lazy_string (length = len)
        return ptr.string (length = len)

    def display_hint (self):
        return 'string'

class Tr1HashtableIterator:
    def __init__ (self, hash):
        self.buckets = hash['_M_buckets']
        self.bucket = 0
        self.bucket_count = hash['_M_bucket_count']
        self.node_type = find_type(hash.type, '_Node').pointer()
        self.node = 0
        while self.bucket != self.bucket_count:
            self.node = self.buckets[self.bucket]
            if self.node:
                break
            self.bucket = self.bucket + 1       

    def __iter__ (self):
        return self

    def next (self):
        if self.node == 0:
            raise StopIteration
        node = self.node.cast(self.node_type)
        result = node.dereference()['_M_v']
        self.node = node.dereference()['_M_next'];
        if self.node == 0:
            self.bucket = self.bucket + 1
            while self.bucket != self.bucket_count:
                self.node = self.buckets[self.bucket]
                if self.node:
                    break
                self.bucket = self.bucket + 1
        return result

class StdHashtableIterator:
    def __init__(self, hash):
        self.node = hash['_M_before_begin']['_M_nxt']
        self.node_type = find_type(hash.type, '__node_type').pointer()

    def __iter__(self):
        return self

    def next(self):
        if self.node == 0:
            raise StopIteration
        elt = self.node.cast(self.node_type).dereference()
        self.node = elt['_M_nxt']
        valptr = elt['_M_storage'].address
        valptr = valptr.cast(elt.type.template_argument(0).pointer())
        return valptr.dereference()

class Tr1UnorderedSetPrinter:
    "Print a tr1::unordered_set"

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val

    def hashtable (self):
        if self.typename.startswith('std::tr1'):
            return self.val
        return self.val['_M_h']

    def to_string (self):
        return '%s with %d elements' % (self.typename, self.hashtable()['_M_element_count'])

    @staticmethod
    def format_count (i):
        return '[%d]' % i

    def children (self):
        counter = itertools.imap (self.format_count, itertools.count())
        if self.typename.startswith('std::tr1'):
            return itertools.izip (counter, Tr1HashtableIterator (self.hashtable()))
        return itertools.izip (counter, StdHashtableIterator (self.hashtable()))

class Tr1UnorderedMapPrinter:
    "Print a tr1::unordered_map"

    def __init__ (self, typename, val):
        self.typename = typename
        self.val = val

    def hashtable (self):
        if self.typename.startswith('std::tr1'):
            return self.val
        return self.val['_M_h']

    def to_string (self):
        return '%s with %d elements' % (self.typename, self.hashtable()['_M_element_count'])

    @staticmethod
    def flatten (list):
        for elt in list:
            for i in elt:
                yield i

    @staticmethod
    def format_one (elt):
        return (elt['first'], elt['second'])

    @staticmethod
    def format_count (i):
        return '[%d]' % i

    def children (self):
        counter = itertools.imap (self.format_count, itertools.count())
        # Map over the hash table and flatten the result.
        if self.typename.startswith('std::tr1'):
            data = self.flatten (itertools.imap (self.format_one, Tr1HashtableIterator (self.hashtable())))
            # Zip the two iterators together.
            return itertools.izip (counter, data)
        data = self.flatten (itertools.imap (self.format_one, StdHashtableIterator (self.hashtable())))
        # Zip the two iterators together.
        return itertools.izip (counter, data)
       

    def display_hint (self):
        return 'map'

class StdForwardListPrinter:
    "Print a std::forward_list"

    class _iterator:
        def __init__(self, nodetype, head):
            self.nodetype = nodetype
            self.base = head['_M_next']
            self.count = 0

        def __iter__(self):
            return self

        def next(self):
            if self.base == 0:
                raise StopIteration
            elt = self.base.cast(self.nodetype).dereference()
            self.base = elt['_M_next']
            count = self.count
            self.count = self.count + 1
            valptr = elt['_M_storage'].address
            valptr = valptr.cast(elt.type.template_argument(0).pointer())
            return ('[%d]' % count, valptr.dereference())

    def __init__(self, typename, val):
        self.val = val
        self.typename = typename

    def children(self):
        nodetype = find_type(self.val.type, '_Node')
        nodetype = nodetype.strip_typedefs().pointer()
        return self._iterator(nodetype, self.val['_M_impl']['_M_head'])

    def to_string(self):
        if self.val['_M_impl']['_M_head']['_M_next'] == 0:
            return 'empty %s' % (self.typename)
        return '%s' % (self.typename)


# A "regular expression" printer which conforms to the
# "SubPrettyPrinter" protocol from gdb.printing.
class RxPrinter(object):
    def __init__(self, name, function):
        super(RxPrinter, self).__init__()
        self.name = name
        self.function = function
        self.enabled = True

    def invoke(self, value):
        if not self.enabled:
            return None

        if value.type.code == gdb.TYPE_CODE_REF:
            if hasattr(gdb.Value,"referenced_value"):
                value = value.referenced_value()

        return self.function(self.name, value)

# A pretty-printer that conforms to the "PrettyPrinter" protocol from
# gdb.printing.  It can also be used directly as an old-style printer.
class Printer(object):
    def __init__(self, name):
        super(Printer, self).__init__()
        self.name = name
        self.subprinters = []
        self.lookup = {}
        self.enabled = True
        self.compiled_rx = re.compile('^([a-zA-Z0-9_:]+)<.*>$')

    def add(self, name, function):
        # A small sanity check.
        # FIXME
        if not self.compiled_rx.match(name + '<>'):
            raise ValueError( 'libstdc++ programming error: "%s" does not match' % name)
        printer = RxPrinter(name, function)
        self.subprinters.append(printer)
        self.lookup[name] = printer

    # Add a name using _GLIBCXX_BEGIN_NAMESPACE_VERSION.
    def add_version(self, base, name, function):
        self.add(base + name, function)
        self.add(base + '__7::' + name, function)

    # Add a name using _GLIBCXX_BEGIN_NAMESPACE_CONTAINER.
    def add_container(self, base, name, function):
        self.add_version(base, name, function)
        self.add_version(base + '__cxx1998::', name, function)

    @staticmethod
    def get_basic_type(type):
        # If it points to a reference, get the reference.
        if type.code == gdb.TYPE_CODE_REF:
            type = type.target ()

        # Get the unqualified type, stripped of typedefs.
        type = type.unqualified ().strip_typedefs ()

        return type.tag

    def __call__(self, val):
        typename = self.get_basic_type(val.type)
        if not typename:
            return None

        # All the types we match are template types, so we can use a
        # dictionary.
        match = self.compiled_rx.match(typename)
        if not match:
            return None

        basename = match.group(1)

        if val.type.code == gdb.TYPE_CODE_REF:
            if hasattr(gdb.Value,"referenced_value"):
                val = val.referenced_value()

        if basename in self.lookup:
            return self.lookup[basename].invoke(val)

        # Cannot find a pretty printer.  Return None.
        return None

libstdcxx_printer = None

class FilteringTypePrinter(object):
    def __init__(self, match, name):
        self.match = match
        self.name = name
        self.enabled = True

    class _recognizer(object):
        def __init__(self, match, name):
            self.match = match
            self.name = name
            self.type_obj = None

        def recognize(self, type_obj):
            if type_obj.tag is None:
                return None

            if self.type_obj is None:
                if not self.match in type_obj.tag:
                    # Filter didn't match.
                    return None
                try:
                    self.type_obj = gdb.lookup_type(self.name).strip_typedefs()
                except:
                    pass
            if self.type_obj == type_obj:
                return self.name
            return None

    def instantiate(self):
        return self._recognizer(self.match, self.name)

def add_one_type_printer(obj, match, name):
    printer = FilteringTypePrinter(match, 'std::' + name)
    gdb.types.register_type_printer(obj, printer)

def register_type_printers(obj):
    global _use_type_printing

    if not _use_type_printing:
        return

    for pfx in ('', 'w'):
        add_one_type_printer(obj, 'basic_string', pfx + 'string')
        add_one_type_printer(obj, 'basic_ios', pfx + 'ios')
        add_one_type_printer(obj, 'basic_streambuf', pfx + 'streambuf')
        add_one_type_printer(obj, 'basic_istream', pfx + 'istream')
        add_one_type_printer(obj, 'basic_ostream', pfx + 'ostream')
        add_one_type_printer(obj, 'basic_iostream', pfx + 'iostream')
        add_one_type_printer(obj, 'basic_stringbuf', pfx + 'stringbuf')
        add_one_type_printer(obj, 'basic_istringstream',
                                 pfx + 'istringstream')
        add_one_type_printer(obj, 'basic_ostringstream',
                                 pfx + 'ostringstream')
        add_one_type_printer(obj, 'basic_stringstream',
                                 pfx + 'stringstream')
        add_one_type_printer(obj, 'basic_filebuf', pfx + 'filebuf')
        add_one_type_printer(obj, 'basic_ifstream', pfx + 'ifstream')
        add_one_type_printer(obj, 'basic_ofstream', pfx + 'ofstream')
        add_one_type_printer(obj, 'basic_fstream', pfx + 'fstream')
        add_one_type_printer(obj, 'basic_regex', pfx + 'regex')
        add_one_type_printer(obj, 'sub_match', pfx + 'csub_match')
        add_one_type_printer(obj, 'sub_match', pfx + 'ssub_match')
        add_one_type_printer(obj, 'match_results', pfx + 'cmatch')
        add_one_type_printer(obj, 'match_results', pfx + 'smatch')
        add_one_type_printer(obj, 'regex_iterator', pfx + 'cregex_iterator')
        add_one_type_printer(obj, 'regex_iterator', pfx + 'sregex_iterator')
        add_one_type_printer(obj, 'regex_token_iterator',
                                 pfx + 'cregex_token_iterator')
        add_one_type_printer(obj, 'regex_token_iterator',
                                 pfx + 'sregex_token_iterator')

    # Note that we can't have a printer for std::wstreampos, because
    # it shares the same underlying type as std::streampos.
    add_one_type_printer(obj, 'fpos', 'streampos')
    add_one_type_printer(obj, 'basic_string', 'u16string')
    add_one_type_printer(obj, 'basic_string', 'u32string')

    for dur in ('nanoseconds', 'microseconds', 'milliseconds',
                'seconds', 'minutes', 'hours'):
        add_one_type_printer(obj, 'duration', dur)

    add_one_type_printer(obj, 'linear_congruential_engine', 'minstd_rand0')
    add_one_type_printer(obj, 'linear_congruential_engine', 'minstd_rand')
    add_one_type_printer(obj, 'mersenne_twister_engine', 'mt19937')
    add_one_type_printer(obj, 'mersenne_twister_engine', 'mt19937_64')
    add_one_type_printer(obj, 'subtract_with_carry_engine', 'ranlux24_base')
    add_one_type_printer(obj, 'subtract_with_carry_engine', 'ranlux48_base')
    add_one_type_printer(obj, 'discard_block_engine', 'ranlux24')
    add_one_type_printer(obj, 'discard_block_engine', 'ranlux48')
    add_one_type_printer(obj, 'shuffle_order_engine', 'knuth_b')

def register_libstdcxx_printers (obj):
    "Register libstdc++ pretty-printers with objfile Obj."

    global _use_gdb_pp
    global libstdcxx_printer

    if _use_gdb_pp:
        gdb.printing.register_pretty_printer(obj, libstdcxx_printer)
    else:
        if obj is None:
            obj = gdb
        obj.pretty_printers.append(libstdcxx_printer)

    register_type_printers(obj)

def build_libstdcxx_dictionary ():
    global libstdcxx_printer

    libstdcxx_printer = Printer("libstdc++-v6")

    # For _GLIBCXX_BEGIN_NAMESPACE_VERSION.
    vers = '(__7::)?'
    # For _GLIBCXX_BEGIN_NAMESPACE_CONTAINER.
    container = '(__cxx1998::' + vers + ')?'

    # libstdc++ objects requiring pretty-printing.
    # In order from:
    # http://gcc.gnu.org/onlinedocs/libstdc++/latest-doxygen/a01847.html
    libstdcxx_printer.add_version('std::', 'basic_string', StdStringPrinter)
    libstdcxx_printer.add_container('std::', 'bitset', StdBitsetPrinter)
    libstdcxx_printer.add_container('std::', 'deque', StdDequePrinter)
    libstdcxx_printer.add_container('std::', 'list', StdListPrinter)
    libstdcxx_printer.add_container('std::', 'map', StdMapPrinter)
    libstdcxx_printer.add_container('std::', 'multimap', StdMapPrinter)
    libstdcxx_printer.add_container('std::', 'multiset', StdSetPrinter)
    libstdcxx_printer.add_version('std::', 'priority_queue',
                                  StdStackOrQueuePrinter)
    libstdcxx_printer.add_version('std::', 'queue', StdStackOrQueuePrinter)
    libstdcxx_printer.add_version('std::', 'tuple', StdTuplePrinter)
    libstdcxx_printer.add_container('std::', 'set', StdSetPrinter)
    libstdcxx_printer.add_version('std::', 'stack', StdStackOrQueuePrinter)
    libstdcxx_printer.add_version('std::', 'unique_ptr', UniquePointerPrinter)
    libstdcxx_printer.add_container('std::', 'vector', StdVectorPrinter)
    # vector<bool>

    # Printer registrations for classes compiled with -D_GLIBCXX_DEBUG.
    libstdcxx_printer.add('std::__debug::bitset', StdBitsetPrinter)
    libstdcxx_printer.add('std::__debug::deque', StdDequePrinter)
    libstdcxx_printer.add('std::__debug::list', StdListPrinter)
    libstdcxx_printer.add('std::__debug::map', StdMapPrinter)
    libstdcxx_printer.add('std::__debug::multimap', StdMapPrinter)
    libstdcxx_printer.add('std::__debug::multiset', StdSetPrinter)
    libstdcxx_printer.add('std::__debug::priority_queue',
                          StdStackOrQueuePrinter)
    libstdcxx_printer.add('std::__debug::queue', StdStackOrQueuePrinter)
    libstdcxx_printer.add('std::__debug::set', StdSetPrinter)
    libstdcxx_printer.add('std::__debug::stack', StdStackOrQueuePrinter)
    libstdcxx_printer.add('std::__debug::unique_ptr', UniquePointerPrinter)
    libstdcxx_printer.add('std::__debug::vector', StdVectorPrinter)

    # These are the TR1 and C++0x printers.
    # For array - the default GDB pretty-printer seems reasonable.
    libstdcxx_printer.add_version('std::', 'shared_ptr', SharedPointerPrinter)
    libstdcxx_printer.add_version('std::', 'weak_ptr', SharedPointerPrinter)
    libstdcxx_printer.add_container('std::', 'unordered_map',
                                    Tr1UnorderedMapPrinter)
    libstdcxx_printer.add_container('std::', 'unordered_set',
                                    Tr1UnorderedSetPrinter)
    libstdcxx_printer.add_container('std::', 'unordered_multimap',
                                    Tr1UnorderedMapPrinter)
    libstdcxx_printer.add_container('std::', 'unordered_multiset',
                                    Tr1UnorderedSetPrinter)
    libstdcxx_printer.add_container('std::', 'forward_list',
                                    StdForwardListPrinter)

    libstdcxx_printer.add_version('std::tr1::', 'shared_ptr', SharedPointerPrinter)
    libstdcxx_printer.add_version('std::tr1::', 'weak_ptr', SharedPointerPrinter)
    libstdcxx_printer.add_version('std::tr1::', 'unordered_map',
                                  Tr1UnorderedMapPrinter)
    libstdcxx_printer.add_version('std::tr1::', 'unordered_set',
                                  Tr1UnorderedSetPrinter)
    libstdcxx_printer.add_version('std::tr1::', 'unordered_multimap',
                                  Tr1UnorderedMapPrinter)
    libstdcxx_printer.add_version('std::tr1::', 'unordered_multiset',
                                  Tr1UnorderedSetPrinter)

    # These are the C++0x printer registrations for -D_GLIBCXX_DEBUG cases.
    # The tr1 namespace printers do not seem to have any debug
    # equivalents, so do no register them.
    libstdcxx_printer.add('std::__debug::unordered_map',
                          Tr1UnorderedMapPrinter)
    libstdcxx_printer.add('std::__debug::unordered_set',
                          Tr1UnorderedSetPrinter)
    libstdcxx_printer.add('std::__debug::unordered_multimap',
                          Tr1UnorderedMapPrinter)
    libstdcxx_printer.add('std::__debug::unordered_multiset',
                          Tr1UnorderedSetPrinter)
    libstdcxx_printer.add('std::__debug::forward_list',
                          StdForwardListPrinter)


    # Extensions.
    libstdcxx_printer.add_version('__gnu_cxx::', 'slist', StdSlistPrinter)

    if True:
        # These shouldn't be necessary, if GDB "print *i" worked.
        # But it often doesn't, so here they are.
        libstdcxx_printer.add_container('std::', '_List_iterator',
                                        StdListIteratorPrinter)
        libstdcxx_printer.add_container('std::', '_List_const_iterator',
                                        StdListIteratorPrinter)
        libstdcxx_printer.add_version('std::', '_Rb_tree_iterator',
                                      StdRbtreeIteratorPrinter)
        libstdcxx_printer.add_version('std::', '_Rb_tree_const_iterator',
                                      StdRbtreeIteratorPrinter)
        libstdcxx_printer.add_container('std::', '_Deque_iterator',
                                        StdDequeIteratorPrinter)
        libstdcxx_printer.add_container('std::', '_Deque_const_iterator',
                                        StdDequeIteratorPrinter)
        libstdcxx_printer.add_version('__gnu_cxx::', '__normal_iterator',
                                      StdVectorIteratorPrinter)
        libstdcxx_printer.add_version('__gnu_cxx::', '_Slist_iterator',
                                      StdSlistIteratorPrinter)

        # Debug (compiled with -D_GLIBCXX_DEBUG) printer
        # registrations.  The Rb_tree debug iterator when unwrapped
        # from the encapsulating __gnu_debug::_Safe_iterator does not
        # have the __norm namespace. Just use the existing printer
        # registration for that.
        libstdcxx_printer.add('__gnu_debug::_Safe_iterator',
                              StdDebugIteratorPrinter)
        libstdcxx_printer.add('std::__norm::_List_iterator',
                              StdListIteratorPrinter)
        libstdcxx_printer.add('std::__norm::_List_const_iterator',
                              StdListIteratorPrinter)
        libstdcxx_printer.add('std::__norm::_Deque_const_iterator',
                              StdDequeIteratorPrinter)
        libstdcxx_printer.add('std::__norm::_Deque_iterator',
                              StdDequeIteratorPrinter)

build_libstdcxx_dictionary ()

register_libstdcxx_printers (None)

C++ STL how to print unicode string and characters

1. essential headers

#include <iostream>
#include <fstream>
#include <algorithm>
#include <sstream>
#include <clocale>
#include <cwchar>   // for mbstate_t
#include <cstddef>  // for std::size_t

2.

in order to use wcout to print wchar_t[] or wstring, you must first run

 setlocale( LC_ALL, "en_US.UTF-8" );
once this is set, you can no longer use cout

3.

in order to use wifstream to read, you must first run imbue() :

 wifstream fin("utf8.txt");

 fin.imbue(std::locale("en_US.UTF-8"));

sample code that works correctly:

#include <stdio.h>

#include <stdlib.h>

#include <string>

#include <iostream>

#include <fstream>

#include <sstream>

#include <locale>

#include <cwchar>   // for mbstate_t

#include <cstddef>  // for std::size_t

using namespace std;

#define def(a,b) typeof(b) a=b

#define dec(a,b) typeof(b) a

int main(int argc, char *argv[]){

    std::locale utf8_locale("en_US.UTF-8");

    wchar_t ss[] = L"你是谁我？asd asdf d";

    wstring ws(L"机器，1234电脑");

    wifstream fin("a.txt");

    fin.imbue(utf8_locale);

    wstring s;

    setlocale( LC_ALL, "en_US.UTF-8" ); // this might fail, you should check return value

    wcout << ws << endl;

    wcout << ss << endl;

    while(fin>>s){

        wcout << L"size=" << s.size() << endl;

        wcout << L"length=" << s.length() << endl;

        for(int x=0; x<s.length(); x++)

            wcout << s[x] << L" ";

        wcout << endl;

    }

    return  0;

}

a.txt contains 1 line:

我是 谁？a sd f

output:

机器，1234电脑

你是谁我？asd asdf d

size=2

length=2

我 是

size=3

length=3

谁 ？ a

size=2

length=2

s d

size=1

length=1

f

Take note that your console must support utf8 and console locale must be set to en_US.UTF-8

To convert between UTF-8 encoded string and wstring, use the following:

#include <codecvt>
#include <string>

// convert UTF-8 string to wstring
std::wstring utf8_string_to_wstring (const std::string& str)
{
    std::wstring_convert<std::codecvt_utf8<wchar_t>> myconv;
    return myconv.from_bytes(str);
}

// convert wstring to UTF-8 string
std::string utf8_wstring_to_string (const std::wstring& str)
{
    std::wstring_convert<std::codecvt_utf8<wchar_t>> myconv;
    return myconv.to_bytes(str);
}

 
However, since <codecvt> is not implemented by gcc/g++ yet, you can only use the following to do the manual conversion:

void utf8toWStr(WStr& dest, const Str& src){
 dest.clear();
 wchar_t w = 0;
 int bytes = 0;
 wchar_t err = L'�';
 for (size_t i = 0; i < src.size(); i++){
  unsigned char c = (unsigned char)src[i];
  if (c <= 0x7f){//first byte
   if (bytes){
    dest.push_back(err);
    bytes = 0;
   }
   dest.push_back((wchar_t)c);
  }
  else if (c <= 0xbf){//second/third/etc byte
   if (bytes){
    w = ((w << 6)|(c & 0x3f));
    bytes--;
    if (bytes == 0)
     dest.push_back(w);
   }
   else
    dest.push_back(err);
  }
  else if (c <= 0xdf){//2byte sequence start
   bytes = 1;
   w = c & 0x1f;
  }
  else if (c <= 0xef){//3byte sequence start
   bytes = 2;
   w = c & 0x0f;
  }
  else if (c <= 0xf7){//3byte sequence start
   bytes = 3;
   w = c & 0x07;
  }
  else{
   dest.push_back(err);
   bytes = 0;
  }
 }
 if (bytes)
  dest.push_back(err);
}

void wstrToUtf8(Str& dest, const WStr& src){
 dest.clear();
 for (size_t i = 0; i < src.size(); i++){
  wchar_t w = src[i];
  if (w <= 0x7f)
   dest.push_back((char)w);
  else if (w <= 0x7ff){
   dest.push_back(0xc0 | ((w >> 6)& 0x1f));
   dest.push_back(0x80| (w & 0x3f));
  }
  else if (w <= 0xffff){
   dest.push_back(0xe0 | ((w >> 12)& 0x0f));
   dest.push_back(0x80| ((w >> 6) & 0x3f));
   dest.push_back(0x80| (w & 0x3f));
  }
  else if (w <= 0x10ffff){
   dest.push_back(0xf0 | ((w >> 18)& 0x07));
   dest.push_back(0x80| ((w >> 12) & 0x3f));
   dest.push_back(0x80| ((w >> 6) & 0x3f));
   dest.push_back(0x80| (w & 0x3f));
  }
  else
   dest.push_back('?');
 }
}

Str wstrToUtf8(const WStr& str){
 Str result;
 wstrToUtf8(result, str);
 return result;
}

WStr utf8toWStr(const Str& str){
 WStr result;
 utf8toWStr(result, str);
 return result;
}

std::ostream& operator<<(std::ostream& f, const WStr& s){
 Str s1;
 wstrToUtf8(s1, s);
 f << s1;
 return f;
}

std::istream& operator>>(std::istream& f, WStr& s){
 Str s1;
 f >> s1;
 utf8toWStr(s, s1);
 return f;
}

The world's best vimrc

# my latest version of .vimrc
# F2 toggle autoindent during paste
# F3/F4 show/hide line number
# F5/F6 set/unset scrollbind
# F7/F8 toggles horizontal vertical layout of 2 file windows
# F9/F10 set/unset wrap
# F12 vimdiff ignore whitespace/tab
# insert mode mouse can scroll screen, set cursor position
# non-insert mode mouse can scroll cursor, terminal mouse highlight automatically save into copy/paste buffer
# d#d : delete # lines, contents no longer go into copy/paste buffer
# c#c : cut # lines, contents go into copy/paste buffer
# mouse can work past the 220th column
# vimdiff : all text visible (evening scheme)

filetype plugin indent on
set mouse=a
set tabstop=4
set shiftwidth=4
set autoindent
set smartindent
set nowrap
set backspace=2
syntax on
autocmd FileType python set tabstop=4|set shiftwidth=4|set noexpandtab

set noexpandtab
set pastetoggle=<F2>
nnoremap <silent> <F8> :TlistToggle<CR>
inoremap <C-J> <C-\><C-O>b
inoremap <C-K> <C-\><C-O>w
set hlsearch
set ttymouse=xterm2
hi comment ctermfg=green
set fileencodings=utf8,gb2312
map <F3> <Esc>:set number<CR>
map <F4> <Esc>:set nonumber<CR>
map <F5> <Esc>:set scrollbind<CR>
map <F6> <Esc>:set noscrollbind<CR>
map <F7> <C-W>t<C-W>K
map <F8> <C-W>t<C-W>H
map <F9> <Esc>:set wrap<CR>
map <F10> <Esc>:set nowrap<CR>
map <F12> <Esc>:set diffopt+=iwhite<CR>
"set viminfo='10,\"100,:20,%,n~/.viminfo
"set indentexpr=''
"set noautoindent
"autocmd BufRead,BufNewFile *.cu set noic cin autoindent
if &diff
    colorscheme evening
endif

if has("mouse_sgr")
    set ttymouse=sgr
else
    set ttymouse=xterm2
end

function! ResCur()
  if line("'\"") <= line("$")
    normal! g`"
    return 1
  endif
endfunction

augroup resCur
  autocmd!
  autocmd BufWinEnter * call ResCur()
augroup END

nnoremap d "_d
vnoremap d "_d
vnoremap p "_p


let g:LargeFile = 1024 * 1024 * 10
augroup LargeFile
 autocmd BufReadPre * let f=getfsize(expand("<afile>")) | if f > g:LargeFile || f == -2 | call LargeFile() | endif
augroup END

function LargeFile()
 " no syntax highlighting etc
 set eventignore+=FileType
 " save memory when other file is viewed
 setlocal bufhidden=unload
 " is read-only (write with :w new_filename)
 " setlocal buftype=nowrite
 " no undo possible
 setlocal noswapfile
 " display message
 autocmd VimEnter *  echo "The file is larger than " . (g:LargeFile / 1024 / 1024) . " MB, so some options are changed (see .vimrc for details)."
endfunction

Dual-mode dual-boot Ubuntu with Windows 7 without burning a CD

The term "Dual-mode dual-boot" refers to a single copy of OS which can run either inside Virtual Machine or on its own standalone. You can launch it by either booting directly into it upon machine restart (usually referred to as "dual-boot") or running it inside a virtual machine. This way of installation is very useful for some people like me because Windows is good for playing large-scale computer games while many programming tasks/packages can only be run in Linux environment and all programs running inside virtual machine has performance penalty.

Dual-booting Ubuntu with Windows7 is no longer a big problem for many people. Just a bit of googling will give you tons of solutions. However, to install dual-mode dual-boot Ubuntu on the same harddisk of the currently running Windows7 is kind of headache especially for 64-bit windows because of the new restriction on raw disk access from Vista onwards ( support.microsoft.com/kb/942448 ) as well as Windows 7 x64 forbids loading kernel drivers without Microsoft digital signature (this can be solved by enabling testsigning in boot config using bcdedit).

I have eventually managed to install Ubuntu 12.04 64-bit on my Windows 7 Ultimate 64-bit using VirtualBox 4.2.4 using a Ubuntu ISO image without burning a physical CD. The copy of Ubuntu is installed inside Windows7's VirtualBox but can run not only inside Windows7's virtualBox but also on its own (restart computer and boot directly into the same copy of Ubuntu).

Warning: DO NOT try this if you are NOT an expert in OS boot principle and harddisk partition, attempting things without understanding what exactly it does may result in total loss of your harddisk data and at the same time rendering your current system unbootable. 

Here're the required stuffs you need to prepare:
1. VirtualBox 4.2.4
2. EasyBCD 2.2
3. EaseUS Partition Master (in case needed)
4. Ubuntu CD (12.04 I used)
5. basic knowledge in Command Prompt and Windows disk management
6. in BIOS, enable Intel virtualization technology, otherwise, 64-bit OS cannot be emulated in virtual machine, only 32-bit OS can be emulated

To get started:
1. Make sure you are in administrator mode with UAC disabled, so every operation has system administrator privilege;

2. run the following command in command prompt to enable test signing mode so that some kernel drivers can be loaded (I'm not sure if it's neccesary)
bcdedit.exe -set TESTSIGNING ON

3. in "Windows disk management" (by right-click Computer -> Manage -> Storage -> Disk Management), make sure you have 2 empty partitions prepared for Ubuntu, 1 for system (suggest >=8GB), 1 for swap (total memory size). Your current harddisk type must be a basic disk (not a 'dynamic disk' or a 'GPT disk').

For example, in the figure above, Drive X is for Ubuntu OS and Y is for swap area (optional but strongly recommended unless you don't need virtual memory in Ubuntu).
Take note that the Ubuntu OS better to be installed on a primary partition, swap can be either on primary or extended partition.

4. The two Ubuntu partitions must be unformatted by Windows7, if formatted, delete and recreate; moreover, you MUST assign letters to the two partitions but don't use them in Windows7. In addition, all other partitions must also have be assigned a letter, including the system reserved partition M in my case.

5. Create the disk image. In command prompt, goto virtual box folder,
C:\Program Files\Oracle\VirtualBox in my case, run this:
VBoxManage internalcommands createrawvmdk -filename C:\rawdisk0.vmdk -rawdisk "\\.\PhysicalDrive0" -partitions 0,1,2,3,4,5,6
The image file 'disk0.vmdk' can be put anywhere you want with folder names not containing any space (probably a bug for that version of VirtualBox). The actual number of partitions "0,1,2,3,4,5,6" depends on your harddisk. There must be no errors in running this command, if there's an error, delete the possibly wrongly created disk0.vmdk and disk0-pt.vmdk file and change parameters and retry.

6. Create a new virtual machine in VirtualBox, use this C:\rawdisk0.vmdk as the disk and load Ubuntu ISO image into CD-DVD drive as shown below.

7. Launch the newly created virtual machine, press F12 to boot into Ubuntu CD, perform normal install of Ubuntu, make sure you install into the right partition and set correct swap area. For the boot loader I install into the Ubuntu partition. Try not to install the boot loader into MBR (master boot record) because Windows might not allow you to write MBR directly; also, you may want to remove Ubuntu and free up the 2 partitions at some later time.

Regardless of whether the boot-loader has been installed successfully or not, use EasyBCD to add a new entry and setting the drive to be the Ubuntu partition (i.e. X in my case) as shown below (it will be changed back to C by EasyBCD for some unknown reason):

8. The installation eventually complete successfully and the resultant copy of Ubuntu 64-bit can be run both inside Windows7's virtualBox and upon machine reboot.

Unfortunately, you cannot run the same copy of Windows7 in Linux's virtualBox because the VirtualBox's environment has incurred too many hardware changes that will require another windows activation.

Warning: DO NEVER reboot into the same copy of Windows7 from the Windows7's virtualBox although you can do so in the boot option shown above. Doing so may result in file system corruption or rendering all existing operating systems unbootable.
To prevent doing so, I strongly encourage you to set default system to Ubuntu in the boot option.

The good thing is that the Ubuntu in virtualBox is able to access windows partitions in read-only mode or full-access mode if they are unmounted (in Windows XP or earlier, you can have full access to any partition even when they are mounted; therefore, making changes to that partition from both inside and outside virtual machine causes file system corruption), to unmount a volume E:, run in command prompt:
mountvol
mountvol E:\ /D
after that, mount it in Ubuntu, make some changes to the partition, un-mount it in Ubuntu and re-mount in Windows 7 as follows:
mountvol E:\ VolumeName
Therefore, you need to note down the VolumeName for E:\ when running mountvol for the 1st time.

Here are some of my food-for-thoughts:
1. Although the new restriction in Vista/Win7 forbids full access to \\.\PhysicalDrive0, it does allow full access to individual mounted but unlocked raw partitions, this installation procedure does not violate this rule as long as you install grub loader into the partition boot record rather than MBR

2. Win7 has some harddisk I/O flush delay, updating MBR or partition boot sector takes a long time (say 20 minutes) unless you reboot Win7. So after installation, you might need to reboot your system in order to enter Ubuntu from VirtualBox.

3. In Win7, assigning a drive letter to a partition causes it to be mounted. Every time you make changes to your harddisk, i.e. mount/unmount some partitions, you might need to delete the entire VM in virtualBox manager together with the .vmdk files, and recreate them. Otherwise, it might fail.

I am writing this blog to notify all Linux/Ubuntu OS developers that it's possible for the same copy of OS to run in multiple different environments, e.g. inside VirtualBox or standalone (upon machine reboot). Therefore, OS developer should be aware of certain features and system attributes which might change because of the change in the execution environment.