Pull from trunk

src/calibre/ebooks/lrf/lrfparser.py

@@ -5,16 +5,16 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 import sys, array, os, re, codecs, logging
 
 from calibre import setup_cli_handlers, sanitize_file_name
-from calibre.utils.config import OptionParser 
+from calibre.utils.config import OptionParser
 from calibre.ebooks.lrf.meta import LRFMetaFile
 from calibre.ebooks.lrf.objects import get_object, PageTree, StyleObject, \
                                          Font, Text, TOCObject, BookAttr, ruby_tags
-                                         
+
 
 class LRFDocument(LRFMetaFile):
-    
+
     class temp(object): pass
-    
+
     def __init__(self, stream):
         LRFMetaFile.__init__(self, stream)
         self.scramble_key = self.xor_key
@@ -23,11 +23,11 @@ class LRFDocument(LRFMetaFile):
         self.image_map = {}
         self.toc = ''
         self.keep_parsing = True
-        
+
     def parse(self):
         self._parse_objects()
         self.metadata = LRFDocument.temp()
-        for a in ('title', 'title_reading', 'author', 'author_reading', 'book_id', 
+        for a in ('title', 'title_reading', 'author', 'author_reading', 'book_id',
                   'classification', 'free_text', 'publisher', 'label', 'category'):
             setattr(self.metadata, a, getattr(self, a))
         self.doc_info = LRFDocument.temp()
@@ -37,7 +37,7 @@ class LRFDocument(LRFMetaFile):
         self.device_info = LRFDocument.temp()
         for a in ('dpi', 'width', 'height'):
             setattr(self.device_info, a, getattr(self, a))
-        
+
     def _parse_objects(self):
         self.objects = {}
         self._file.seek(self.object_index_offset)
@@ -68,15 +68,15 @@ class LRFDocument(LRFMetaFile):
                 attr = h[0]
                 if hasattr(obj, attr):
                     self.ruby_tags[attr] = getattr(obj, attr)
-    
+
     def __iter__(self):
         for pt in self.page_trees:
             yield pt
-        
+
     def write_files(self):
         for obj in self.image_map.values() + self.font_map.values():
-            open(obj.file, 'wb').write(obj.stream)            
+            open(obj.file, 'wb').write(obj.stream)
-        
+
     def to_xml(self, write_files=True):
         bookinfo = u'<BookInformation>\n<Info version="1.1">\n<BookInfo>\n'
         bookinfo += u'<Title reading="%s">%s</Title>\n'%(self.metadata.title_reading, self.metadata.title)
@@ -113,7 +113,7 @@ class LRFDocument(LRFMetaFile):
                 pages += unicode(page)
             pages += close
         traversed_objects = [int(i) for i in re.findall(r'objid="(\w+)"', pages)] + [pt_id]
-        
+
         objects = u'\n<Objects>\n'
         styles  = u'\n<Style>\n'
         for obj in self.objects:
@@ -131,16 +131,16 @@ class LRFDocument(LRFMetaFile):
         if write_files:
             self.write_files()
         return '<BBeBXylog version="1.0">\n' + bookinfo + pages + styles + objects + '</BBeBXylog>'
-        
+
 def option_parser():
     parser = OptionParser(usage=_('%prog book.lrf\nConvert an LRF file into an LRS (XML UTF-8 encoded) file'))
     parser.add_option('--output', '-o', default=None, help=_('Output LRS file'), dest='out')
-    parser.add_option('--dont-output-resources', default=True, action='store_false', 
+    parser.add_option('--dont-output-resources', default=True, action='store_false',
-                      help=_('Do not save embedded image and font files to disk'), 
+                      help=_('Do not save embedded image and font files to disk'),
                       dest='output_resources')
     parser.add_option('--verbose', default=False, action='store_true', dest='verbose')
     return parser
-    
+
 def main(args=sys.argv, logger=None):
     parser = option_parser()
     opts, args = parser.parse_args(args)

src/calibre/ebooks/lrf/objects.py

@@ -970,7 +970,12 @@ class Canvas(LRFStream):
         stream = cStringIO.StringIO(self.stream)
         while stream.tell() < len(self.stream):
             tag = Tag(stream)
-            self._contents.append(PutObj(self._document.objects, *struct.unpack("<HHI", tag.contents)))
+            try:
+                self._contents.append(
+                    PutObj(self._document.objects,
+                        *struct.unpack("<HHI", tag.contents)))
+            except struct.error:
+                print 'Canvas object has errors, skipping.'
 
     def __unicode__(self):
         s = '\n<%s objid="%s" '%(self.__class__.__name__, self.id,)

src/calibre/ebooks/mobi/reader.py

@@ -313,8 +313,10 @@ class MobiReader(object):
             self.read_embedded_metadata(root, metadata_elems[0], guide)
         for elem in guides + metadata_elems:
             elem.getparent().remove(elem)
+        fname = self.name.encode('ascii', 'replace')
+        fname = re.sub(r'[\x08\x15\0]+', '', fname)
         htmlfile = os.path.join(output_dir,
-                                sanitize_file_name(self.name)+'.html')
+                                sanitize_file_name(fname)+'.html')
         try:
             for ref in guide.xpath('descendant::reference'):
                 if ref.attrib.has_key('href'):
@@ -396,8 +398,8 @@ class MobiReader(object):
                     'xx-large' : '6',
                     }
         mobi_version = self.book_header.mobi_version
-        style_map = {}
         for i, tag in enumerate(root.iter(etree.Element)):
+            tag.attrib.pop('xmlns', '')
             if tag.tag in ('country-region', 'place', 'placetype', 'placename',
                            'state', 'city', 'street', 'address', 'content'):
                 tag.tag = 'div' if tag.tag == 'content' else 'span'

src/calibre/utils/smtp.py

@@ -81,7 +81,12 @@ def sendmail(msg, from_, to, localhost=None, verbose=0, timeout=30,
         for x in to:
             return sendmail_direct(from_, x, msg, timeout, localhost, verbose)
     import smtplib
-    cls = smtplib.SMTP if encryption == 'TLS' else smtplib.SMTP_SSL
+    class SMTP_SSL(smtplib.SMTP_SSL): # Workaround for bug in smtplib.py
+        def _get_socket(self, host, port, timeout):
+            smtplib.SMTP_SSL._get_socket(self, host, port, timeout)
+            return self.sock
+
+    cls = smtplib.SMTP if encryption == 'TLS' else SMTP_SSL
     timeout = None # Non-blocking sockets sometimes don't work
     port = int(port)
     s = cls(timeout=timeout, local_hostname=localhost)
@@ -93,6 +98,8 @@ def sendmail(msg, from_, to, localhost=None, verbose=0, timeout=30,
         s.starttls()
         s.ehlo()
     if username is not None and password is not None:
+        if encryption == 'SSL':
+            s.sock = s.file.sslobj
         s.login(username, password)
     s.sendmail(from_, to, msg)
     return s.quit()

src/calibre/utils/terminfo.py

@@ -7,22 +7,22 @@ import sys, re, os
 class TerminalController:
     """
     A class that can be used to portably generate formatted output to
-    a terminal.  
+    a terminal.
-    
+
     `TerminalController` defines a set of instance variables whose
     values are initialized to the control sequence necessary to
     perform a given action.  These can be simply included in normal
     output to the terminal:
-    
+
     >>> term = TerminalController()
     >>> print 'This is '+term.GREEN+'green'+term.NORMAL
-    
+
     Alternatively, the `render()` method can used, which replaces
     '${action}' with the string required to perform 'action':
-    
+
     >>> term = TerminalController()
     >>> print term.render('This is ${GREEN}green${NORMAL}')
-    
+
     If the terminal doesn't support a given action, then the value of
     the corresponding instance variable will be set to ''.  As a
     result, the above code will still work on terminals that do not
@@ -30,11 +30,11 @@ class TerminalController:
     Also, this means that you can test whether the terminal supports a
     given action by simply testing the truth value of the
     corresponding instance variable:
-    
+
     >>> term = TerminalController()
     >>> if term.CLEAR_SCREEN:
     ...     print 'This terminal supports clearing the screen.'
-    
+
     Finally, if the width and height of the terminal are known, then
     they will be stored in the `COLS` and `LINES` attributes.
     """
@@ -44,35 +44,35 @@ class TerminalController:
     DOWN = ''            #: Move the cursor down one line
     LEFT = ''            #: Move the cursor left one char
     RIGHT = ''           #: Move the cursor right one char
-    
+
     # Deletion:
     CLEAR_SCREEN = ''    #: Clear the screen and move to home position
     CLEAR_EOL = ''       #: Clear to the end of the line.
     CLEAR_BOL = ''       #: Clear to the beginning of the line.
     CLEAR_EOS = ''       #: Clear to the end of the screen
-    
+
     # Output modes:
     BOLD = ''            #: Turn on bold mode
     BLINK = ''           #: Turn on blink mode
     DIM = ''             #: Turn on half-bright mode
     REVERSE = ''         #: Turn on reverse-video mode
     NORMAL = ''          #: Turn off all modes
-    
+
     # Cursor display:
     HIDE_CURSOR = ''     #: Make the cursor invisible
     SHOW_CURSOR = ''     #: Make the cursor visible
-    
+
     # Terminal size:
     COLS = None          #: Width of the terminal (None for unknown)
     LINES = None         #: Height of the terminal (None for unknown)
-    
+
     # Foreground colors:
     BLACK = BLUE = GREEN = CYAN = RED = MAGENTA = YELLOW = WHITE = ''
-    
+
     # Background colors:
     BG_BLACK = BG_BLUE = BG_GREEN = BG_CYAN = ''
     BG_RED = BG_MAGENTA = BG_YELLOW = BG_WHITE = ''
-    
+
     _STRING_CAPABILITIES = """
     BOL=cr UP=cuu1 DOWN=cud1 LEFT=cub1 RIGHT=cuf1
     CLEAR_SCREEN=clear CLEAR_EOL=el CLEAR_BOL=el1 CLEAR_EOS=ed BOLD=bold
@@ -80,7 +80,7 @@ class TerminalController:
     HIDE_CURSOR=cinvis SHOW_CURSOR=cnorm""".split()
     _COLORS = """BLACK BLUE GREEN CYAN RED MAGENTA YELLOW WHITE""".split()
     _ANSICOLORS = "BLACK RED GREEN YELLOW BLUE MAGENTA CYAN WHITE".split()
-    
+
     def __init__(self, term_stream=sys.stdout):
         """
         Create a `TerminalController` and initialize its attributes
@@ -92,24 +92,24 @@ class TerminalController:
         # Curses isn't available on all platforms
         try: import curses
         except: return
-        
+
         # If the stream isn't a tty, then assume it has no capabilities.
         if os.environ.get('CALIBRE_WORKER', None) is not None or not hasattr(term_stream, 'isatty') or not term_stream.isatty(): return
-        
+
         # Check the terminal type.  If we fail, then assume that the
         # terminal has no capabilities.
         try: curses.setupterm()
         except: return
-        
+
         # Look up numeric capabilities.
         self.COLS = curses.tigetnum('cols')
         self.LINES = curses.tigetnum('lines')
-        
+
         # Look up string capabilities.
         for capability in self._STRING_CAPABILITIES:
             (attrib, cap_name) = capability.split('=')
             setattr(self, attrib, self._tigetstr(cap_name) or '')
-        
+
         # Colors
         set_fg = self._tigetstr('setf')
         if set_fg:
@@ -127,7 +127,7 @@ class TerminalController:
         if set_bg_ansi:
             for i,color in zip(range(len(self._ANSICOLORS)), self._ANSICOLORS):
                 setattr(self, 'BG_'+color, curses.tparm(set_bg_ansi, i) or '')
-    
+
     def _tigetstr(self, cap_name):
         # String capabilities can include "delays" of the form "$<2>".
         # For any modern terminal, we should be able to just ignore
@@ -135,7 +135,7 @@ class TerminalController:
         import curses
         cap = curses.tigetstr(cap_name) or ''
         return re.sub(r'\$<\d+>[/*]?', '', cap)
-    
+
     def render(self, template):
         """
         Replace each $-substitutions in the given template string with
@@ -143,7 +143,7 @@ class TerminalController:
         '' (if it's not).
         """
         return re.sub(r'\$\$|\${\w+}', self._render_sub, template)
-    
+
     def _render_sub(self, match):
         s = match.group()
         if s == '$$': return s
@@ -156,20 +156,20 @@ class TerminalController:
 class ProgressBar:
     """
     A 3-line progress bar, which looks like::
-    
+
     Header
     20% [===========----------------------------------]
     progress message
-    
+
     The progress bar is colored, if the terminal supports color
     output; and adjusts to the width of the terminal.
-    
+
     If the terminal doesn't have the required capabilities, it uses a
     simple progress bar.
     """
     BAR = '%3d%% ${GREEN}[${BOLD}%s%s${NORMAL}${GREEN}]${NORMAL}\n'
     HEADER = '${BOLD}${CYAN}%s${NORMAL}\n\n'
-    
+
     def __init__(self, term, header, no_progress_bar = False):
         self.term, self.no_progress_bar = term, no_progress_bar
         self.fancy = self.term.CLEAR_EOL and self.term.UP and self.term.BOL
@@ -177,12 +177,14 @@ class ProgressBar:
             self.width = self.term.COLS or 75
             self.bar = term.render(self.BAR)
             self.header = self.term.render(self.HEADER % header.center(self.width))
+            if isinstance(self.header, unicode):
+                self.header = self.header.encode('utf-8')
             self.cleared = 1 #: true if we haven't drawn the bar yet.
-        
+
     def update(self, percent, message=''):
         if isinstance(message, unicode):
             message = message.encode('utf-8', 'replace')
-        
+
         if self.no_progress_bar:
             if message:
                 print message
@@ -203,8 +205,8 @@ class ProgressBar:
             else:
                 print '%d%%'%(percent*100), message
             sys.stdout.flush()
-            
+
-    
+
     def clear(self):
         if self.fancy and not self.cleared:
             sys.stdout.write(self.term.BOL + self.term.CLEAR_EOL +