HTML detector and switcher

HTML messages are detected and if present, the top of the message
textedit displays a clickable area that switches HTML rendering on and
off.

Fixes #13

src/bitmessageqt/__init__.py

@@ -35,6 +35,7 @@ from addaddressdialog import *
 from newsubscriptiondialog import *
 from regenerateaddresses import *
 from newchandialog import *
+from safehtmlparser import *
 from specialaddressbehavior import *
 from emailgateway import *
 from settings import *
@@ -4026,10 +4027,9 @@ class MyForm(settingsmixin.SMainWindow):
             data = self.getCurrentMessageId()
             if data != False:
                 message = "Error occurred: could not load message from disk."
-        message = unicode(message, 'utf-8)')
         messageTextedit.setCurrentFont(QtGui.QFont())
         messageTextedit.setTextColor(QtGui.QColor())
-        messageTextedit.setPlainText(message)
+        messageTextedit.setContent(message)
 
     def tableWidgetAddressBookItemChanged(self):
         currentRow = self.ui.tableWidgetAddressBook.currentRow()

src/bitmessageqt/bitmessageui.py

@@ -8,6 +8,7 @@
 # WARNING! All changes made in this file will be lost!
 
 from PyQt4 import QtCore, QtGui
+from messageview import MessageView
 import settingsmixin
 
 try:
@@ -123,7 +124,7 @@ class Ui_MainWindow(object):
         self.tableWidgetInbox.verticalHeader().setVisible(False)
         self.tableWidgetInbox.verticalHeader().setDefaultSectionSize(26)
         self.verticalSplitter_7.addWidget(self.tableWidgetInbox)
-        self.textEditInboxMessage = QtGui.QTextEdit(self.inbox)
+        self.textEditInboxMessage = MessageView(self.inbox)
         self.textEditInboxMessage.setBaseSize(QtCore.QSize(0, 500))
         self.textEditInboxMessage.setReadOnly(True)
         self.textEditInboxMessage.setObjectName(_fromUtf8("textEditInboxMessage"))
@@ -428,7 +429,7 @@ class Ui_MainWindow(object):
         self.tableWidgetInboxSubscriptions.verticalHeader().setVisible(False)
         self.tableWidgetInboxSubscriptions.verticalHeader().setDefaultSectionSize(26)
         self.verticalSplitter_4.addWidget(self.tableWidgetInboxSubscriptions)
-        self.textEditInboxMessageSubscriptions = QtGui.QTextEdit(self.subscriptions)
+        self.textEditInboxMessageSubscriptions = MessageView(self.subscriptions)
         self.textEditInboxMessageSubscriptions.setBaseSize(QtCore.QSize(0, 500))
         self.textEditInboxMessageSubscriptions.setReadOnly(True)
         self.textEditInboxMessageSubscriptions.setObjectName(_fromUtf8("textEditInboxMessageSubscriptions"))
@@ -527,7 +528,7 @@ class Ui_MainWindow(object):
         self.tableWidgetInboxChans.verticalHeader().setVisible(False)
         self.tableWidgetInboxChans.verticalHeader().setDefaultSectionSize(26)
         self.verticalSplitter_8.addWidget(self.tableWidgetInboxChans)
-        self.textEditInboxMessageChans = QtGui.QTextEdit(self.chans)
+        self.textEditInboxMessageChans = MessageView(self.chans)
         self.textEditInboxMessageChans.setBaseSize(QtCore.QSize(0, 500))
         self.textEditInboxMessageChans.setReadOnly(True)
         self.textEditInboxMessageChans.setObjectName(_fromUtf8("textEditInboxMessageChans"))

src/bitmessageqt/messageview.py

@@ -0,0 +1,44 @@
+from PyQt4 import QtCore, QtGui
+
+from safehtmlparser import *
+
+class MessageView(QtGui.QTextEdit):
+    MODE_PLAIN = 0
+    MODE_HTML = 1
+    TEXT_PLAIN = "HTML detected, click here to display"
+    TEXT_HTML = "Click here to disable HTML"
+    
+    def __init__(self, parent = 0):
+        super(MessageView, self).__init__(parent)
+        self.mode = MessageView.MODE_PLAIN 
+        self.html = None
+    
+    def mousePressEvent(self, event):
+        #text = textCursor.block().text()
+        if event.button() == QtCore.Qt.LeftButton and self.html.has_html and self.cursorForPosition(event.pos()).block().blockNumber() == 0:
+            if self.mode == MessageView.MODE_PLAIN:
+                self.showHTML()
+            else:
+                self.showPlain()
+        else:
+            super(MessageView, self).mousePressEvent(event)
+
+    def showPlain(self):
+        self.mode = MessageView.MODE_PLAIN
+        out = self.html.raw
+        if self.html.has_html:
+            out = "<div align=\"center\" style=\"text-decoration: underline;\"><b>" + QtGui.QApplication.translate("MessageView", MessageView.TEXT_PLAIN) + "</b></div><br/>" + out
+        self.setHtml(QtCore.QString(out))    
+
+    def showHTML(self):
+        self.mode = MessageView.MODE_HTML
+        out = self.html.sanitised
+        out = "<div align=\"center\" style=\"text-decoration: underline;\"><b>" + QtGui.QApplication.translate("MessageView", MessageView.TEXT_HTML) + "</b></div><br/>" + out
+        self.setHtml(QtCore.QString(out))
+
+    def setContent(self, data):
+        self.html = SafeHTMLParser()
+        self.html.allow_picture = True
+        self.html.feed(data)
+        self.html.close()
+        self.showPlain()

src/bitmessageqt/safehtmlparser.py

@@ -0,0 +1,97 @@
+from HTMLParser import HTMLParser
+import inspect
+from urllib import quote, quote_plus
+
+class SafeHTMLParser(HTMLParser):
+    # from html5lib.sanitiser
+    acceptable_elements = ['a', 'abbr', 'acronym', 'address', 'area',
+                           'article', 'aside', 'audio', 'b', 'big', 'blockquote', 'br', 'button',
+                           'canvas', 'caption', 'center', 'cite', 'code', 'col', 'colgroup',
+                           'command', 'datagrid', 'datalist', 'dd', 'del', 'details', 'dfn',
+                           'dialog', 'dir', 'div', 'dl', 'dt', 'em', 'event-source', 'fieldset',
+                           'figcaption', 'figure', 'footer', 'font', 'header', 'h1',
+                           'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'i', 'img', 'ins',
+                           'keygen', 'kbd', 'label', 'legend', 'li', 'm', 'map', 'menu', 'meter',
+                           'multicol', 'nav', 'nextid', 'ol', 'output', 'optgroup', 'option',
+                           'p', 'pre', 'progress', 'q', 's', 'samp', 'section', 'select',
+                           'small', 'sound', 'source', 'spacer', 'span', 'strike', 'strong',
+                           'sub', 'sup', 'table', 'tbody', 'td', 'textarea', 'time', 'tfoot',
+                           'th', 'thead', 'tr', 'tt', 'u', 'ul', 'var', 'video']
+    def __init__(self, *args, **kwargs):
+        HTMLParser.__init__(self, *args, **kwargs)
+        self.elements = set()
+        self.sanitised = u""
+        self.raw = u""
+        self.has_html = False
+        
+    def reset_safe(self):
+        self.raw = u""
+        self.sanitised = u""
+
+    def add_if_acceptable(self, tag, attrs = None):
+        if not tag in self.acceptable_elements:
+            return
+        self.sanitised += "<"
+        if inspect.stack()[1][3] == "handle_endtag":
+            self.sanitised += "/"
+        self.sanitised += tag
+        if attrs is not None:
+            for attr in attrs:
+                if tag == "img" and attr[0] == "src" and not self.allow_picture:
+                    attr[1] = ""
+                self.sanitised += " " + quote_plus(attr[0]) + "=\"" + attr[1] + "\""
+        if inspect.stack()[1][3] == "handle_startendtag":
+            self.sanitised += "/"
+        self.sanitised += ">"
+    
+    def add_raw(self, tag, attrs = None):
+        self.raw += "&lt;"
+        if inspect.stack()[1][3] == "handle_endtag":
+            self.raw += "/"
+        self.raw += tag
+        if attrs is not None:
+            for attr in attrs:
+                if tag == "img" and attr[0] == "src" and not self.allow_picture:
+                    attr[1] = ""
+                self.raw += " " + attr[0] + "=&quot;" + attr[1] + "&quot;"
+        if inspect.stack()[1][3] == "handle_startendtag":
+            self.raw += "/"
+        self.raw += "&gt;"
+        
+    def handle_starttag(self, tag, attrs):
+        if tag in self.acceptable_elements:
+            self.has_html = True
+        self.add_if_acceptable(tag, attrs)
+        self.add_raw(tag, attrs)
+
+    def handle_endtag(self, tag):
+        self.add_if_acceptable(tag)
+        self.add_raw(tag)
+        
+    def handle_startendtag(self, tag, attrs):
+        if tag in self.acceptable_elements:
+            self.has_html = True
+        self.add_if_acceptable(tag, attrs)
+        self.add_raw(tag, attrs)
+    
+    def handle_data(self, data):
+        self.sanitised += unicode(data, 'utf-8', 'replace')
+        tmp = data.replace("\n", "<br/>")
+        self.raw += unicode(tmp, 'utf-8', 'replace')
+        
+    def handle_charref(self, name):
+        self.sanitised += "&#" + name + ";"
+        self.raw += quote("&#" + name + ";")
+    
+    def handle_entityref(self, name):
+        self.sanitised += "&" + name + ";"
+        self.raw += quote("&" + name + ";")
+        
+    def is_html(self, text = None, allow_picture = False):
+        if text:
+            self.reset()
+            self.reset_safe()
+            self.feed(text)
+            self.close()
+            self.allow_picture = allow_picture
+        return self.has_html