diff --git a/qutebrowser/misc/mhtml.py b/qutebrowser/misc/mhtml.py
index 3afc32c02..795b102e6 100644
--- a/qutebrowser/misc/mhtml.py
+++ b/qutebrowser/misc/mhtml.py
@@ -20,11 +20,13 @@
 """Utils for writing a MHTML file."""
 
 import functools
+import quopri
 import io
 
 from collections import namedtuple
 from base64 import b64encode
 from urllib.parse import urljoin
+from uuid import uuid4
 
 from PyQt5.QtCore import QUrl
 from PyQt5.QtNetwork import QNetworkRequest, QNetworkReply
@@ -36,18 +38,44 @@ _File = namedtuple("_File",
                    "content content_type content_location transfer_encoding")
 
 
+def _chunked_base64(data, maxlen=76, linesep=b"\r\n"):
+    """Just like b64encode, except that it breaks long lines.
+
+    Args:
+        maxlen: Maximum length of a line, not including the line separator.
+        linesep: Line separator to use as bytes.
+    """
+    encoded = b64encode(data)
+    result = []
+    for i in range(0, len(encoded), maxlen):
+        result.append(encoded[i:i+maxlen])
+    return linesep.join(result)
+
+def _rn_quopri(data):
+    """Return a quoted-printable representation of data."""
+    orig_funcs = (quopri.b2a_qp, quopri.a2b_qp)
+    # Workaround for quopri mixing \n and \r\n
+    quopri.b2a_qp = quopri.a2b_qp = None
+    encoded = quopri.encodestring(data)
+    quopri.b2a_qp, quopri.a2b_qp = orig_funcs
+    return encoded.replace(b"\n", b"\r\n")
+
+
 E_NONE = (None, lambda x: x)
 """No transfer encoding, copy the bytes from input to output"""
 
-E_BASE64 = ("BASE64", b64encode)
+E_BASE64 = ("base64", _chunked_base64)
 """Encode the file using base64 encoding"""
 
+E_QUOPRI = ("quoted-printable", _rn_quopri)
+"""Encode the file using MIME quoted-printable encoding."""
+
 
 class MHTMLWriter(object):
     """A class for aggregating multiple files and outputting them to a MHTML
     file."""
 
-    BOUNDARY = b"qute-mhtml"
+    BOUNDARY = b"---qute-mhtml-" + str(uuid4()).encode("ascii")
 
     def __init__(self, root_content=None, content_location=None,
                  content_type=None):
@@ -90,6 +118,9 @@ class MHTMLWriter(object):
         self._output_root_file(fp)
         for file_data in self._files.values():
             self._output_file(fp, file_data)
+        fp.write(b"\r\n--")
+        fp.write(self.BOUNDARY)
+        fp.write(b"--")
 
     def _output_header(self, fp):
         if self.content_location is None:
@@ -99,34 +130,34 @@ class MHTMLWriter(object):
 
         fp.write(b"Content-Location: ")
         fp.write(self.content_location.encode("utf-8"))
-        fp.write(b'\nContent-Type: multipart/related;boundary="')
+        fp.write(b'\r\nContent-Type: multipart/related;boundary="')
         fp.write(self.BOUNDARY)
         fp.write(b'";type="')
         fp.write(self.content_type.encode("utf-8"))
-        fp.write(b'"\n\n')
+        fp.write(b'"\r\n\r\n')
 
     def _output_root_file(self, fp):
         root_file = _File(
             content=self.root_content, content_type=self.content_type,
-            content_location=self.content_location, transfer_encoding=E_BASE64
+            content_location=self.content_location, transfer_encoding=E_QUOPRI,
         )
         self._output_file(fp, root_file)
 
     def _output_file(self, fp, file_struct):
         fp.write(b"--")
         fp.write(self.BOUNDARY)
-        fp.write(b"\nContent-Location: ")
+        fp.write(b"\r\nContent-Location: ")
         fp.write(file_struct.content_location.encode("utf-8"))
         if file_struct.content_type is not None:
-            fp.write(b"\nContent-Type: ")
+            fp.write(b"\r\nContent-Type: ")
             fp.write(file_struct.content_type.encode("utf-8"))
         encoding_name, encoding_func = file_struct.transfer_encoding
         if encoding_name:
-            fp.write(b"\nContent-Transfer-Encoding: ")
+            fp.write(b"\r\nContent-Transfer-Encoding: ")
             fp.write(encoding_name.encode("utf-8"))
-        fp.write(b"\n\n")
+        fp.write(b"\r\n\r\n")
         fp.write(encoding_func(file_struct.content))
-        fp.write(b"\n\n")
+        fp.write(b"\r\n\r\n")
 
 
 def start_download(dest):
@@ -163,13 +194,15 @@ def start_download(dest):
         pending_downloads.remove(item)
         mime = item.raw_headers.get(b"Content-Type", b"")
         mime = mime.decode("ascii", "ignore")
-        writer.add_file(name, item.fileobj.getvalue(), mime)
+        encode = E_QUOPRI if mime.startswith("text/") else E_BASE64
+        writer.add_file(name, item.fileobj.getvalue(), mime, encode)
         if pending_downloads:
             return
         finish_file()
 
-    def error(item, *args):
+    def error(name, item, *args):
         pending_downloads.remove(item)
+        writer.add_file(name, b"")
         if pending_downloads:
             return
         finish_file()
@@ -190,7 +223,6 @@ def start_download(dest):
             # Might be a local <script> tag or something else
             continue
         absolute_url_str = urljoin(web_url_str, element_url)
-        name = absolute_url_str if element_url.startswith("//") else element_url
         # Prevent loading an asset twice
         if absolute_url_str in loaded_urls:
             continue
@@ -204,8 +236,8 @@ def start_download(dest):
                                     auto_remove=True)
         pending_downloads.add(item)
         item.finished.connect(
-            functools.partial(finished, name, item))
+            functools.partial(finished, absolute_url_str, item))
         item.error.connect(
-            functools.partial(error, item))
+            functools.partial(error, absolute_url_str, item))
         item.cancelled.connect(
-            functools.partial(error, item))
+            functools.partial(error, absolute_url_str, item))