Fix(Py3): `urllib` writes its files as bytes

[bookmarks_db.git] / Robots / bkmk_rurllib_py3.py

diff --git a/Robots/bkmk_rurllib_py3.py b/Robots/bkmk_rurllib_py3.py
index b39288770fa2212fe0cbfbaba010505c436faa1d..73f9b77ca3ef19f0106cc3bd3dc3e16902b5a818 100644 (file)
--- a/Robots/bkmk_rurllib_py3.py
+++ b/Robots/bkmk_rurllib_py3.py
@@ -11,6 +11,8 @@ __license__ = "GNU GPL"
 __all__ = ['robot_urllib_py3']
 
 
+import http.client
+import socket
 import sys
 import urllib.request
 
@@ -41,6 +43,8 @@ class MyURLopener(urllib.request.URLopener):
     http_error_303 = http_error_301
     # Error 307 -- relocated (temporarily)
     http_error_307 = http_error_301
+    # Error 308 -- relocated (permanently)
+    http_error_308 = http_error_301
 
     # Error 401 -- authentication required
     def http_error_401(self, url, fp, errcode, errmsg, headers, data=None):
@@ -53,24 +57,27 @@ class MyURLopener(urllib.request.URLopener):
             fp.close()
         raise IOError(('http error', errcode, errmsg, headers))
 
+    def open(self, fullurl, data=None, timeout=socket._GLOBAL_DEFAULT_TIMEOUT):
+        return urllib.request.URLopener.open(self, fullurl, data)
 
-urllib.request._opener = MyURLopener()
+
+urllib.request._opener = opener = MyURLopener()
 
 # Fake headers to pretend this is a real browser
 _user_agent = "Mozilla/5.0 (X11; U; Linux 2.6 i686; en)"
 " Gecko/20001221 Firefox/2.0.0"
-urllib.request._opener.addheaders[0] = ('User-Agent', _user_agent)
+opener.addheaders[0] = ('User-Agent', _user_agent)
 _x_user_agent = "bookmarks_db (Python %d.%d.%d; urllib/%s)" % (
    sys.version_info[0], sys.version_info[1],
    sys.version_info[2], urllib.request.__version__
 )
-urllib.request._opener.addheader('X-User-Agent', _x_user_agent)
-urllib.request._opener.addheader('Referer', '')
+opener.addheader('X-User-Agent', _x_user_agent)
+opener.addheader('Referer', '')
 
-urllib.request._opener.addheader('Accept', '*/*')
-urllib.request._opener.addheader('Accept-Language', 'ru,en')
-urllib.request._opener.addheader('Cache-Control', 'max-age=300')
-urllib.request._opener.addheader('Connection', 'close')
+opener.addheader('Accept', '*/*')
+opener.addheader('Accept-Language', 'ru,en')
+opener.addheader('Cache-Control', 'max-age=300')
+opener.addheader('Connection', 'close')
 
 
 urllib_ftpwrapper = urllib.request.ftpwrapper
@@ -91,26 +98,50 @@ class robot_urllib_py3(robot_base):
     def get(self, bookmark, url, accept_charset=False):
         try:
             # Set fake referer to the base URL
-            urllib.request._opener.addheaders[2] = ('Referer', url)
+            opener.addheaders[2] = ('Referer', url)
 
             if accept_charset and bookmark.charset:
-                urllib.request._opener.addheader('Accept-Charset', bookmark.charset)
+                opener.addheader('Accept-Charset', bookmark.charset)
             try:
                 fname, headers = urllib.request.urlretrieve(url)
             finally:
                 if accept_charset and bookmark.charset:
                     # Remove Accept-Charset
-                    del urllib.request._opener.addheaders[-1]
-
-            infile = open(fname, 'rt')
-            content = infile.read()
-            infile.close()
-
+                    del opener.addheaders[-1]
+
+            possible_encodings = []
+            for encoding in (
+                    bookmark.charset,
+                    sys.getfilesystemencoding(),
+                    'utf-8',
+            ):
+                if encoding and encoding not in possible_encodings:
+                    possible_encodings.append(encoding)
+            content = e = None
+            infile = open(fname, 'rb')
+            try:
+                content = infile.read()
+            except Exception as e:
+                content = None
+            finally:
+                infile.close()
+
+            if content is None:
+                e = str(e)
+                return (
+                    'ERROR: ' + e,
+                    None, None, None, None
+                )
             return None, None, None, headers, content
 
         except RedirectException as e:
             return None, e.errcode, e.newurl, None, None
 
+        except (OSError, http.client.IncompleteRead) as e:
+            error = str(e)
+            self.log('   Error: %s' % error)
+            return error, None, None, None, None
+
         except IOError as e:
             if (e[0] == "http error") and (e[1] == -1):
                 error = None
@@ -125,7 +156,7 @@ class robot_urllib_py3(robot_base):
 
     def get_ftp_welcome(self):
         global ftpcache_key
-        _welcome = urllib.request._opener.ftpcache[ftpcache_key].ftp.welcome
+        _welcome = opener.ftpcache[ftpcache_key].ftp.welcome
         # I am assuming there are no duplicate ftp URLs in db.
         # If there are - ftpcache_key in next line is invalid.
         ftpcache_key = None
@@ -134,3 +165,4 @@ class robot_urllib_py3(robot_base):
     def finish_check_url(self, bookmark):
         robot_base.finish_check_url(self, bookmark)
         urllib.request.urlcleanup()
+        urllib.request._opener = opener