Improve url/searchterm checking.

qutebrowser/utils/__init__.py

@@ -19,10 +19,12 @@
 
 import re
 import sys
+import socket
 import os.path
 import platform
 import logging
 import subprocess
+import ipaddress
 import urllib.parse
 
 from PyQt5.QtCore import QUrl, QT_VERSION_STR, PYQT_VERSION_STR, qVersion
@@ -36,16 +38,24 @@ def qurl(url):
     if isinstance(url, QUrl):
         logging.debug("url is already a qurl")
         return url
-    elif '.' in url or is_about_url(url):  # probably an address
+    return QUrl.fromUserInput(url)
+
+
+def fuzzy_url(url):
+    """Returns a QUrl based on an user input which is URL or search term."""
+    u = url.toString() if isinstance(url, QUrl) else url
+    if is_url(u):
+        # probably an address
         logging.debug("url is a fuzzy address")
-        newurl = QUrl.fromUserInput(url)
+        newurl = QUrl.fromUserInput(u)
     else:  # probably a search term
         logging.debug("url is a fuzzy search term")
         try:
-            newurl = QUrl.fromUserInput(_get_search_url(url))
+            newurl = QUrl.fromUserInput(_get_search_url(u))
         except ValueError:
-            newurl = QUrl.fromUserInput(url)
+            newurl = QUrl.fromUserInput(u)
-    logging.debug('Converting {} to qurl -> {}'.format(url, newurl.url()))
+    logging.debug('Converting fuzzy term {} to url -> {}'.format(
+        u, newurl.url()))
     return newurl
 
 
@@ -104,9 +114,67 @@ def version():
 
 def is_about_url(url):
     """Return True if url is an about:... or other special URL."""
-    if isinstance(url, QUrl):
+    u = url.toString() if isinstance(url, QUrl) else url
-        url = url.toString()
+    return u.replace('http://', '').startswith('about:')
-    return url.replace('http://', '').startswith('about:')
+
+
+def is_url(url):
+    """Return True if url seems to be a valid URL."""
+    # FIXME Importing this here fixes some weird dependency problems.
+    import qutebrowser.utils.config as config
+    logging.debug('Checking if "{}" is an URL'.format(url))
+    if ' ' in url:
+        # An URL will never contain a space
+        logging.debug('Contains space -> no url')
+        return False
+    elif config.config.getboolean('general', 'addressbar_dns_lookup'):
+        logging.debug('Checking via DNS')
+        return _is_url_dns(url)
+    else:
+        logging.debug('Checking via naive check')
+        return _is_url_naive(url)
+
+
+def _is_url_naive(url):
+    """Naive check if given url string is really an url."""
+    PROTOCOLS = ['http://', 'https://']
+    ip = _get_netloc(url)
+    if not ip:
+        is_ip = False
+    else:
+        try:
+            ipaddress.ip_address(ip)
+        except ValueError:
+            is_ip = False
+        else:
+            is_ip = True
+    return (any([url.startswith(proto) for proto in PROTOCOLS]) or
+            '.' in url or is_about_url(url) or url == 'localhost'
+            or is_ip)
+
+
+def _is_url_dns(url):
+    """Check if an url string is really an url via DNS."""
+    # FIXME we could probably solve this in a nicer way by attempting to open
+    # the page in the webview, and then open the search if that fails.
+    netloc = _get_netloc(url)
+    if not netloc:
+        return False
+    try:
+        socket.gethostbyname(netloc)
+    except socket.gaierror:
+        return False
+    else:
+        return True
+
+
+def _get_netloc(url):
+    """Gets the host part of an url."""
+    # FIXME better way to do this?
+    if '://' in url:
+        return urllib.parse.urlsplit(url).netloc
+    else:
+        return urllib.parse.urlsplit('http://' + url).netloc
 
 
 def _git_str():

qutebrowser/utils/config.py

@@ -39,6 +39,7 @@ default_config = """
     ignorecase = true
     wrapsearch = true
     startpage = http://www.duckduckgo.com/
+    addressbar_dns_lookup = false
 
     [searchengines]
     duckduckgo = https://duckduckgo.com/?q={}

qutebrowser/widgets/browser.py

@@ -428,19 +428,19 @@ class BrowserTab(QWebView):
 
         url -- The URL to load, as string or QUrl.
         """
-        qurl = utils.qurl(url)
+        u = utils.fuzzy_url(url)
-        logging.debug('New title: {}'.format(qurl.url()))
+        logging.debug('New title: {}'.format(u.url()))
-        self.titleChanged.emit(qurl.url())
+        self.titleChanged.emit(u.url())
-        if utils.is_about_url(qurl):
+        if utils.is_about_url(u):
             try:
-                content = about.handle(qurl.toString())
+                content = about.handle(u.toString())
             except AttributeError:
-                return self.load(qurl)
+                return self.load(u)
             else:
-                self.setUrl(qurl)
+                self.setUrl(u)
                 self.setContent(content, 'text/html')
         else:
-            return self.load(qurl)
+            return self.load(u)
 
     def link_handler(self, url):
         """Handle a link.