From 8eae7772653daec53d6ad6ee98f0be9e8c9c5857 Mon Sep 17 00:00:00 2001
From: Arthur Bols <arthur@bols.dev>
Date: Sun, 28 Mar 2021 19:53:14 +0200
Subject: [PATCH] Fix issues

---
 client.py                 |  1 +
 client/command.py         | 27 +--------------------------
 client/responsehandler.py |  6 ++++--
 httplib/httpsocket.py     |  8 ++++++--
 httplib/parser.py         | 12 ++++++++++--
 5 files changed, 22 insertions(+), 32 deletions(-)

diff --git a/client.py b/client.py
index 9d19d7c..d0af628 100644
--- a/client.py
+++ b/client.py
@@ -16,6 +16,7 @@ def main():
 
     arguments = parser.parse_args()
 
+    # Setup logging
     logging.basicConfig(level=logging.INFO - (10 * arguments.verbose), format="[%(levelname)s] %(message)s")
     logging.debug("Arguments: %s", arguments)
 
diff --git a/client/command.py b/client/command.py
index 3774f52..caf0ac8 100644
--- a/client/command.py
+++ b/client/command.py
@@ -1,7 +1,6 @@
 import logging
 from abc import ABC, abstractmethod
 from typing import Dict
-from urllib.parse import urlparse
 
 from client.httpclient import HTTPClient
 from httplib import parser
@@ -21,7 +20,7 @@ def create(method: str, url: str, port):
     @param port: The port for the command
     """
 
-    uri = parser.get_uri(url)
+    uri = parser.uri_from_url(url)
     if method == "GET":
         return GetCommand(uri, port)
     elif method == "HEAD":
@@ -125,8 +124,6 @@ class AbstractCommand(ABC):
             if not sub_request:
                 client.close()
 
-
-
     def _get_preamble(self, client):
         """
         Returns the preamble (start-line and headers) of the response of this command.
@@ -157,28 +154,6 @@ class AbstractCommand(ABC):
     def _build_message(self, message: str) -> bytes:
         return (message + "\r\n").encode(FORMAT)
 
-    def parse_uri(self):
-        """
-        Parses the URI and returns the hostname and path.
-        @return: A tuple of the hostname and path.
-        """
-        parsed = urlparse(self.uri)
-
-        # If there is no netloc, the url is invalid, so prepend `//` and try again
-        if parsed.netloc == "":
-            parsed = urlparse("http://" + self.uri)
-
-        host = parsed.netloc
-        path = parsed.path
-        if len(path) == 0 or path[0] != '/':
-            path = "/" + path
-
-        port_pos = host.find(":")
-        if port_pos >= 0:
-            host = host[:port_pos]
-
-        return host, path
-
 
 class HeadCommand(AbstractCommand):
     """
diff --git a/client/responsehandler.py b/client/responsehandler.py
index 841a1de..6d87ccb 100644
--- a/client/responsehandler.py
+++ b/client/responsehandler.py
@@ -269,7 +269,9 @@ class HTMLDownloadHandler(DownloadHandler):
                     to_replace.append((url_start, url_end, local_path))
 
             except Exception as e:
-                logging.error("Failed to download image: %s, skipping...", target, exc_info=e)
+                logging.error("Failed to download image: %s, skipping...", target)
+                logging.debug("", exc_info=e)
+                processed[target] = None
 
         # reverse the list so urls at the bottom of the html file are processed first.
         # Otherwise, our start and end positions won't be correct.
@@ -277,7 +279,7 @@ class HTMLDownloadHandler(DownloadHandler):
         for (start, end, path) in to_replace:
             html = html[:start] + path + html[end:]
 
-        logging.info("Saving to HTML '%s'", parser.get_relative_save_path(target_path))
+        logging.info("Saving HTML to '%s'", parser.get_relative_save_path(target_path))
         with open(target_path, 'w', encoding=FORMAT) as file:
             file.write(html)
 
diff --git a/httplib/httpsocket.py b/httplib/httpsocket.py
index 25cd80f..ad6544a 100644
--- a/httplib/httpsocket.py
+++ b/httplib/httpsocket.py
@@ -82,11 +82,15 @@ class HTTPSocket:
 
 
 class HTTPException(Exception):
-    """ Base class for HTTP exceptions """
+    """
+    Base class for HTTP exceptions
+    """
 
 
 class InvalidResponse(HTTPException):
-    """ Response message cannot be parsed """
+    """
+    Response message cannot be parsed
+    """
 
     def __init(self, message):
         self.message = message
diff --git a/httplib/parser.py b/httplib/parser.py
index 781b7fd..12520ed 100644
--- a/httplib/parser.py
+++ b/httplib/parser.py
@@ -174,13 +174,21 @@ def parse_uri(uri: str):
     return host, port, path
 
 
-def get_uri(url: str):
+def uri_from_url(url: str):
     """
     Returns a valid URI of the specified URL.
     """
     parsed = urlsplit(url)
 
-    result = f"http://{parsed.netloc}{parsed.path}"
+    if parsed.hostname is None:
+        url = f"http://{url}"
+        parsed = urlsplit(url)
+
+    path = parsed.path
+    if path == "":
+        path = "/"
+
+    result = f"http://{parsed.netloc}{path}"
     if parsed.query != "":
         result = f"{result}?{parsed.query}"