Update

httplib/parser.py

@@ -1,6 +1,6 @@
 import logging
 import re
-from urllib.parse import urlparse
+from urllib.parse import urlparse, urlsplit
 
 from httplib.exceptions import InvalidStatusLine, InvalidResponse, BadRequest
 from httplib.httpsocket import HTTPSocket
@@ -42,6 +42,26 @@ def get_status_line(client: HTTPSocket):
     return version, status, reason
 
 
+def parse_status_line(line: str):
+    split = list(filter(None, line.strip().split(" ", 2)))
+    if len(split) < 3:
+        raise InvalidStatusLine(line)  # TODO fix exception
+
+    (http_version, status, reason) = split
+
+    if not _is_valid_http_version(http_version):
+        raise InvalidStatusLine(line)
+    version = http_version[:4]
+
+    if not re.match(r"\d{3}", status):
+        raise InvalidStatusLine(line)
+    status = int(status)
+    if status < 100 or status > 999:
+        raise InvalidStatusLine(line)
+
+    return version, status, reason
+
+
 def parse_request_line(client: HTTPSocket):
     line, (method, target, version) = _get_start_line(client)
 
@@ -119,7 +139,7 @@ def parse_request_headers(client: HTTPSocket):
                 raise BadRequest()
 
         headers[key] = value
-    
+
     return headers
 
 
@@ -157,6 +177,38 @@ def get_headers(client: HTTPSocket):
 
     return result
 
+def parse_headers(lines):
+    headers = []
+    # first header after the status-line may not contain a space
+    for line in lines:
+        line = next(lines)
+        if line[0].isspace():
+            continue
+        else:
+            break
+
+    for line in lines:
+        if line in ("\r\n", "\n", " "):
+            break
+
+        if line[0].isspace():
+            headers[-1] = headers[-1].rstrip("\r\n")
+
+        headers.append(line.lstrip())
+
+    result = {}
+    header_str = "".join(headers)
+    for line in header_str.splitlines():
+        pos = line.find(":")
+
+        if pos <= 0 or pos >= len(line) - 1:
+            continue
+
+        (header, value) = map(str.strip, line.split(":", 1))
+        check_next_header(result, header, value)
+        result[header.lower()] = value.lower()
+
+    return result
 
 def check_next_header(headers, next_header: str, next_value: str):
     if next_header == "content-length":
@@ -166,3 +218,25 @@ def check_next_header(headers, next_header: str, next_value: str):
         if not next_value.isnumeric() or int(next_value) <= 0:
             logging.error("Invalid content-length value: %r", next_value)
             raise InvalidResponse()
+
+
+def parse_uri(uri: str):
+    parsed = urlsplit(uri)
+
+    # If there is no netloc, the given string is not a valid URI, so split on /
+    if parsed.hostname:
+        host = parsed.hostname
+        path = parsed.path
+        if parsed.query != '':
+            path = f"{path}?{parsed.query}"
+    else:
+        (host, path) = uri.split("/", 1)
+
+    if ":" in host:
+        host, port = host.split(":", 1)
+    elif parsed.scheme == "https":
+        port = 443
+    else:
+        port = 80
+
+    return host, port, path