Address challenge feedback from Discord

Author: zardus

challenges/computing-101/building-a-web-server/common/run.py

@@ -13,6 +13,7 @@
 import signal
 import fcntl
 import ctypes
+import time
 import tempfile
 import pathlib
 import contextlib
@@ -32,7 +33,7 @@
 3..10   bind(3, {sa_family=AF_INET, sin_port=htons(<bind_port>), sin_addr=inet_addr("<bind_address>")}, 16) = 0
 4..10   listen(3, 0) = 0
 5..10   accept(3, NULL, NULL) = 4
-6..8    read(4, <read_request>, <read_request_count>) = <read_request_result>
+7..8    read(4, <read_request>, <read_request_count>) = <read_request_result>
 7..8    open("<open_path>", O_RDONLY) = 5
 7..8    read(5, <read_file>, <read_file_count>) = <read_file_result>
 7..8    close(5) = 0
@@ -354,12 +355,31 @@ def connect():
         pass
 
 
-def validate_connect():
-    session = retry_session()
-    try:
-        session.get("http://localhost", timeout=1)
-    except requests.exceptions.RequestException as e:
-        return request_failure("Connect", e)
+def validate_static_response():
+    expected = b"HTTP/1.0 200 OK\r\n\r\n"
+
+    for attempt in range(5):
+        try:
+            client = socket.create_connection(("localhost", 80), timeout=1)
+            break
+        except ConnectionRefusedError as e:
+            if attempt == 4:
+                return f"Static response: Failed to connect ({type(e).__name__}: {e})"
+            time.sleep(0.1)
+
+    with client:
+        response = b""
+        try:
+            while len(response) < len(expected):
+                chunk = client.recv(len(expected) - len(response))
+                if not chunk:
+                    break
+                response += chunk
+        except TimeoutError as e:
+            return f"Static response: Timed out ({type(e).__name__}: {e})"
+
+    if response != expected:
+        return "Static response: Response not correct"
 
 
 def validate_get(data=None):
@@ -401,7 +421,7 @@ def challenge():
         3: "bind an address to a socket",
         4: "listen on a socket",
         5: "accept a connection",
-        6: "respond to an http request",
+        6: "send a static HTTP response to a client",
         7: "respond to a GET request for the contents of a specified file",
         8: "accept multiple requests",
         9: "concurrently accept multiple requests",
@@ -451,14 +471,14 @@ def target():
 
         operation_names = {
             connect: "connect",
-            validate_connect: "validated connect",
+            validate_static_response: "static HTTP response",
             validate_get: "HTTP GET request",
             validate_post: "HTTP POST request",
         }
 
         operations = {
             5: [connect],
-            6: [validate_connect],
+            6: [validate_static_response],
             7: [validate_get],
             8: [validate_get, connect],
             9: [validate_get, connect],

challenges/computing-101/building-a-web-server/static-response/tests_private/test_solve.sh

@@ -1,7 +1,6 @@
-The available space in `/home/hacker` in this container is a measly **1 gigabyte**.
-In this level you will clog up `/home/hacker` with so much junk that even a tiny 1 megabyte file can't be created.
-When this happens, your workspace becomes unusable.
-We'll practice inducing this in this challenge, and then expand on it a bit later.
+In this challenge, `/home/hacker` is backed by a small temporary filesystem.
+You will clog it with so much junk that even a tiny 1 megabyte file can't be created.
+Then you will clean up the junk and restore the available space.
 
 How to fill the disk?
 There are so many ways.
@@ -22,9 +21,9 @@ y
 hacker@dojo:~$
 ```
 
-The `yes` outputs `y` over and over forever.
+The `yes` command outputs `y` over and over forever.
 The typical usage is to automate confirmation prompts ("Are you sure you want to delete this file?") using piping, but we'll use it here to make a massive file full of "y" lines.
-Just redirect `yes` to a file in your home directory, and you'll fill your disk in a minute or two!
+Just redirect `yes` to a file in your home directory, and it will keep growing until the temporary filesystem fills up!
 
 This challenge forces you to fill the disk and then clean up.
 The process:
@@ -36,14 +35,8 @@ The process:
 
 ----
 **NOTE:**
-For this challenge, `/home/hacker` is a small, temporary home directory.
 Your existing home-directory files will not be available here, and files you save here will not persist after the challenge ends.
 
 **Why two stages?**
-Your home directory persists across challenge instances.
-If we let you keep it full, your pwn.college will stop working.
-This is _by far_ the most common cause of weird issues on pwn.college!
-
-**HELP IT BROKE!**
-If you fill the disk and don't clean it up afterwards, you'll need to `ssh` in to fix things (by removing that file).
-This is a bit tricky, but we describe how to do it under "Connecting over SSH" in the [Getting Started](/welcome/welcome) module.
+Filling a filesystem is only half the exercise.
+The second stage teaches you to clean up after exhausting a system resource, even though this sandbox keeps the experiment away from your persistent home directory.

challenges/linux-luminarium/destruction/disk-doomsday/DESCRIPTION.md

@@ -31,3 +31,4 @@ Commands succeeded!
 
 Now, you try it!
 This process' `/challenge/pwn` must be piped into `/challenge/college`, but you'll need to intercept the data to see what `pwn` needs from you!
+Use `tee` to duplicate your first attempt into a file, inspect the saved output, and then retry the pipeline with what you learned.

challenges/linux-luminarium/piping/tee/DESCRIPTION.md

@@ -12,7 +12,7 @@ read SS < <(tail -n1)
 
 if [ "$SS" != "SUPERSECRET:$(tail -c+20 /flag | head -c4)" ]
 then
-	fold -s <<< "The input to 'college' does not contain the correct secret code! This code should be provided by the 'pwn' command. HINT: use 'tee' to intercept the output of 'pwn' and figure out what the code needs to be."
+	fold -s <<< "The input to 'college' does not contain the correct secret code! This code should be provided by the 'pwn' command. HINT: use 'tee' to save the first attempt to a file, inspect the saved output, and then retry the pipeline with what you learned."
 	exit 2
 fi