Skip to content

Latest commit

 

History

History
20 lines (12 loc) · 891 Bytes

README.md

File metadata and controls

20 lines (12 loc) · 891 Bytes

Echo

  • Round: 30 (2023/01)
  • Category: Web
  • Points: 125
  • Solves: 12

Description

An echo service that employs microservices architecture!

Solution

Content-Length is the message length in str, but it should be bytes, so we can smuggle an extra request using Unicode characters. To force it to be handled correctly in gunicorn you need to align the request to chunk_size.

solve.py

But it turns out you don't need to align the second request to chunk_size as other people who solved this challenge told me. After some testing, it turns how it is gunicorn's worker type affecting this:

When I use gunicorn -k gevent --keep-alive 1 --bind 0.0.0.0:7777 api:app then there is no need to align anything, but if you are using gunicorn -k gthread -t 2 --keep-alive 1 --bind 0.0.0.0:7777 api:app then you must align it as it probably does what Chromium does.