Fix decoding gzip

This is how urllib3 does it.
vfaronov · Feb 1, 2018 · 1456cda · 1456cda
1 parent cc6a3b5
commit 1456cda
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/CHANGELOG.rst b/CHANGELOG.rst
@@ -15,6 +15,7 @@ Added
 
 Fixed
 -----
+- ``gzip``-encoded payloads are now decompressed more reliably.
 - When `analyzing TCP streams`_, HTTPolice now uses a stricter heuristic
   for detecting HTTP/1.x streams, producing fewer spurious `1006`_/`1009`_
   notices.

diff --git a/httpolice/codings.py b/httpolice/codings.py
@@ -2,15 +2,14 @@
 
 """Decoding content and transfer codings."""
 
-import gzip
-import io
 import zlib
 
 import brotli
 
 
 def decode_gzip(data):
-    return gzip.GzipFile(fileobj=io.BytesIO(data)).read()
+    # Just ``decompress(data, 16 + zlib.MAX_WBITS)`` doesn't work.
+    return zlib.decompressobj(16 + zlib.MAX_WBITS).decompress(data)
 
 
 def decode_deflate(data):