1 files changed, 13 insertions, 25 deletions
diff --git a/importpkg.py b/importpkg.py
index ce4a446..7bca70b 100755
--- a/importpkg.py
+++ b/importpkg.py
@@ -1,4 +1,4 @@
-#!/usr/bin/python
+#!/usr/bin/python3
 """This tool reads a Debian package from stdin and emits a yaml stream on
 stdout.  It does not access a database. Therefore it can be run in parallel and
 on multiple machines. The generated yaml contains multiple documents. The first
@@ -8,42 +8,33 @@ And finally a document consisting of the string "commit" is emitted."""
 import argparse
 import hashlib
 import sys
+import urllib.request
 import zlib
-try:
-    from urllib.request import urlopen
-except ImportError:
-    from urllib2 import urlopen
 
 import yaml
 
-from dedup.debpkg import DebExtractor, decodetarname, get_tar_hashes
+from dedup.debpkg import DebExtractor, get_tar_hashes
 from dedup.hashing import DecompressedHash, SuppressingHash, HashedStream, \
         HashBlacklistContent
 from dedup.compression import GzipDecompressor
 from dedup.image import GIFHash, PNGHash
 
-boring_content = set(("", "\n"))
+boring_content = set((b"", b"\n"))
 
 def sha512_nontrivial():
     return HashBlacklistContent(hashlib.sha512(), boring_content)
 
 def gziphash():
-    hashobj = DecompressedHash(GzipDecompressor(), hashlib.sha512())
+    hashobj = hashlib.sha512()
+    hashobj = DecompressedHash(GzipDecompressor(), hashobj, "gzip_sha512")
     hashobj = SuppressingHash(hashobj, (ValueError, zlib.error))
-    hashobj.name = "gzip_sha512"
     return HashBlacklistContent(hashobj, boring_content)
 
 def pnghash():
-    hashobj = PNGHash(hashlib.sha512())
-    hashobj = SuppressingHash(hashobj, (ValueError,))
-    hashobj.name = "png_sha512"
-    return hashobj
+    return SuppressingHash(PNGHash(hashlib.sha512()), (ValueError,))
 
 def gifhash():
-    hashobj = GIFHash(hashlib.sha512())
-    hashobj = SuppressingHash(hashobj, (ValueError,))
-    hashobj.name = "gif_sha512"
-    return hashobj
+    return SuppressingHash(GIFHash(hashlib.sha512()), (ValueError,))
 
 class ProcessingFinished(Exception):
     pass
@@ -63,7 +54,7 @@ class ImportpkgExtractor(DebExtractor):
         # deb822 currently returns :any dependencies raw. see #670679
         deprelations = info.relations.get("depends", []) + \
                        info.relations.get("pre-depends", [])
-        depends = set(dep[0]["name"].split(u':', 1)[0]
+        depends = set(dep[0]["name"].split(':', 1)[0]
                       for dep in deprelations if len(dep) == 1)
         self.callback(dict(package=info["package"], source=source,
                            version=info["version"],
@@ -73,22 +64,19 @@ class ImportpkgExtractor(DebExtractor):
         for name, size, hashes in get_tar_hashes(tarfileobj,
                                                  self.hash_functions):
             try:
-                name = decodetarname(name)
-            except UnicodeDecodeError:
+                name.encode("utf8", "strict")
+            except UnicodeEncodeError:
                 print("warning: skipping filename with encoding error")
                 continue # skip files with non-utf8 encoding for now
             self.callback(dict(name=name, size=size, hashes=hashes))
         raise ProcessingFinished()
 
 def main():
-    try:
-        stdin = sys.stdin.buffer
-    except AttributeError: # python2
-        stdin = sys.stdin
     parser = argparse.ArgumentParser()
     parser.add_argument("-H", "--hash", action="store",
                         help="verify that stdin hash given sha256 hash")
-    parser.add_argument("input", nargs='?', default=stdin, type=urlopen,
+    parser.add_argument("input", nargs='?', default=sys.stdin.buffer,
+                        type=urllib.request.urlopen,
                         help="read from this location instead of stdin")
     args = parser.parse_args()
     dumper = yaml.SafeDumper(sys.stdout)