switch to md5; adjust format a bit

This commit is contained in:
Benjamin Peterson 2017-09-07 12:13:26 -07:00
parent 435f751ee5
commit d027ce144b
1 changed files with 6 additions and 8 deletions

View File

@ -59,13 +59,12 @@ Python will begin to recognize two magic number variants for every pyc
version. One magic number will correspond to the current pyc format and the
other to "hash-based" pycs introduced by this PEP.
In hash-based pycs, the second field in the pyc header (currently the
"timestamp" field) will contain the SipHash_ (with a hardcoded key) of the
contents of the source file. Another a fast hash like MD5 or BLAKE2_ would also
work. We choose SipHash because Python already has a builtin implementation of
it from :pep:`456`. The third field in the pyc header (currently the "source
size" field) will become a bitset of flags. We define the lowest flag in this
bitset called ``check_source``
In hash-based pycs, the second 32-bit field in the pyc header (currently the
"timestamp" field) will become a bitset of flags. We define the lowest flag in
this bitset called ``check_source``. This bitset field is followed by a 128-bit
MD5 hash field. We choose MD5 because it is fast and sufficiently collision
resistant for this application. No strong security properties are required
here. (A truncated BLAKE2_ hash would also work.)
When Python encounters a hash-based pyc, its behavior depends on the setting of
the ``check_source`` flag. If the ``check_source`` flag is set, Python will
@ -94,7 +93,6 @@ References
.. _reproducible build: https://reproducible-builds.org/
.. _Bazel: https://bazel.build/
.. _BLAKE2: https://blake2.net/
.. _SipHash: https://131002.net/siphash/
.. [#frozensets] http://benno.id.au/blog/2013/01/15/python-determinism
.. [#interning] http://bugzilla.opensuse.org/show_bug.cgi?id=1049186
.. _magic number: https://docs.python.org/3/library/importlib.html#importlib.util.MAGIC_NUMBER