| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* drop python 3.4 support
Our dependencies have started dropping support so it is becoming
difficult to test. Additionally, Python 3.4 represents <2% of our
downloads, so taking on a large maintenance burden to maintain support
isn't a good use of limited time. Accordingly, we're dropping testing
infrastructure and migrating our abi3 wheels to py35+.
* use removed instead of dropped
|
| |
|
|
|
|
| |
* Fixes #5010 -- test and build 3.8 wheels
* try using isolated_build = True to work around a failure
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Remove non-test dependencies on asn1crypto.
cryptography.io actually contains two OpenSSL bindings right now, the
expected cffi one, and an optional one hidden in asn1crypto. asn1crypto
contains a lot of things that cryptography.io doesn't use, including a
BER parser and a hand-rolled and not constant-time EC implementation.
Instead, check in a much small DER-only parser in cryptography/hazmat. A
quick benchmark suggests this parser is also faster than asn1crypto:
from __future__ import absolute_import, division, print_function
import timeit
print(timeit.timeit(
"decode_dss_signature(sig)",
setup=r"""
from cryptography.hazmat.primitives.asymmetric.utils import decode_dss_signature
sig=b"\x30\x2d\x02\x15\x00\xb5\xaf\x30\x78\x67\xfb\x8b\x54\x39\x00\x13\xcc\x67\x02\x0d\xdf\x1f\x2c\x0b\x81\x02\x14\x62\x0d\x3b\x22\xab\x50\x31\x44\x0c\x3e\x35\xea\xb6\xf4\x81\x29\x8f\x9e\x9f\x08"
""",
number=10000))
Python 2.7:
asn1crypto: 0.25
_der.py: 0.098
Python 3.5:
asn1crypto: 0.17
_der.py: 0.10
* Remove test dependencies on asn1crypto.
The remaining use of asn1crypto was some sanity-checking of
Certificates. Add a minimal X.509 parser to extract the relevant fields.
* Add a read_single_element helper function.
The outermost read is a little tedious.
* Address flake8 warnings
* Fix test for long-form vs short-form lengths.
Testing a zero length trips both this check and the non-minimal long
form check. Use a one-byte length to cover the missing branch.
* Remove support for negative integers.
These never come up in valid signatures. Note, however, this does
change public API.
* Update src/cryptography/hazmat/primitives/asymmetric/utils.py
Co-Authored-By: Alex Gaynor <alex.gaynor@gmail.com>
* Review comments
* Avoid hardcoding the serialization of NULL in decode_asn1.py too.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Drop setup.py test support.
It's fragile and relies on eggs. While our downstream redistributors don't love this (sorry!), it seems like the right tradeoff.
* Remove unused
* Added a changelog entry
* typo
* line length
|
| |
|
| |
It already defaults to x-rst, but this silences a warning
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* allow bytearrays for key/iv for symmetric encryption
* bump pypy/cffi requirements
* update docs, fix some tests
* old openssl is naught but pain
* revert a typo
* use trusty for old pypy
* better error msg again
* restore match
|
| |
|
|
|
|
|
|
|
|
| |
* remove idna as a primary dependency
* empty commit
* dynamodb test fix (thanks to Matt Bullock)
* review feedback
|
| | |
|
| |
|
|
|
|
|
|
| |
* ci is suffering again, is it the new pytest or hypothesis?
* now try 3.79.2 hypothesis
* it's a blacklist party
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* update pytest config
pytest 3.8.0 was just released and officially deprecates some of the way
we do pytest marks. They introduced a new way to do this in 3.6 so this
PR switches to that mechanism and updates our minimum pytest requirement
* update the stubs
* also update wycheproof test config to remove deprecated paths
* don't need this any more
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* test against python 3.7 for windows
* update docs to say we test on 3.7
* more succinct
* maybe make this actually work.
* link properly
* moar changes
|
| | |
|
| |
|
|
|
|
| |
* Remove cffi branch for pypy that's not needed
* simplify further
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
* install only the minimum required to build our docs for rtd
We don't need to install sphinxcontrib-spelling or doc8 in rtd.
* move another package
|
| |
|
|
|
|
| |
* temporarily pin cffi to avoid bug related to virtualenv
* right, don't duplicate the name
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* clearly error out on older setuptools
* use the right thing
|
| |
|
|
|
|
| |
* Fixes #3947 -- remove the docutils.conf
* include minimum versions
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* expunge python 2.6
* how did THAT happen
* remove another unsupported python from the tox envlist
* hypothesis can now be unconditionally imported
|
| |
|
|
| |
This makes it so we don't have to pin sphinx while we wait for
sphinxcontrib-spelling to properly handle smart quotes
|
| | |
|
| |
|
|
|
|
|
|
| |
* Switch to using the other style of environment markers
* We don't do py3.3 any more!
* changelog
|
| | |
|
| |
|
| |
See upstream: https://github.com/pytest-dev/pytest/issues/2644
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* remove python 3.3 from CI
* remove 3.3 everywhere, not just from CI
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
* Bump the minimum PyPy/cffi version and simplify as a result
* unused imports
* grumble, fix
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* time to remove commoncrypto, fare thee well
* remove even more
* update the changelog
* remove more things
* don't need this function
* remove CAST5 CTR tests since that was only supported in commoncrypto
* assert a thing
|
| | |
|
| | |
|
| |
|
| |
This is sort of a pre-req for #3508 (but not really, because setuptools needs to be this version before setup.py even runs for that PR). Because we still support older setuptools versions in our pkg_resources code, this is mostly a gentle nudge to the ecosystem.
|
| |
|
|
| |
performance/memory usage (#3411)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* add support for update_into on CipherContext
This allows you to provide your own buffer (like recv_into) to improve
performance when repeatedly calling encrypt/decrypt on large payloads.
* another skip_if
* more skip_if complexity
* maybe do this right
* correct number of args
* coverage for the coverage gods
* add a cffi minimum test tox target and travis builder
This tests against macOS so we capture some commoncrypto branches
* extra arg
* need to actually install py35
* fix
* coverage for GCM decrypt in CC
* no longer relevant
* 1.8 now
* pep8
* dramatically simplify
* update docs
* remove unneeded test
* changelog entry
* test improvements
* coverage fix
* add some comments to example
* move the comments to their own line
* fix and move comment
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* replace pyasn1 with asn1crypto
* allow trailing bytes
* fix x509 test
* update CHANGELOG.rst
* fix assert
* make asn1crypto code more idiomatic
* find tag
* final clean-up
* leave trailing byte logic unchanged
* document dependency change
* spelling
* fix spelling
|
