bpo-31904: Add _crypt module support for VxWorks RTOS

[pxinwr]

This is the successive PR after #11968. This PR adds _crypt module support for VxWorks RTOS.
VxWorks has no APIs crypt() and crypt_r() provided. Instead, DES_crypt() is defined in openssl/des.h to replace the crypt(). It has the same prototype as crypt(). This extension module _crypt will not be built if VxWorks users don't enable the OPENSSL features. See

cpython/setup.py

Lines 976 to 989 in f2f55e7

	defdetect_crypt(self):
	# crypt module.
	ifself.compiler.find_library_file(self.lib_dirs, 'crypt'):
	libs= ['crypt']
	else:
	libs= []
	ifnotVXWORKS:
	self.add(Extension('_crypt', ['_cryptmodule.c'],
	libraries=libs))
	elifself.compiler.find_library_file(self.lib_dirs, 'OPENSSL'):
	libs= ['OPENSSL']
	self.add(Extension('_crypt', ['_cryptmodule.c'],
	libraries=libs))

More and full support on modules for VxWorks will continuously be added by the coming PRs.
VxWorks is a product developed and owned by Wind River. For VxWorks introduction or more details, go to https://www.windriver.com/products/vxworks/
Wind River will have a dedicated engineering team to contribute to the support as maintainers.
We already have a working buildbot worker internally, but has not bound to master. We will check the process for the buildbot, then add it.

https://bugs.python.org/issue31904

[pxinwr]

@serhiy-storchaka I've fixed the problem you mentioned. Could you help to review again and do the merge if the code is OK?

[vstinner]

@tiran: Do you think that using OpenSSL DES_crypt() to implement crypt.crypt() is acceptable?

[vstinner]

Oh, the commit 32f5fdd already added OpenSSL dependency to _crypt Python module in setup.py. IMHO it would be better to not add it before this change, but well, since the commit is already pushed, it's too late and it's ok :-)

[tiran]

-1 for implementing crypt as DES.

Do you really need the crypt module or are you simply implementing it for the sake of having a complete Python stdlib?

I'm planning to deprecate and drop the crypt module. It's not very useful these days and most hashing algorithms are weak. Especially DES is super weak.

[bedevere-bot]

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

[pxinwr]

@tiran Thanks for your comments. I am pinging my PM for the necessity of crypt module. Before I get the decision, could you tell what modules can be used to replace the crypt module?

[tiran]

It depends on your use case. For passwords the hashlib module provides PBKDF2-HMAC and scrypt. Both algorithms are much better than crypt. There are also external packages like bcrypt, cryptography, and passlib, that provide secure password derivation algorithms.

The crypt module is really only useful for legacy systems or for use with the spwd module. The spwd module must not be used on any modern Unix system that uses PAM.

[serhiy-storchaka]

Agree. If VxWorks RTOS for some (likely good) reasons does not provide the system crypt() function, it may be better to not provide it in Python. The crypt module is optional.

[pxinwr]

@tiran@serhiy-storchaka Thanks for your comments. My PM has confirmed that VxWorks can drop the crypt module support. So as you required, I have made the requested changes; please review again.

[bedevere-bot]

Thanks for making the requested changes!

@serhiy-storchaka, @tiran: please review the changes made to this pull request.

[pxinwr]

Already checked the CI failures. That should be caused by other PRs other than this one.

[pxinwr]

Closing it and reopening it is to rerun the CI to update the regression result.

[vstinner]

I would prefer to get a new PR. It's surprising that a PR to add VxWorks support to crypt became "remove crypt". For example, you didn't update the PR title.

Please open a new PR with requested changes.

[bedevere-bot]

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

[pxinwr]

As @vstinner requested, a new PR #12833 has been created to replace this one. And all the issues raised here have been fixed in the new PR #12833