Add warning for base64 encode functions

[Eddy114514]

Add a wrapper in base64.py that wrap b64encode, b32encode, b16encode, encodebytes, allowing them to warn user the behavior difference between python2 and python3. Methods' name are listed in the warning message.

warning is as following:

Add test functions in test_base64.py to test whether warning messages shows up.

[nanjekyejoannah]

Main points: Use Py3kWarning Exception and don't prepend Pygrate2

[ltratt]

Are there situations in which _name couldn't be in __all__? I think these functions are always available, on all platforms?

[Eddy114514]

In order to solve this compactivity issue, would using globals() instead of __all__ be a good solution to it?
`

for _name in ["b64encode", "b32encode", "b16encode", "encodebytes"]:
    if _name in globals():
       globals()[_name] = _warn_encode(globals()[_name], _name)

`

[ltratt]

@Eddy114514 Still wondering about this one.

[Eddy114514]

@ltratt After checking with documentations, I think "b64/b32/b16encode" exists in both 2.6 and 3.x, but I should remove encodebytes as it doesn't exist in python 2.

After removing it, I think it is suitable to remove the if-statement

[Eddy114514]

@ltratt Hi, I am wondering if my new implementation resolve the concern. thanks

[ltratt]

@Eddy114514 If you could ping me when you've made updates, it'll help me review them quicker (otherwise I won't notice). Reviewing now.

[Eddy114514]

@Eddy114514 If you could ping me when you've made updates, it'll help me review them quicker (otherwise I won't notice). Reviewing now.

@ltratt Understood, thanks for pointing it out, I will make sure to ping you with updates going forward, sorry about it.

[Eddy114514]

add more test cases to cover b32,16 in test_base64

[Eddy114514]

In order to solve this compactivity issue, would using globals() instead of __all__ be a good solution to it?
`

for _name in ["b64encode", "b32encode", "b16encode", "encodebytes"]:
    if _name in globals():
       globals()[_name] = _warn_encode(globals()[_name], _name)

`

[Eddy114514]

@ltratt After checking with documentations, I think "b64/b32/b16encode" exists in both 2.6 and 3.x, but I should remove encodebytes as it doesn't exist in python 2.

After removing it, I think it is suitable to remove the if-statement

[ltratt]

Please squash. [Note: this doesn't mean you have to squash things down to a single commit, though you can if you want! It does mean that I'd like you to squash down to what you consider a sensible number of commits. More details of this process at https://ykjit.github.io/yk/contributing/prs.html]

[Eddy114514]

Please squash. [Note: this doesn't mean you have to squash things down to a single commit, though you can if you want! It does mean that I'd like you to squash down to what you consider a sensible number of commits. More details of this process at https://ykjit.github.io/yk/contributing/prs.html]

@ltratt Thank you for your instruction, I have squash my commits

Agreed OK

[ltratt]

Hmm, I'm not quite sure why that C code isn't building anymore. It's not directly related to this PR, but it will block it being merged. @Eddy114514 are you able to replicate this issue locally?

[Eddy114514]

Hmm, I'm not quite sure why that C code isn't building anymore. It's not directly related to this PR, but it will block it being merged. @Eddy114514 are you able to replicate this issue locally?

@ltratt I am unable to replicate this issue locally, I can successfully build and run pygrate 2 (and test) locally by following steps (make distclean, ./configure, make -j). However it do will throw some warning and note during "make -j". As following:

[ltratt]

This might be because we're using a new Debian in CI. Will confirm in the morning with luck!

[ltratt]

OK, that failure is a genuine error with this PR I think. @Eddy114514 hopefully you can see the builbot log above.

[Eddy114514]

OK, that failure is a genuine error with this PR I think. @Eddy114514 hopefully you can see the builbot log above.

@ltratt thanks for checking, after checking the log I figured out the problem. My test cases in test_base64 assume python will run with -3 flag such that it will check warnings. I will remove them and only those in py3kwarn

[Eddy114514]

@ltratt I have squashed my change, the failed test file in log is test_base64, I tested again (for it and for all) in local. I think it should be ready to merge. Thank you for your help.

[ltratt]

Please squash.

[Eddy114514]

Please squash.

@ltratt squashed to 1 now.