[Python-Dev] PEP 594: Removing dead batteries from the standard library

[Christian Heimes]

On 21/05/2019 00.13, Antoine Pitrou wrote:
> On Tue, 21 May 2019 00:06:35 +0200
> Christian Heimes <christian at python.org> wrote:
>> On 20/05/2019 23.27, Antoine Pitrou wrote:
>>> NNTP is still quite used (often through GMane, but probably not only) so
>>> I'd question the removal of nntplib.  
>>
>> Is NNTP support important enough to keep the module in the standard library?
> 
> I'd phrase the question differently: is NNTP dead enough, or nntplib
> painful enough to maintain, that it's worth removing it from the stdlib?

The module itself does not create much work. But its tests are a regular source of pain and instabilities. The tests for nntplib depend on external NNTP servers. These servers are sometimes down, very slow, or don't work over IPv6. I'm sure that Pablo and Victor can tell you some war stories. I briefly mentioned the issues in the PEP, too.

> If the stdlib didn't have NNTP support, obviously nobody would suggest
> adding it nowadays.  But it has that support, and there are certainly
> uses of it in the wild, so we must take that into account.
> 
>>> If the wave module depends on the audioop module, and if the wave
>>> module is kept in the stdlib, then the audioop module can't be removed.  
>>
>> No, it can be removed. I explained the situation in the "wave" section of the PEP.
> 
> My bad.  I had skipped that.

The section in audioop was confusing. I have updated the audioop paragraph and the crypt paragraph with your feedback. I'll keep the PR https://github.com/python/peps/pull/1063 option for a couple of days to collect more feedback.

Christian