[Python-Dev] Choosing an official stance towards module deprecation in Python 3

[Terry Reedy]

On 9/8/2015 12:59 PM, Brett Cannon wrote:
> There are two discussions going on in the issue tracker about
> deprecating some modules and it has led to the inevitable discussion of
> Python 2/3 compatibility (I'm not even going to bother mentioning the
> issue #s as this thread is not about the modules specifically but module
> deprecation in general). Because I'm tired of rehashing the same
> discussion every time a module deprecation comes up I would like to make
> an official decision that we can follow any time we decide to deprecate
> a module.

Would you consider the 'official decision' as applying to idlelib 
modules?  The issue here is not especially related to 2.7. PEP 434 
formalized the idea that idlelib is mostly private, but acknowledged 
that there might be unauthorized module uses here and there. The first 
example is removing idlelib.idlever, now unused by Idle itself (#21499).

> The approaches to module deprecation I have seen are:
> 1. Nothing changes to the deprecation process; you deprecate a module
> and remove it in one to two releases
> 2. Deprecate the module but with no plans for removal until Python 2.7
> reaches its EOL (I have been calling this Python 4)

For either 1 or 2, the 2.7 code should get a py3 warning.

> 3. Document the deprecation but no actual code deprecation

Number 3 creates a discrepancy between code and doc.  This is normally 
considered a bug, and might be seen as such by someone reading both 
unless a comment were inserted into the code.

> I'm personally in the #2 camp. I want users to be fully aware that the
> module in question is not being updated and possibly not even getting
> non-critical bugfixes, but it's still there in Python 3 in order to make
> sure that you can have code which straddles Python 2 & 3 more easily.

There will be people who convert after 2020, but the final 3.x before 
removal will always be available.  I also see this as the best of 
imperfect choices.

> I don't like #1 because when the module exists in python 2.7 as it makes
> transitioning from Python 2 a bit harder. Leaving code in the stdlib as
> deprecated isn't going to kill us, especially if we make it clear the
> code only exists for transitioning purposes and you very well may have
> to work around any bugs you come across (and yes, this means changing my
> stance for the formatter module's deprecation).

Idlelib currently has about 6 module-naming conventions, and it was 
omitted from the 3.0 renaming. Even without an immediate mass PEP8 
rename (#24225), which would help Idle maintenance, new files from 
refactoring and ttk conversions will result in a large number of 
eventually unused old files. I would prefer to remove them before 2020.

> I don't like #3 because if you have an API memorized or you copied some
> code you found online you very well may not realize a module is
> deprecated. It's not difficult to silence a deprecation warning and you
> can make it so that even if you use -Werror your deprecated module
> imports continue to work without throwing an exception while all other
> deprecations do throw an exception.
>
> Whatever decision we come to I will update PEP 4 and then personally go
> through the various deprecated modules in Python 3.6 and tweak them as
> necessary to follow whatever policy we come up with.

-- 
Terry Jan Reedy