Re: DRAFT Introduce CFFI-based Python bindings

[Tomi Ollila <tomi.ollila@iki.fi>]

On Wed, Nov 29 2017, Floris Bruynooghe wrote:

> David Bremner <david@tethera.net> writes:
>
>> Floris Bruynooghe <flub@devork.be> writes:
>>
>>>
>>> Lastly there are some downsides to the choices I made:
>>> - I ended up going squarely for CPython 3.6+.  Choosing Python
>>>   3 allowed better API design, e.g. with keyword-only parameters
>>>   etc.  Choosing CPython 3.4+ restricts the madness that can
>>>   happen with __del__ and gives some newer (tho now unused)
>>>   features in weakref.finalizer.
>>> - This is no longer drop-in compatible.
>>> - I haven't got to a stage where my initial goal of speed has
>>>   been proven yet.
>>
>> I guess you'll have to convince the maintainers / users of alot and afew
>> that this makes sense before we go much further. I'd point out that
>> Debian stable is only at python 3.5, so that makes me a bit wary of this
>> (being able to run the test suite on debian stable and similar aged
>> distros useful for me, and I suspect other developers).
>>
>> I know there are issues with memory management in the current bindings,
>> so that may be a strong reason to push to python 3.6; it seems to need
>> more investigation at the moment.
>
> So on earlier Python versions, sure this is possible at not too much
> cost.
>
> - Python 3.4+ would just cost the use of some f-strings.  Not major, was
>   just nice to use.
> - Python <3.4 afaik would only need a tweak to the Database.tags and
>   Message.tags properties.  I *think* swapping the caching of these
>   using a weakref should suffice and not break the brittle
>   Python-libnotmuch memory management.
>   Mind you I think Python 3.0-3.3 are pretty old and not much point in
>   supporting them.  But this would also apply for 2.7 support.
> - Python 2.7 is probably the worst, in that keyword-only arguments would
>   be gone.  If python 2.7 is required I'd be much keener to have another
>   go at a drop-in replacement with the memory safety features and then
>   build the "notdb" API on top off it.  But for that to be worth it
>   people need to be convinced enough that maintaining a CFFI version is
>   nicer than a ctypes version I guess.

IMO Python 3.4+ would be OK, if python 2 support can be dropped. 
Even Ubuntu 14.04 has python 3.4. One notable distribution that has
Python 3.3 by default is RHEL 7, but there seems to be quite a few
packaged alternatives available...

>
> Kind regards,
> Floris


Tomi