[Matplotlib-devel] What minimum version of python3?

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

[Matplotlib-devel] What minimum version of python3?

tcaswell
These are derived from today's call:

Pro 3.6:
 - by the time we release, 3.7 will be out so will support "two most recent python"
    - this gives 3 years of support on master of each python version (assuming python sticks 
 - ordered, faster dictionaries (as implementation detail in 3.6, guaranteed in 3.7) 
 - guaranteed ordered kwargs / class definitions (do precedence of conflicting kwaargs)
 - fstrings
 - fspath / pathlib protocol

Con 3.6:
 - feels a bit agressive as 3.7 is not out yet
 - might get ahead of some LTS releases
   - but older version of Matplotlib will still work
   - user-space environments go along way to fixing this
     - conda
   - ubuntu 18.04 is planned to be py3.6 by default

Pro 3.5:
 - more conservative approach
 - pypy3 support (but they will support 3.6 eventually)

The consensus on the call and on gitter seems to be for 3.6 as the minimum. 

Tom 


_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

Matthew Brett
On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]> wrote:

> These are derived from today's call:
>
> Pro 3.6:
>  - by the time we release, 3.7 will be out so will support "two most recent
> python"
>     - this gives 3 years of support on master of each python version
> (assuming python sticks
>  - ordered, faster dictionaries (as implementation detail in 3.6, guaranteed
> in 3.7)
>  - guaranteed ordered kwargs / class definitions (do precedence of
> conflicting kwaargs)
>  - fstrings
>  - fspath / pathlib protocol
>
> Con 3.6:
>  - feels a bit agressive as 3.7 is not out yet
>  - might get ahead of some LTS releases
>    - but older version of Matplotlib will still work
>    - user-space environments go along way to fixing this
>      - conda
>      -
> https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>    - ubuntu 18.04 is planned to be py3.6 by default
>
> Pro 3.5:
>  - more conservative approach
>  - pypy3 support (but they will support 3.6 eventually)
>
> The consensus on the call and on gitter seems to be for 3.6 as the minimum.

Man - that seems pretty extreme.   Are the 3.6 features really worth
it?  There must be a lot of Linuces at Python 3.5, even at the most
recent distribution.  Was there by any chance a majority of conda
users on the call?

Cheers,

Matthew
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

Nathan Goldbaum
Ubuntu 14.04 (which admittedly goes EOL in April) is still on python 3.4, for example.

On Mon, Feb 12, 2018 at 5:28 PM Matthew Brett <[hidden email]> wrote:
On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]> wrote:
> These are derived from today's call:
>
> Pro 3.6:
>  - by the time we release, 3.7 will be out so will support "two most recent
> python"
>     - this gives 3 years of support on master of each python version
> (assuming python sticks
>  - ordered, faster dictionaries (as implementation detail in 3.6, guaranteed
> in 3.7)
>  - guaranteed ordered kwargs / class definitions (do precedence of
> conflicting kwaargs)
>  - fstrings
>  - fspath / pathlib protocol
>
> Con 3.6:
>  - feels a bit agressive as 3.7 is not out yet
>  - might get ahead of some LTS releases
>    - but older version of Matplotlib will still work
>    - user-space environments go along way to fixing this
>      - conda
>      -
> https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>    - ubuntu 18.04 is planned to be py3.6 by default
>
> Pro 3.5:
>  - more conservative approach
>  - pypy3 support (but they will support 3.6 eventually)
>
> The consensus on the call and on gitter seems to be for 3.6 as the minimum.

Man - that seems pretty extreme.   Are the 3.6 features really worth
it?  There must be a lot of Linuces at Python 3.5, even at the most
recent distribution.  Was there by any chance a majority of conda
users on the call?

Cheers,

Matthew
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel

_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

vincent.adrien@gmail.com
Well, I do not know how accurate is the following script:
https://github.com/Mariatta/python_versions_and_distros
but looking at the list it returned on 2017-10-15, it looks like Python
3.6+ may be quite common **outside of the usual left-behind
distributions ala Debian stable or CentOS**.

Adrien,
conda lover since he had to work with CentOS ^^...

On 02/12/2018 03:48 PM, Nathan Goldbaum wrote:

> Ubuntu 14.04 (which admittedly goes EOL in April) is still on python
> 3.4, for example.
>
> On Mon, Feb 12, 2018 at 5:28 PM Matthew Brett <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]
>     <mailto:[hidden email]>> wrote:
>      > These are derived from today's call:
>      >
>      > Pro 3.6:
>      >  - by the time we release, 3.7 will be out so will support "two
>     most recent
>      > python"
>      >     - this gives 3 years of support on master of each python version
>      > (assuming python sticks
>      >  - ordered, faster dictionaries (as implementation detail in 3.6,
>     guaranteed
>      > in 3.7)
>      >  - guaranteed ordered kwargs / class definitions (do precedence of
>      > conflicting kwaargs)
>      >  - fstrings
>      >  - fspath / pathlib protocol
>      >
>      > Con 3.6:
>      >  - feels a bit agressive as 3.7 is not out yet
>      >  - might get ahead of some LTS releases
>      >    - but older version of Matplotlib will still work
>      >    - user-space environments go along way to fixing this
>      >      - conda
>      >      -
>      >
>     https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>      >    - ubuntu 18.04 is planned to be py3.6 by default
>      >
>      > Pro 3.5:
>      >  - more conservative approach
>      >  - pypy3 support (but they will support 3.6 eventually)
>      >
>      > The consensus on the call and on gitter seems to be for 3.6 as
>     the minimum.
>
>     Man - that seems pretty extreme.   Are the 3.6 features really worth
>     it?  There must be a lot of Linuces at Python 3.5, even at the most
>     recent distribution.  Was there by any chance a majority of conda
>     users on the call?
>
>     Cheers,
>
>     Matthew
>     _______________________________________________
>     Matplotlib-devel mailing list
>     [hidden email] <mailto:[hidden email]>
>     https://mail.python.org/mailman/listinfo/matplotlib-devel
>
>
>
> _______________________________________________
> Matplotlib-devel mailing list
> [hidden email]
> https://mail.python.org/mailman/listinfo/matplotlib-devel
>

_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

Chris Barker - NOAA Federal
How many people are using the district-supplied python these days?

Do they need the latest MPL?

I have literally no idea.

-CHB

Sent from my iPhone

> On Feb 12, 2018, at 3:54 PM, "[hidden email]" <[hidden email]> wrote:
>
> Well, I do not know how accurate is the following script:
> https://github.com/Mariatta/python_versions_and_distros
> but looking at the list it returned on 2017-10-15, it looks like Python 3.6+ may be quite common **outside of the usual left-behind distributions ala Debian stable or CentOS**.
>
> Adrien,
> conda lover since he had to work with CentOS ^^...
>
>> On 02/12/2018 03:48 PM, Nathan Goldbaum wrote:
>> Ubuntu 14.04 (which admittedly goes EOL in April) is still on python 3.4, for example.
>> On Mon, Feb 12, 2018 at 5:28 PM Matthew Brett <[hidden email] <mailto:[hidden email]>> wrote:
>>    On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]
>>    <mailto:[hidden email]>> wrote:
>>     > These are derived from today's call:
>>     >
>>     > Pro 3.6:
>>     >  - by the time we release, 3.7 will be out so will support "two
>>    most recent
>>     > python"
>>     >     - this gives 3 years of support on master of each python version
>>     > (assuming python sticks
>>     >  - ordered, faster dictionaries (as implementation detail in 3.6,
>>    guaranteed
>>     > in 3.7)
>>     >  - guaranteed ordered kwargs / class definitions (do precedence of
>>     > conflicting kwaargs)
>>     >  - fstrings
>>     >  - fspath / pathlib protocol
>>     >
>>     > Con 3.6:
>>     >  - feels a bit agressive as 3.7 is not out yet
>>     >  - might get ahead of some LTS releases
>>     >    - but older version of Matplotlib will still work
>>     >    - user-space environments go along way to fixing this
>>     >      - conda
>>     >      -
>>     >
>>    https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>>     >    - ubuntu 18.04 is planned to be py3.6 by default
>>     >
>>     > Pro 3.5:
>>     >  - more conservative approach
>>     >  - pypy3 support (but they will support 3.6 eventually)
>>     >
>>     > The consensus on the call and on gitter seems to be for 3.6 as
>>    the minimum.
>>    Man - that seems pretty extreme.   Are the 3.6 features really worth
>>    it?  There must be a lot of Linuces at Python 3.5, even at the most
>>    recent distribution.  Was there by any chance a majority of conda
>>    users on the call?
>>    Cheers,
>>    Matthew
>>    _______________________________________________
>>    Matplotlib-devel mailing list
>>    [hidden email] <mailto:[hidden email]>
>>    https://mail.python.org/mailman/listinfo/matplotlib-devel
>> _______________________________________________
>> Matplotlib-devel mailing list
>> [hidden email]
>> https://mail.python.org/mailman/listinfo/matplotlib-devel
>
> _______________________________________________
> Matplotlib-devel mailing list
> [hidden email]
> https://mail.python.org/mailman/listinfo/matplotlib-devel
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

tcaswell
We are talking about a release scheduled for July/August 2018 when 3.7 will already be out and 3.6 will have been out for ~20 months so support coverage will only get better.

From the point of view of distributions, I am mostly worried about running into an issue with a distribution that will not be able to package  Matplotlib.  Are there any distributions that _will_ want to package mpl 3.0 in August and will _not_ have 3.6 available in August?

If people are using system python / Matplotlib on older versions of OSs, then they are using the old version that is packaged and are OK.  If they want to build user-space envs there are a bunch of options (pyenv, conda, activestate) which all current support python 3.6 as well as PPAs / IUS / .. for system level 3.6.

I think it is incumbent on those advocating we support 3.5 to produce any sort of estimate of the harm for dropping 3.5.

I am advocating a policy of 'last 2 python feature releases' (which is what I remember from when I first started working on Matplotlib), what is the alternative proposal?

As I have said before, I think the community's expectations have been warped by having to support and provide new features for an almost decade old version of python.

As a final fallback, the 2.2 series will continue to support 3.4 and 3.5 until 2020.

Tom

On Mon, Feb 12, 2018 at 9:00 PM Chris Barker - NOAA Federal <[hidden email]> wrote:
How many people are using the district-supplied python these days?

Do they need the latest MPL?

I have literally no idea.

-CHB

Sent from my iPhone

> On Feb 12, 2018, at 3:54 PM, "[hidden email]" <[hidden email]> wrote:
>
> Well, I do not know how accurate is the following script:
> https://github.com/Mariatta/python_versions_and_distros
> but looking at the list it returned on 2017-10-15, it looks like Python 3.6+ may be quite common **outside of the usual left-behind distributions ala Debian stable or CentOS**.
>
> Adrien,
> conda lover since he had to work with CentOS ^^...
>
>> On 02/12/2018 03:48 PM, Nathan Goldbaum wrote:
>> Ubuntu 14.04 (which admittedly goes EOL in April) is still on python 3.4, for example.
>> On Mon, Feb 12, 2018 at 5:28 PM Matthew Brett <[hidden email] <mailto:[hidden email]>> wrote:
>>    On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]
>>    <mailto:[hidden email]>> wrote:
>>     > These are derived from today's call:
>>     >
>>     > Pro 3.6:
>>     >  - by the time we release, 3.7 will be out so will support "two
>>    most recent
>>     > python"
>>     >     - this gives 3 years of support on master of each python version
>>     > (assuming python sticks
>>     >  - ordered, faster dictionaries (as implementation detail in 3.6,
>>    guaranteed
>>     > in 3.7)
>>     >  - guaranteed ordered kwargs / class definitions (do precedence of
>>     > conflicting kwaargs)
>>     >  - fstrings
>>     >  - fspath / pathlib protocol
>>     >
>>     > Con 3.6:
>>     >  - feels a bit agressive as 3.7 is not out yet
>>     >  - might get ahead of some LTS releases
>>     >    - but older version of Matplotlib will still work
>>     >    - user-space environments go along way to fixing this
>>     >      - conda
>>     >      -
>>     >
>>    https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>>     >    - ubuntu 18.04 is planned to be py3.6 by default
>>     >
>>     > Pro 3.5:
>>     >  - more conservative approach
>>     >  - pypy3 support (but they will support 3.6 eventually)
>>     >
>>     > The consensus on the call and on gitter seems to be for 3.6 as
>>    the minimum.
>>    Man - that seems pretty extreme.   Are the 3.6 features really worth
>>    it?  There must be a lot of Linuces at Python 3.5, even at the most
>>    recent distribution.  Was there by any chance a majority of conda
>>    users on the call?
>>    Cheers,
>>    Matthew
>>    _______________________________________________
>>    Matplotlib-devel mailing list
>>    [hidden email] <mailto:[hidden email]>
>>    https://mail.python.org/mailman/listinfo/matplotlib-devel
>> _______________________________________________
>> Matplotlib-devel mailing list
>> [hidden email]
>> https://mail.python.org/mailman/listinfo/matplotlib-devel
>
> _______________________________________________
> Matplotlib-devel mailing list
> [hidden email]
> https://mail.python.org/mailman/listinfo/matplotlib-devel
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel

_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

tcaswell
We discussed the minimum python again on the call this past week and settled on:

 - support minor versions of python initially released in the previous 3 years
 - support minor versions of python initially released in the previous 3 years or oldest that supports the minimum python version (which ever is higher)

We will bump the minimum python an numpy versions as we can every minor and major release.

We did not discuss other dependencies, but I propose: 
 - for system/c-dependencies (libpng, freetype, GUI frameworks, latex, gs, ffmpeg) support as old as practical
 - for python dependencies with compiled extensions 3 years or the oldest that support our minimum python + numpy 
 - for python dependencies without complied extensions at least 2 years or the oldest that supports our minimum python.  

Bump these minimum versions as they no longer support our minimum python or we want to use new features in them.

We should never bump minimum versions on a patch release.

For mpl 3.0 (July 2018) this means python 3.5 (released Sept 2015) and numpy 1.10 (release in Oct 2015).  Python 3.4 was released Mar 2014 and  Numpy 1.9 was released Sept 2014

For mpl 3.1 (January 2019) this means python 3.6 (released Dec 2016) and numpy 1.11 (March 2016).

Tom

On Mon, Feb 12, 2018 at 11:38 PM Thomas Caswell <[hidden email]> wrote:
We are talking about a release scheduled for July/August 2018 when 3.7 will already be out and 3.6 will have been out for ~20 months so support coverage will only get better.

From the point of view of distributions, I am mostly worried about running into an issue with a distribution that will not be able to package  Matplotlib.  Are there any distributions that _will_ want to package mpl 3.0 in August and will _not_ have 3.6 available in August?

If people are using system python / Matplotlib on older versions of OSs, then they are using the old version that is packaged and are OK.  If they want to build user-space envs there are a bunch of options (pyenv, conda, activestate) which all current support python 3.6 as well as PPAs / IUS / .. for system level 3.6.

I think it is incumbent on those advocating we support 3.5 to produce any sort of estimate of the harm for dropping 3.5.

I am advocating a policy of 'last 2 python feature releases' (which is what I remember from when I first started working on Matplotlib), what is the alternative proposal?

As I have said before, I think the community's expectations have been warped by having to support and provide new features for an almost decade old version of python.

As a final fallback, the 2.2 series will continue to support 3.4 and 3.5 until 2020.

Tom

On Mon, Feb 12, 2018 at 9:00 PM Chris Barker - NOAA Federal <[hidden email]> wrote:
How many people are using the district-supplied python these days?

Do they need the latest MPL?

I have literally no idea.

-CHB

Sent from my iPhone

> On Feb 12, 2018, at 3:54 PM, "[hidden email]" <[hidden email]> wrote:
>
> Well, I do not know how accurate is the following script:
> https://github.com/Mariatta/python_versions_and_distros
> but looking at the list it returned on 2017-10-15, it looks like Python 3.6+ may be quite common **outside of the usual left-behind distributions ala Debian stable or CentOS**.
>
> Adrien,
> conda lover since he had to work with CentOS ^^...
>
>> On 02/12/2018 03:48 PM, Nathan Goldbaum wrote:
>> Ubuntu 14.04 (which admittedly goes EOL in April) is still on python 3.4, for example.
>> On Mon, Feb 12, 2018 at 5:28 PM Matthew Brett <[hidden email] <mailto:[hidden email]>> wrote:
>>    On Mon, Feb 12, 2018 at 10:40 PM, Thomas Caswell <[hidden email]
>>    <mailto:[hidden email]>> wrote:
>>     > These are derived from today's call:
>>     >
>>     > Pro 3.6:
>>     >  - by the time we release, 3.7 will be out so will support "two
>>    most recent
>>     > python"
>>     >     - this gives 3 years of support on master of each python version
>>     > (assuming python sticks
>>     >  - ordered, faster dictionaries (as implementation detail in 3.6,
>>    guaranteed
>>     > in 3.7)
>>     >  - guaranteed ordered kwargs / class definitions (do precedence of
>>     > conflicting kwaargs)
>>     >  - fstrings
>>     >  - fspath / pathlib protocol
>>     >
>>     > Con 3.6:
>>     >  - feels a bit agressive as 3.7 is not out yet
>>     >  - might get ahead of some LTS releases
>>     >    - but older version of Matplotlib will still work
>>     >    - user-space environments go along way to fixing this
>>     >      - conda
>>     >      -
>>     >
>>    https://www.digitalocean.com/community/tutorials/how-to-install-python-3-and-set-up-a-local-programming-environment-on-centos-7
>>     >    - ubuntu 18.04 is planned to be py3.6 by default
>>     >
>>     > Pro 3.5:
>>     >  - more conservative approach
>>     >  - pypy3 support (but they will support 3.6 eventually)
>>     >
>>     > The consensus on the call and on gitter seems to be for 3.6 as
>>    the minimum.
>>    Man - that seems pretty extreme.   Are the 3.6 features really worth
>>    it?  There must be a lot of Linuces at Python 3.5, even at the most
>>    recent distribution.  Was there by any chance a majority of conda
>>    users on the call?
>>    Cheers,
>>    Matthew
>>    _______________________________________________
>>    Matplotlib-devel mailing list
>>    [hidden email] <mailto:[hidden email]>
>>    https://mail.python.org/mailman/listinfo/matplotlib-devel
>> _______________________________________________
>> Matplotlib-devel mailing list
>> [hidden email]
>> https://mail.python.org/mailman/listinfo/matplotlib-devel
>
> _______________________________________________
> Matplotlib-devel mailing list
> [hidden email]
> https://mail.python.org/mailman/listinfo/matplotlib-devel
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel

_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

Matthew Brett
Hi,

On Sat, Mar 3, 2018 at 6:40 PM, Thomas Caswell <[hidden email]> wrote:

> We discussed the minimum python again on the call this past week and settled
> on:
>
>  - support minor versions of python initially released in the previous 3
> years
>  - support minor versions of python initially released in the previous 3
> years or oldest that supports the minimum python version (which ever is
> higher)
>
> We will bump the minimum python an numpy versions as we can every minor and
> major release.
>
> We did not discuss other dependencies, but I propose:
>  - for system/c-dependencies (libpng, freetype, GUI frameworks, latex, gs,
> ffmpeg) support as old as practical
>  - for python dependencies with compiled extensions 3 years or the oldest
> that support our minimum python + numpy
>  - for python dependencies without complied extensions at least 2 years or
> the oldest that supports our minimum python.
>
> Bump these minimum versions as they no longer support our minimum python or
> we want to use new features in them.
>
> We should never bump minimum versions on a patch release.
>
> For mpl 3.0 (July 2018) this means python 3.5 (released Sept 2015) and numpy
> 1.10 (release in Oct 2015).  Python 3.4 was released Mar 2014 and  Numpy 1.9
> was released Sept 2014
>
> For mpl 3.1 (January 2019) this means python 3.6 (released Dec 2016) and
> numpy 1.11 (March 2016).

That seems very reasonable to me ...

Cheers,

Matthew
_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel
Reply | Threaded
Open this post in threaded view
|

Re: What minimum version of python3?

tcaswell

On Sat, Mar 3, 2018 at 1:50 PM Matthew Brett <[hidden email]> wrote:
Hi,

On Sat, Mar 3, 2018 at 6:40 PM, Thomas Caswell <[hidden email]> wrote:
> We discussed the minimum python again on the call this past week and settled
> on:
>
>  - support minor versions of python initially released in the previous 3
> years
>  - support minor versions of python initially released in the previous 3
> years or oldest that supports the minimum python version (which ever is
> higher)
>
> We will bump the minimum python an numpy versions as we can every minor and
> major release.
>
> We did not discuss other dependencies, but I propose:
>  - for system/c-dependencies (libpng, freetype, GUI frameworks, latex, gs,
> ffmpeg) support as old as practical
>  - for python dependencies with compiled extensions 3 years or the oldest
> that support our minimum python + numpy
>  - for python dependencies without complied extensions at least 2 years or
> the oldest that supports our minimum python.
>
> Bump these minimum versions as they no longer support our minimum python or
> we want to use new features in them.
>
> We should never bump minimum versions on a patch release.
>
> For mpl 3.0 (July 2018) this means python 3.5 (released Sept 2015) and numpy
> 1.10 (release in Oct 2015).  Python 3.4 was released Mar 2014 and  Numpy 1.9
> was released Sept 2014
>
> For mpl 3.1 (January 2019) this means python 3.6 (released Dec 2016) and
> numpy 1.11 (March 2016).

That seems very reasonable to me ...

Cheers,

Matthew

_______________________________________________
Matplotlib-devel mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/matplotlib-devel