How keep Python 3 moving forward

[Mark Lawrence]

An article by Brett Cannon that I thought might be of interest
http://nothingbutsnark.svbtle.com/my-view-on-the-current-state-of-python-3

 

[Roy Smith]

An article by Brett Cannon that I thought might be of interest
http://nothingbutsnark.svbtle.com/my-view-on-the-current-state-of-python-3

Thanks for the pointer. I installed and ran caniusepython3. It tells
me:

Finding and checking dependencies ...
[WARNING] rpclib not found

You need 19 projects to transition to Python 3.
Of those 19 projects, 17 have no direct dependencies blocking their
transition:

beanstalkc
dateglob
diamond
django-multi-sessions
django-timedeltafield
dnspython
ecks
fabric
gevent (which is blocking grequests)
hash_ring
httmock
jellyfish
boto (which is blocking mrjob)
paste
pyephem
python-cjson
suds

That's a big list. A few of those we could probably work around or
replace with a different module without too much pain. But, between
gevent, boto, fabric, and suds, any idea of migrating is a total
non-starter for us. I imagine they're all working on ports, but I'll
check back in a year and see how things stand.

 

[Ethan Furman]

Thanks for the pointer. I installed and ran caniusepython3. It tells
me:
[snip]

That's a big list. A few of those we could probably work around or
replace with a different module without too much pain. But, between
gevent, boto, fabric, and suds, any idea of migrating is a total
non-starter for us. I imagine they're all working on ports, but I'll
check back in a year and see how things stand.

Don't imagine. Send 'em an email! Let them know there is one more user who'd like a Python 3 port.

 

[Roy Smith]

Thanks for the pointer. I installed and ran caniusepython3. It tells
me:
[snip]

That's a big list. A few of those we could probably work around or
replace with a different module without too much pain. But, between
gevent, boto, fabric, and suds, any idea of migrating is a total
non-starter for us. I imagine they're all working on ports, but I'll
check back in a year and see how things stand.

Don't imagine. Send 'em an email! Let them know there is one more user
who'd like a Python 3 port.

But, you're assuming I want that. I don't. What I have now works.

I'm not trying to be difficult here. I'm just looking at all the things
I could be doing with my time that will improve my product and/or help
grow the business. "Transition to Python 3" isn't even on the list.

 

[wxjmfauth]

Le vendredi 23 mai 2014 22:16:10 UTC+2, Mark Lawrence a écrit :

An article by Brett Cannon that I thought might be of interest

http://nothingbutsnark.svbtle.com/my-view-on-the-current-state-of-python-3



--

My fellow Pythonistas, ask not what our language can do for you, ask

what you can do for our language.



Mark Lawrence



---

This email is free from viruses and malware because avast! Antivirus protection is active.

http://www.avast.com

=========
=========

Quote:
""" And with Python 3.4 I really have not heard anyone complain that they wouldn't like to use Python 3 instead of Python 2. """

Or the devs do not wish to listen.

Python 3 will never work.

jmf

 

[Marko Rauhamaa]

It works for me.

Instead of focusing on bringing legacy libraries to Python3 (for which
there never seems to be a critical need), Python3 needs a brand new
killer module/application/library that is only available on Python3.

Asyncio is a baby step in that direction.


Marko

 

[Rustom Mody]

Instead of focusing on bringing legacy libraries to Python3 (for which
there never seems to be a critical need), Python3 needs a brand new
killer module/application/library that is only available on Python3.

I think those times are over
http://en.wikipedia.org/wiki/Long_tail

Asyncio is a baby step in that direction.

No issue with that

 

[Grant Edwards]

Python 3 will never work.

Neither will color TV.

All that phase detection stuff and that shadow-mask thing?

And airplanes? Bah!

Completely ridiculous.

 

[Devin Jeanpierre]

blindanagram <[email protected]>:
Instead of focusing on bringing legacy libraries to Python3 (for which
there never seems to be a critical need), Python3 needs a brand new
killer module/application/library that is only available on Python3.

Asyncio is a baby step in that direction.

Yikes! Backwards incompatibility is a poor excuse for NIH syndrome.

Don't reinvent the wheel, please. If there's an existing
implementation of a thing, that can save you a lot of work. Even if it
ties you to Python 2, that's worth it, most of the time. If you want
to migrate to Python 3, help that library forward, rather than trying
to make some bespoke replacement you think will be a killer app.

-- Devin

 

[Ethan Furman]

Yikes! Backwards incompatibility is a poor excuse for NIH syndrome.

Don't reinvent the wheel, please. If there's an existing
implementation of a thing, that can save you a lot of work. Even if it
ties you to Python 2, that's worth it, most of the time. If you want
to migrate to Python 3, help that library forward, rather than trying
to make some bespoke replacement you think will be a killer app.

+1

 

[Stefan Behnel]

Devin Jeanpierre, 24.05.2014 18:03:

Yikes! Backwards incompatibility is a poor excuse for NIH syndrome.

Don't reinvent the wheel, please. If there's an existing
implementation of a thing, that can save you a lot of work. Even if it
ties you to Python 2, that's worth it, most of the time. If you want
to migrate to Python 3, help that library forward, rather than trying
to make some bespoke replacement you think will be a killer app.

http://www.joelonsoftware.com/articles/fog0000000069.html

There might still be something that doesn't exist yet, and if you start
working on that, going with Py3 is certainly the right way. For everything
that's there already, however, reusing working, tested code is way better.
And making it work in Py3.

Stefan

 

[Marko Rauhamaa]

If you want to migrate to Python 3, help that library forward, rather
than trying to make some bespoke replacement you think will be a
killer app.

Few people have Python 3 as an objective. What I'm saying is that if
Python 3 had something everybody wants and nothing else provides, the
people will come, even the legacy libraries will be ported then.


Marko

 

[Mark Lawrence]

Few people have Python 3 as an objective. What I'm saying is that if
Python 3 had something everybody wants and nothing else provides, the
people will come, even the legacy libraries will be ported then.

Legacy libraries are being ported, as shown by the green on the Python
Wall of Superpowers here https://python3wos.appspot.com/ (remember this
used to be called the Python Wall of Shame, or something like that).
What makes you think that they're not being ported?

 

[Terry Reedy]

Few people have Python 3 as an objective. What I'm saying is that if
Python 3 had something everybody wants and nothing else provides, the
people will come, even the legacy libraries will be ported then.

I cannot think of anything beyond the core that 'everybody' wants.
However, Python 3.3 has unicode that works for all characters on all
platforms, and some people want that.

 

[wxjmfauth]

Le dimanche 25 mai 2014 02:27:11 UTC+2, Terry Reedy a écrit :

I cannot think of anything beyond the core that 'everybody' wants.

However, Python 3.3 has unicode that works for all characters on all

platforms, and some people want that.

========
========

Python and unicode: a buggy hobbyist toy.
Voilà. Nothing either good or bad.

jmf

 

[Jurko Gospodnetić]

Hi Roy.

You need 19 projects to transition to Python 3.
Of those 19 projects, 17 have no direct dependencies blocking their
transition:

[...snipped...]
suds

That's a big list. A few of those we could probably work around or
replace with a different module without too much pain. But, between
gevent, boto, fabric, and suds, any idea of migrating is a total
non-starter for us. I imagine they're all working on ports, but I'll
check back in a year and see how things stand.

FYI, the suds-jurko fork works on Python 3. And since I'm not aware
of any other actively maintained fork, should I find more free time in
the future I might rename it to suds and try to convert it to a formal
successor to suds.

Hope this helps.

Best regards,
Jurko Gospodnetić

 

[Stefan Behnel]

Roy Smith, 24.05.2014 01:57:

I installed and ran caniusepython3. It tells me:

Finding and checking dependencies ...
[WARNING] rpclib not found

You need 19 projects to transition to Python 3.
Of those 19 projects, 17 have no direct dependencies blocking their
transition:

beanstalkc
dateglob
diamond
django-multi-sessions
django-timedeltafield
dnspython
ecks
fabric
gevent (which is blocking grequests)
hash_ring
httmock
jellyfish
boto (which is blocking mrjob)
paste
pyephem
python-cjson
suds

That's a big list. A few of those we could probably work around or
replace with a different module without too much pain. But, between
gevent, boto, fabric, and suds, any idea of migrating is a total
non-starter for us. I imagine they're all working on ports, but I'll
check back in a year and see how things stand.

Ubuntu provides a (partial) Py3 port of boto. And I don't really see why
you would consider fabric a dependency that keeps you from switching to
Py3. In many cases, you can just keep running it in Py2 as you did before.

Taking a closer look at the "big list" that caniusepython3 spits out will
usually make it shrink to a manageable size. Meaning, the blind size of
that list is not an excuse for anything.

Stefan

 

[Ethan Furman]

Python and unicode: a buggy hobbyist toy.
Voilà. Nothing either good or bad.

I thought this was a moderated list. What exactly are the moderators doing?

 

[Chris Angelico]

I thought this was a moderated list. What exactly are the moderators doing?

It's not a moderated list. We just collectively ignore the (few)
people who aren't saying anything worth reading.

ChrisA

 

[Mark Lawrence]

I thought this was a moderated list. What exactly are the moderators
doing?

I don't think the list is moderated. I do think this guy has had thirty
strikes rather than three, so isn't it time he was finally given out?