Mercurial for Python3

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
10 messages Options
Reply | Threaded
Open this post in threaded view
|

Mercurial for Python3

Detlev Offenbach
Hello,

are there any plans to port Mercurial to Python3?

Regards,
Detlev
--
Detlev Offenbach
[hidden email]
_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Augie Fackler-3
On Mon, May 9, 2011 at 12:29 PM, Detlev Offenbach
<[hidden email]> wrote:
> Hello,
>
> are there any plans to port Mercurial to Python3?

Not in the immediate term, as there aren't any obvious benefits for us
and there are huge pain points because we need to support back to
Python 2.4. There's also no observable demand for Python 3.

>
> Regards,
> Detlev
> --
> Detlev Offenbach
> [hidden email]
> _______________________________________________
> Mercurial mailing list
> [hidden email]
> http://selenic.com/mailman/listinfo/mercurial
>
_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Matt Mackall
In reply to this post by Detlev Offenbach
On Mon, 2011-05-09 at 19:29 +0200, Detlev Offenbach wrote:
> Hello,
>
> are there any plans to port Mercurial to Python3?

Not in the near future. The reasons are many:

- We have to continue supporting enterprise distributions with 2.x for
the next several years
- Supporting both simultaneously has proven extremely difficult (I
estimate at least a developer-year of effort)
- Python3 doesn't have any features we care about
- Very few users are asking for Python3
- We don't have a public, supported Python API, so there aren't many
legitimate users:

http://mercurial.selenic.com/wiki/MercurialApi

If what you want is a stable API you can use from Python 3, you should
use os.system(). Or you should wait for us to finish the command server
API, which is one of this year's GSoC projects.

--
Mathematics is the supreme nostalgia of our time.


_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

James Mansion-2
On Mon, 09 May 2011 18:57:45 +0100, Matt Mackall <[hidden email]> wrote:
> - Very few users are asking for Python3

I've changed completely and I'm prepared to give up Twisted to do it.
<shrug>
_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Matt Mackall
On Mon, 2011-05-09 at 20:53 +0100, James Mansion wrote:
> On Mon, 09 May 2011 18:57:45 +0100, Matt Mackall <[hidden email]> wrote:
> > - Very few users are asking for Python3
>
> I've changed completely and I'm prepared to give up Twisted to do it.

The difference is that Twisted is a library. Mercurial emphatically is
not.

--
Mathematics is the supreme nostalgia of our time.


_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Arne Babenhauserheide-2
In reply to this post by James Mansion-2
On Monday 09 May 2011 20:53:04 James Mansion wrote:
> On Mon, 09 May 2011 18:57:45 +0100, Matt Mackall <[hidden email]> wrote:
> > - Very few users are asking for Python3
>
> I've changed completely and I'm prepared to give up Twisted to do it.
> <shrug>

I also changed for most of my projects, but I can live with Mercurial sticking
to 2.x a bit longer.

I play with using it from Python for a wiki from time to time, though, and
there not running in py3 can hurt (in py3 unicode handling is much nicer than
in py2).

If you remember my hg for servers evaluation¹, you’ll know the speed difference
you get by just exchanging dispatch for commit.commit()…

¹: http://draketo.de/proj/hg-vs-git-server/test-results.html - last plot on
the page.

Best wishes,
Arne
_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial

signature.asc (325 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Yonggang Luo
In reply to this post by Augie Fackler-3
There is observable demand for python 3.
The Unicode is most demanded part.
And I don't think fixut8 plugin is an good solution to deal with such an problem
For non-Latin developers, hg can not replace svn, the only reason is the UNICODE.

_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Matt Mackall
On Thu, 2011-06-02 at 18:54 -0700, 罗勇刚 wrote:
> There is observable demand for python 3.
> The Unicode is most demanded part.
> And I don't think fixut8 plugin is an good solution to deal with such an
> problem
> For non-Latin developers, hg can not replace svn, the only reason is the
> UNICODE.

Even if we spent the year it would take to port Mercurial to Python 3...
it would not change how we handle Unicode at all. It's unrelated.

--
Mathematics is the supreme nostalgia of our time.


_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Maxim Khitrov-2
In reply to this post by Matt Mackall
On Mon, May 9, 2011 at 1:57 PM, Matt Mackall <[hidden email]> wrote:

> On Mon, 2011-05-09 at 19:29 +0200, Detlev Offenbach wrote:
>> Hello,
>>
>> are there any plans to port Mercurial to Python3?
>
> Not in the near future. The reasons are many:
>
> - We have to continue supporting enterprise distributions with 2.x for
> the next several years
> - Supporting both simultaneously has proven extremely difficult (I
> estimate at least a developer-year of effort)
> - Python3 doesn't have any features we care about
> - Very few users are asking for Python3
> - We don't have a public, supported Python API, so there aren't many
> legitimate users:
>
> http://mercurial.selenic.com/wiki/MercurialApi
>
> If what you want is a stable API you can use from Python 3, you should
> use os.system(). Or you should wait for us to finish the command server
> API, which is one of this year's GSoC projects.

Just curious if any of the reasons have changed over the past eight
months - any chance of Python 3 support in 2012? For me, it would
eliminate the need to maintain Python 2.7 next to 3.2 on my servers.

- Max
_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial
Reply | Threaded
Open this post in threaded view
|

Re: Mercurial for Python3

Matt Mackall
On Sat, 2012-01-07 at 21:21 -0500, Maxim Khitrov wrote:

> On Mon, May 9, 2011 at 1:57 PM, Matt Mackall <[hidden email]> wrote:
> > On Mon, 2011-05-09 at 19:29 +0200, Detlev Offenbach wrote:
> >> Hello,
> >>
> >> are there any plans to port Mercurial to Python3?
> >
> > Not in the near future. The reasons are many:
> >
> > - We have to continue supporting enterprise distributions with 2.x for
> > the next several years
> > - Supporting both simultaneously has proven extremely difficult (I
> > estimate at least a developer-year of effort)
> > - Python3 doesn't have any features we care about
> > - Very few users are asking for Python3
> > - We don't have a public, supported Python API, so there aren't many
> > legitimate users:
> >
> > http://mercurial.selenic.com/wiki/MercurialApi
> >
> > If what you want is a stable API you can use from Python 3, you should
> > use os.system(). Or you should wait for us to finish the command server
> > API, which is one of this year's GSoC projects.
>
> Just curious if any of the reasons have changed over the past eight
> months - any chance of Python 3 support in 2012? For me, it would
> eliminate the need to maintain Python 2.7 next to 3.2 on my servers.

Given it's a developer-year of effort and no such developer has shown
up, I'd have to say no. You're much more likely to see a Py2.8 fork
appear: that's a much easier project.

--
Mathematics is the supreme nostalgia of our time.


_______________________________________________
Mercurial mailing list
[hidden email]
http://selenic.com/mailman/listinfo/mercurial