Procedure to request adding a module to the standard library - or initiating a vote on it

[Helmut Jarausch]

Hi,

I'd like to request adding the module

http://pypi.python.org/pypi/regex

to Python's standard library in the (near) future or to even replace the
current 're' module by it.

Personally I'm in need for fuzzy regular expressions and I don't see how
to do this easily and efficiently without this module.

For a long term project I also need some "guarantee" that this
functionality will exist in future.

So, is there a (formal) procedure for such a request or for initiating
some sort of vote on it?

I know there is a "Benevolent Dictator" for Python.
Should I try to contact him personally?

Many thanks for a hint,

Helmut Jarausch

RWTH Aachen University
Germany

[Peter Otten]

Helmut Jarausch wrote:

> I'd like to request adding the module
>
> http://pypi.python.org/pypi/regex
>
> to Python's standard library in the (near) future or to even replace the
> current 're' module by it.
>
> Personally I'm in need for fuzzy regular expressions and I don't see how
> to do this easily and efficiently without this module.
>
> For a long term project I also need some "guarantee" that this
> functionality will exist in future.

There has been a discussion about that particular module, and it would have
gone into 3.3 if it were not for the low "bus count" (only the author having
working knowledge of the code).

> So, is there a (formal) procedure for such a request or for initiating
> some sort of vote on it?
>
> I know there is a "Benevolent Dictator" for Python.
> Should I try to contact him personally?

I don't think that will help. From PEP 408:

"""
As part of the same announcement, Guido explicitly accepted Matthew
Barnett's 'regex' module [4] as a provisional addition to the standard
library for Python 3.3 (using the 'regex' name, rather than as a drop-in
replacement for the existing 're' module).
"""

[Mark Lawrence]

My understanding is that the main stumbling block is there's no one
committed to being the maintainer of the code. If that commitment is
not made, and I belive it's required to be several (10?) years, then the
code may never get into the standard library. I'll happily stand
corrected if anyone knows any better.

--
Cheers.

Mark Lawrence.

[Helmut Jarausch]

On Tue, 07 Aug 2012 13:15:29 +0200, Peter Otten wrote:

> I don't think that will help. From PEP 408:
>
> """
> As part of the same announcement, Guido explicitly accepted Matthew
> Barnett's 'regex' module [4] as a provisional addition to the standard
> library for Python 3.3 (using the 'regex' name, rather than as a drop-in
> replacement for the existing 're' module).
> """

What is a "provisional addition"? Python-3.3 (20120708) doesn't have
such a module.

Many thanks,
Helmut.

[Mark Lawrence]

This is a new concept, as an example see
http://bugs.python.org/issue14814#msg164795 which refers to the
implementation of PEP 3144, the ipaddress module.

--
Cheers.

Mark Lawrence.

[Peter Otten]

Helmut Jarausch wrote:

> On Tue, 07 Aug 2012 13:15:29 +0200, Peter Otten wrote:
>
>> I don't think that will help. From PEP 408:
>>
>> """
>> As part of the same announcement, Guido explicitly accepted Matthew
>> Barnett's 'regex' module [4] as a provisional addition to the standard
>> library for Python 3.3 (using the 'regex' name, rather than as a drop-in
>> replacement for the existing 're' module).
>> """
>
> What is a "provisional addition"?

My understanding is that Guido agreed for it to be added under a different
name (i. e. as regex, not re) and with no guarantees that it will remain in
Python versions above 3.3.

> Python-3.3 (20120708) doesn't have such a module.

Indeed. Guido allowed adding the module, but some core devs remained
skeptical and no one put in the legwork to add it. It may still be added in
3.4 if a core developer is willing to familiarise with the codebase and
promises to help with maintenance, but the chances seem low by now.

[Terry Reedy]

On 8/7/2012 6:13 AM, Helmut Jarausch wrote:

> I'd like to request adding the module
> http://pypi.python.org/pypi/regex
> to Python's standard library in the (near) future

As near as I can tell, the author is lukewarm about the prospect.

To respond the general question:

The author of a module should be warm to hot about the idea and must be
willing to move development into the stdlib source tree and conform to
Python's release schedule, which may be too slow for actively developed
modules.

(If the module wraps a well-known and maintained external C library, the
wrapper must go into the source tree. Then the docs says that the stdlib
module wraps something we are not responsible for.)

There should be community support for the module as one of the best of
its kind.

Someone has to write a PEP. There must be developer support to review
the code, api, and documentation. Author must allow changes. (The new
ipaddress module has had changes to all three, including considerable
doc expansion. Some were to make it accessible to beginners rather than
only ip experts, others to make it conform to current 3.x stdlib
standards and best practices. For instance, 2.x style list returns were
changed to 3.x style iterator returns. )

There must be commitment for the author or substitute for maintenance.

> For a long term project I also need some "guarantee" that this
> functionality will exist in future.

That is the point of the last requirement.

>
> So, is there a (formal) procedure for such a request or for initiating
> some sort of vote on it?

'voting' is fuzzy. Community support. Some support from developers. Best
no strong opposition from a senior core developer, or at least more that
one. Final decision is always by GvR, but he often delegates decisions
to other developers, especially in an area of his non-expertise.

> I know there is a "Benevolent Dictator" for Python.
> Should I try to contact him personally?

No. If there is author and community support, the next step is a PEP or
discussion on python-ideas list (which Guido reads even if he does not
write much).

--
Terry Jan Reedy