python-dev summary: Sept. 1-15

A.M. Kuchling
Tue, 19 Sep 2000 21:19:53 -0400

Python-dev summary, September 1-15, 2000

Released at last!

Python 1.6final was released on September 5.

Python 2.0beta1 was released, also on September 5.  GvR's announcement:
"A unique event in all the history of Python: two releases on the same day!"
Python 2.0 home page:

Remember, the "What's New in Python 2.0" article goes into detail on 2.0's 
changes and new features.

Current plans are for a beta2 release around September 26th, and 
issue the final release two weeks later on October 10th.

Squabbling over the meaning of "print >>None" occupied some time; this
is the same as "print >>sys.stdout".  Vladimir Marangozov pointed out
that this feels wrong: "In Python, this form surely looks & feels like
the Unix cat /dev/null, that is, since None doesn't have a 'write'
method, the print statement is expected to either raise an exception
or be specialized for None to mean "the print statement has no
effect". The deliberate choice of sys.stdout is not obvious."

The BDFL's reaction: "You don't have to make a strong association with
sys.stdout.  When the file expression is None, the whole ">>file, "
part disappears!"


The licensing squabble become highly visible outside the community,
stemming from a debian-devel posting.  The basic problem is that one
term in CNRI's 1.6 license specifies that the contract should be
interpreted using Virginia law, because that's where CNRI is located.
The FSF's lawyer believes that this term is incompatible with the GPL,
because the GPL doesn't allow adding restrictions, and this counts as
a restriction.  

GvR's posting: "At this point, our best shot seems to be to arrange a
meeting between CNRI's lawyer and Stallman's lawyer.  Without the
lawyers there, we never seem to be able to get a commitment to an
agreement.  CNRI is willing to do this; Stallman's lawyer (Eben
Moglen; he's a law professor at Columbia U, not NYU as I previously
mentioned) is even harder to get a hold of than Stallman himself, so
it may be a while."

GvR's reaction:


Other matters

XML support for 2.0 poses a problem, because it adds a package named
'xml' to the standard library, but this package name was already in
use by the package being developed by the XML-SIG.  There's a hack in
Lib/xml/ that tries to delegate imports to an _xmlplus
package if one is installed, but it doesn't work:

Discussion continued for a while, but no clear resolution to the
problem has been made.  Possible fixes might be to use __import__
instead of the imp module, renaming the packages to xmlbase and xml,
or dropping the xml/ package from Python 2.0 completely.

An long-standing error in the docs for the re module was found by Paul
Prescod.  The docs claimed that the .start() and .end() methods of
MatchObjects return None when the group didn't participate in the
match, but the code actually returns -1.  The fix was to change the
docs to match the code, not the other way around.  GvR: "I'd suggest
fixing SRE and the docs, because -1 is a more useful indicator for 'no
match' than None: it has the same type as valid indices.  It makes it
easier to adapt to static typing later."

Greg Ward wrote a Distutils setup script for IDLE:

And the usual pile of minor fixes.  

Related Links

Temporary archive of python-dev summaries:

Python-dev archives:

Python project page on SourceForge:

Python Enhancement Proposals (PEPs) summarize recent and past
proposals for enhancements:

The python-dev summaries are intended to inform the larger Python
community of currently ongoing development.  To comment on material in
this python-dev summary, simply post to comp.lang.python or
<>.  Give your posting a meaningful subject
line, and if it's about a PEP, include the PEP number (e.g. Subject:
PEP 201 - Lockstep iteration) All python-dev members are interested in
seeing ideas discussed by the community, so don't hesitate to take a
stance on a PEP if you can share an informed opinion.