Mailman 3 October 2002 - Python-Dev

RE: [Python-Dev] test_pep277 vs Win98
by Delaney, Timothy Oct. 8, 2002

Oct. 8, 2002

> From: martin(a)v.loewis.de [mailto:martin@v.loewis.de] > > Tim Peters <tim.one(a)comcast.net> writes: > > > > Furthermore, how would I implement > os.path.supports_unicode_filenames > > > for ntpath? It seems that something must be exported from os. > > > > This sounds like, for now, whether a platform supports this > or not is a > > matter of hand-forcing a config symbol, where only a few > Windows flavors are > > currently justified in saying "yup, works here". > > Repeating my question: How can I hand-force a config symbol to True on > NT+ and to False on W9x? Sounds like an exceptional situation to me. try: try to use problem filename(s) yup_works_here = True except IOError: # or whatever is appropriate yup_works_here = False The test cases then simply test that it works for all cases Tim Delaney

1 0

Re: [snake-farm] Autodetect /usr/ccs/bin?
by Guido van Rossum Oct. 7, 2002

Oct. 7, 2002

[Anders Qvist] > Several machines in the snake farm fail to build libpython2.3 because > they don't have ar in their path. ar, ld and the like live in > /usr/ccs/bin, which seems to be some kind of standard (for generous > values of standard) as both SunOS 5.8, AIX 4.3 and HP/UX 11.00 seems > ot have this dir. > > Should the configure script try with /usr/ccs/bin/ar and ld if none is > explicitly supplied? This may be a good idea, but I don't know zit about this area. I'm cc'ing python-dev, maybeo someone there understands this. --Guido van Rossum (home page: http://www.python.org/~guido/)

4 3

Unclear on the way forward with unsigned integers
by Mark Hammond Oct. 7, 2002

Oct. 7, 2002

I'm a little confused by the new world order for working with integers in extension modules. At the end of the day, my question is this: Assume my extension module has an unsigned integer it wishes to return to Python. Further, assume that this unsigned integer is not really an integer as such, but more a set of bits, or some other value "encoded" in 32 bits, such as an enum. (To put it another way, the "signedness" of this value seems more random than chosen) How should I create the object to return to Python? For a concrete example. Let's say I want to return the value of the Win32 function GetVersion(). The documentation for this function declares it is an unsigned 32 bit value. The documentation then explains that to decode this value, specific bits in the value should be examined. It then expounds on this with C sample code that relies on this unsigned behaviour by using a simple "> 0x80000000" comparison to check the high bit! I see 2 choices for returning this value: * Use PyInt_FromLong() - this will give me a *signed* Python integer, but with an identical bit pattern. * Use PyLong_FromUnsignedLong() - this function will correctly be signed, but may no longer fit in 32 bits. Now, I think I am trying to stay too close to the hardware for a language like Python, but something just seems wrong with promoting my nice 32 bit value to a Python long, simply for the sake of retaining the sign for a value that the whole concept of "signed" doesn't make much sense (as it doesn't in this case, or in the case of enums etc). Any suggestions or general advice? While this case seems quite trivial, I am starting to face this issue more and more, especially as I am seeing these lovely "FutureWarnings" from all my lovely 32 bit hexadecimal constants <wink/frown> Thanks, Mark.

6 11

Python 2.2.2 beta 1 docs finalized
by Fred L. Drake, Jr. Oct. 7, 2002

Oct. 7, 2002

I've posted the Python 2.2.2b1 docs online; please report any (new) problems via SourceForge with a priority of at least 6. I'd like to thank Raymond for putting so much effort into the documentation for 2.2.2; he's really doing a great job! Thanks! -Fred -- Fred L. Drake, Jr. <fdrake at acm.org> PythonLabs at Zope Corporation

1 0

Re: PEP239 (Rational Numbers) Reference Implementation and new issues
by Paul F Dubois Oct. 7, 2002

Oct. 7, 2002

The question was raised as to why the Numeric community really wanted the literal notation 1.0+2.0j for complex numbers. The motivation was not speed particularly; nobody is going to construct a large complex array out of literals. It is true that inside a loop a constant might be a problem but you could always lift it by hand. I'd say rather that the real motivation was ease of learning for those who knew similar languages. There were essentially three ways this was done in other languages that our users would already grok: (1.0, 2.0) -- Fortran; a non-starter for Python 1.0 + 2.0i or 1.0 + 2.0j -- most (all?) existing tools for similar purposes use one of these. We all felt that Python would be perceived as "inferior" to these other programs if you had to do something clumsy. I haven't followed the rational discussion but obviously there is no other long-standing and highly used notation for rationals except 1.0/2.0, is there?

2 1

Re: [Python-Dev] Python 2.2.2 beta release on Monday!
by Michael Hudson Oct. 7, 2002

Oct. 7, 2002

Guido van Rossum <guido(a)python.org> writes: > I plan to do a Python 2.2.2 beta release this Monday (if Fred's and > Tim's time is available to do their part). > > Please finish your changes to 2.2.2 before the weekend is over! Timezones to the rescue... there are definitely a few things I want to do before a beta release. there-goes-the-coffee-break-ly y'rs M. -- In case you're not a computer person, I should probably point out that "Real Soon Now" is a technical term meaning "sometime before the heat-death of the universe, maybe". -- Scott Fahlman <sef(a)cs.cmu.edu>

1 0

PEP 282 Implementation
by Vinay Sajip Oct. 7, 2002

Oct. 7, 2002

I've uploaded my logging module, the proposed implementation for PEP 282, for committer review, to the SourceForge patch manager: http://sourceforge.net/tracker/index.php?func=detail&aid=578494&group_id=547 0&atid=305470 I've assigned it to Mark Hammond as (a) he had posted some comments to Trent Mick's original PEP posting, and (b) Barry Warsaw advised not assigning to PythonLabs people on account of their current workload. The file logging.py is (apart from some test scripts) all that's supposed to go into Python 2.3. The file logging-0.4.6.tar.gz contains the module, an updated version of the PEP (which I mailed to Barry Warsaw on 26th June), numerous test/example scripts, TeX documentation etc. You can also refer to http://www.red-dove.com/python_logging.html Here's hoping for a speedy review :-) Regards, Vinay Sajip

5 12

Patch to fix Distutils bdist_rpm on Red Hat 8.0
by Sean Reifschneider Oct. 7, 2002

Oct. 7, 2002

At the 8.0 release, Red Hat has split the "rpm" command such that the build options reside only in the "rpmbuild" command. Actually, this happened a while ago, as "rpmbuild" was available on 7.3. However, at 8.0 the "rpm" command would no longer do building and it fails out. I've submitted patch #619493 to correct this. If /usr/bin/rpmbuild or /bin/rpmbuild exists, the rpm command is changed from "rpm" to "rpmbuild". This probably should go into the 2.2.2 release. I'm submitting a bug to Red Hat as well, as they probably should release a fixed python package to address this as well. Sean -- "The phrase ``ship it!'' is one with long and deep resonances in my benighted and antiquated big-software-dept career." -- John Shipman, 1998 Sean Reifschneider, Inimitably Superfluous <jafo(a)tummy.com> tummy.com, ltd. - Linux Consulting since 1995. Qmail, Python, SysAdmin

1 0

Anders Qvist: Re: [snake-farm] Autodetect /usr/ccs/bin?
by Guido van Rossum Oct. 6, 2002

Oct. 6, 2002

I've no time to research this further right now, so I'm forwarding Anders' mail here. It seems that Martin disagrees with Pike's solution though. --Guido van Rossum (home page: http://www.python.org/~guido/) ------- Forwarded Message Date: Sun, 06 Oct 2002 22:03:57 +0200 From: Anders Qvist <quest(a)lysator.liu.se> To: Guido van Rossum <guido(a)python.org> Subject: Re: [snake-farm] Autodetect /usr/ccs/bin? On Sun, Oct 06, 2002 at 09:29:24AM -0400, Guido van Rossum wrote: > [Anders Qvist] > > Several machines in the snake farm fail to build libpython2.3 because > > they don't have ar in their path. ar, ld and the like live in > > /usr/ccs/bin, which seems to be some kind of standard (for generous > > values of standard) as both SunOS 5.8, AIX 4.3 and HP/UX 11.00 seems > > ot have this dir. > > > > Should the configure script try with /usr/ccs/bin/ar and ld if none is > > explicitly supplied? > > This may be a good idea, but I don't know zit about this area. I'm > cc'ing python-dev, maybeo someone there understands this. Apparently, the pike people do this. They seem to think it's the right thing to do. From their configure.in: case "$pike_cv_sys_os" in ... Solaris) LDSHARED="/usr/ccs/bin/ld -G";; ... They also add it to their PATH during various circumstances. Have a look at: http://community.roxen.com/_internal/cvsview!0/8223/1.633/0/configure.in - -- Anders "Quest" Qvist "We've all heard that a million monkeys banging on a million typewriters will eventually reproduce the entire works of Shakespeare. Now, thanks to the Internet, we know this is not true." -- Robert Wilensky ------- End of Forwarded Message

3 3

C3 implementation
by Samuele Pedroni Oct. 6, 2002

Oct. 6, 2002

I have an apparently working implementation of C3 in C. test_descr/test_descrtut pass apart some tests that involves hierarchies with MRO order disagreements. In such cases my impl raise TypeError: MRO order disagreement exceptions. Some of them it's clear to me how to fix them myself, for others someone should think about what was the original intention of the test. I can add some tests about monotonicity etc, with examples from the paper and the discussion we had here. Should I submit this as a patch on SF? at least is a start. PS: 1. the implementation is worst-case quadratic in the total length of the superclasses' mros, like the Python one I have presented here or the one used in Goo. I don't think this has relevance in practice, it's a trade-off versus even more complicated logic and data structure usage.

2 1