Mailman 3 January 2001 - Python-Dev

Sets: elt in dict, lst.include - really begs for a PEP
by Skip Montanaro 30 Jan '01

30 Jan '01

After reading through this thread and noticing (but not paying close attention to) all the related posts on c.l.py (subject: "in for dicts"), it seems to me that the whole "if/for something in dict" thing needds to be hashed out in a PEP. There were a fair amount of "Python's changing too fast" rants when 2.0 was released. Adding a major feature such as this at the 2.1 stage is only going to generate that many more rants. The fact that it was easy for Thomas to implement "if key in dict" doesn't make the overall concept less controversial. There are apparently lots of varying opinions about what's reasonable. This topic seems related to PEP 212 (Loop Counter Iteration) and PEP 218 (Adding a Built-In Set Object Type), but may well warrant its own. That said, I have plenty enough on my plate trying to keep Mojam afloat these days, so I can't step into the crevass, just observe that it looks to me like a very long ways to the bottom... ;-) Skip

4 4

re: Making mutable objects readonly
by Greg Wilson 30 Jan '01

30 Jan '01

> Barry Warsaw: > lock(x) ...? :) Greg Wilson: -1 --- everyone will assume it's mutual exclusion, rather than immutability.

1 0

Re: Python 2.1 slower than 2.0
by A.M. Kuchling 29 Jan '01

29 Jan '01

On Sat, 27 Jan 2001 18:28:02 +0100, Andreas Jung <andreas(a)andreas-jung.com> wrote: >Is there a reason why 2.1 runs significantly slower ? >Both Python versions were compiled with -g -O2 only. [CC'ing to python-dev] Confirmed: [amk@mira Python-2.0]$ ./python Lib/test/pystone.py Pystone(1.1) time for 10000 passes = 3.14 This machine benchmarks at 3184.71 pystones/second [amk@mira Python-2.0]$ python2.1 Lib/test/pystone.py Pystone(1.1) time for 10000 passes = 3.81 This machine benchmarks at 2624.67 pystones/second The ceval.c changes seem a likely candidate to have caused this. Anyone want to run Marc-Andre's microbenchmarks and see how the numbers have changed? --amk

8 13

Does autoconfig detect INSTALL incorrectly?
by Neil Schemenauer 29 Jan '01

29 Jan '01

Why is the configure.in file set to always use "install-sh"? There is a comment that says: # Install just never works :-( I don't think that statement is accurate. /usr/bin/install works quite well on my machine. The only commments I can find in the changelog are: revision 1.16 date: 1995/01/20 14:12:16; author: guido; state: Exp; lines: +27 -2 add INSTALL_PROGRAM and INSTALL_DATA; check for getopt and: revision 1.5 date: 1994/08/19 15:33:51; author: guido; state: Exp; lines: +14 -6 Simplify value of INSTALL (always 'cp'). Is there any reason why the autoconf macro AC_PROG_INSTALL is not used? The documentation seems to indicate that is does what we want. Neil

2 1

Function Hash: Check it in?
by Moshe Zadka 29 Jan '01

29 Jan '01

Attached is an example Python session after I patched the intepreter. The test-suite passes all right. I want an OK to check this in. Here is the patch: Index: Objects/funcobject.c =================================================================== RCS file: /cvsroot/python/python/dist/src/Objects/funcobject.c,v retrieving revision 2.33 diff -c -r2.33 funcobject.c *** Objects/funcobject.c 2001/01/25 20:06:59 2.33 --- Objects/funcobject.c 2001/01/27 19:13:08 *************** *** 347,358 **** 0, /*tp_print*/ 0, /*tp_getattr*/ 0, /*tp_setattr*/ ! (cmpfunc)func_compare, /*tp_compare*/ (reprfunc)func_repr, /*tp_repr*/ 0, /*tp_as_number*/ 0, /*tp_as_sequence*/ 0, /*tp_as_mapping*/ ! (hashfunc)func_hash, /*tp_hash*/ 0, /*tp_call*/ 0, /*tp_str*/ (getattrofunc)func_getattro, /*tp_getattro*/ --- 347,358 ---- 0, /*tp_print*/ 0, /*tp_getattr*/ 0, /*tp_setattr*/ ! 0, /*tp_compare*/ (reprfunc)func_repr, /*tp_repr*/ 0, /*tp_as_number*/ 0, /*tp_as_sequence*/ 0, /*tp_as_mapping*/ ! 0, /*tp_hash*/ 0, /*tp_call*/ 0, /*tp_str*/ (getattrofunc)func_getattro, /*tp_getattro*/ Python 2.1a1 (#1, Jan 27 2001, 21:01:24) [GCC 2.95.3 20010111 (prerelease)] on linux2 Type "copyright", "credits" or "license" for more information. >>> def foo(): ... pass ... >>> def bar(): ... pass ... >>> hash(foo) 135484636 >>> hash(bar) 135481676 >>> foo == bar 0 >>> d = {} >>> d[foo] =1 >>> def temp(): ... print "baz" ... >>> foo.func_code = temp.func_code >>> d[foo] 1 -- Moshe Zadka <sig(a)zadka.site.co.il> This is a signature anti-virus. Please stop the spread of signature viruses! Fingerprint: 4BD1 7705 EEC0 260A 7F21 4817 C7FC A636 46D0 1BD6

6 11

RE: Python-Dev digest, Vol 1 #1124 - 13 msgs
by Greg Wilson 29 Jan '01

29 Jan '01

> Greg Wilson writes: > > This would allow multi-dimensional structures > > (e.g. NumPy arrays) to do things like: > > for (i, j, k) in array: > > and: > > for ((i, j, k), v) in array: # three indices and value > Charles Waldman asks: > And what if I had, for example, a 3-dimensional array where the values > are 3-tuples? Would "for (i,j,k) in array" refer to the > indices or the values? Greg Wilson writes: That would be up to the module's implementer --- my idea was to have the 'for' loop provide more information to the object being iterated over, so that it could "do the right thing" (just as objects do right now with "x[i]"). Greg

1 0

Re: Re: Sets: elt in dict, lst.include
by Greg Wilson 29 Jan '01

29 Jan '01

> > > [Ping] > > > dict[key] = 1 > > > if key in dict: ... > > > for key in dict: ... > "Tim Peters" <tim.one(a)home.com> > "if (k, v) in dict" is clearly useless... > I can live with "x in list" checking the values and "x in dict" > checking the keys. But I can *not* live with "x in dict" equivalent > to "dict.has_key(x)" if "for x in dict" would mean "for x in dict.items()". > I also think that defining "x in dict" but not "for x in dict" will be > confusing. [Greg] Quick poll (four people): if the expression "if a in b" works, then all four expected "for a in b" to work as well. This is also my intuition; are there any exceptions in really existing Python? > [Guido] > for key in dict: ... # ... over keys > for key:value in dict: ... # ... over items [Greg] I'm probably revealing my ignorance of Python's internals here, but can the iteration protocol be extended so that the object (in this case, the dict) is told the number and type(s) of the values the loop is expecting? With: for key in dict: ... the dict would be asked for one value; with: for (key, value) in dict: the dict would be told that a two-element tuple was expected, and so on. This would allow multi-dimensional structures (e.g. NumPy arrays) to do things like: for (i, j, k) in array: # please give me three indices and: for ((i, j, k), v) in array: # three indices and value > [Guido] > for index:value in list: ... # ... over zip(range(len(list), list) How do you feel about: for i in seq.keys(): # strings, tuples, etc. "keys()" is kind of strange ("indices" or something would be more natural), *but* this allows uniform iteration over all built-in collections: def showem(c): for i in c.keys(): print i, c[i] Greg

2 1

What happened to Setup.local's functionality?
by Skip Montanaro 29 Jan '01

29 Jan '01

I just remembered Modules/Setup.local. I peeked at mine and noticed it had been zeroed out. I then copied a version of it over from another machine and reran make a couple times. Makesetup ran but nothing mentioned in Setup.local got built. I don't think 2.1 can be released without providing a way for users to recover from this change. I didn't see anything obvious in setup.py. Am I missing something? Skip

4 4

RE: [Python-Dev] problems building Python 2.1a1 on QNX 4.25
by Goodger, David 29 Jan '01

29 Jan '01

Marc-Andre Lemburg's patch to posixpath.py clears up the path problem. Thanks! MACHDEP is qnxJ for QNX 4.25, qnxG for QNX 4.23. I don't know what it is for QNX 6 (Neutrino). Perhaps test for MACHDEP[:3]=='qnx'? I'm still stuck at 'python setup.py build': unable to execute ld: no such file or directory running build running build_ext building 'struct' extension skipping //5/tmp/py/Python-2.1a1/Modules/structmodule.c (build/temp.qnx-J-PCI-2.1/structmodule.o up-to-date) ld build/temp.qnx-J-PCI-2.1/structmodule.o -L/usr/local/lib -o build/lib.qnx-J-PCI-2.1/struct.so error: command 'ld' failed with exit status 1 make: *** [sharedmods] Error 1 Armin Steinhoff said "QNX4 doesn't support dynamic loading". Is this compatible with distutils? If not, is there a workaround? Neil Schemenauer asked, "what should LDSHARED say for QNX?". I don't know. Python 2.0 compiled OK, and its makefile says LDSHARED=ld. However, Modules/Setup has no uncommented "*shared*" line. Those of us who rely on Python to get our work done, and who don't have the bandwidth for the implementation complexities, owe a lot to everyone who makes it possible to compile Python out-of-the-box. Very much appreciated. Thank you! David Goodger Systems Administrator & Programmer, Advanced Systems Automation Tooling Systems Inc., Automation Systems Division direct: (519) 653-4483 ext. 7121 fax: (519) 650-6695 e-mail: dgoodger(a)atsautomation.com

2 1

problems building Python 2.1a1 on QNX 4.25
by Goodger, David 29 Jan '01

29 Jan '01

[CC'ing to Armin Steinhoff, who maintains pyqnx on SourceForge.] I'm having trouble building Python 2.1a1 on QNX 4.25. Caveat: my C is very rusty (long live Python!), I don't know my way around configure, and am not familiar with Python's Makefile. Python 2.0 compiled fine (with a couple of tweaks), but I'm getting caught by the new way of building things. Please help if you can! Many thanks in advance. Here's an excerpt of my efforts: # cd /tmp/py # gunzip -c < python-2.1a1.tgz | tar -rf - # cd Python-2.1a1 # ./configure 2>&1 | tee ../configure.1 # make 2>&1 | tee ../make.1 ... ./python //5/tmp/py/Python-2.1a1/setup.py build 'import site' failed; use -v for traceback Traceback (most recent call last): File "//5/tmp/py/Python-2.1a1/setup.py", line 4, in ? import sys, os, string, getopt ImportError: No module named string Running ./python results in stack overflow. The old QNX instructions in README recommend editing Modules/Makefile: LDFLAGS= -N 64k # make 2>&1 | tee ../make.2 Same error as first make. But now the stack doesn't overflow. # python 'import site' failed; use -v for traceback Python 2.1a1 (#2, Jan 26 2001, 11:38:55) [C] on qnxJ Type "copyright", "credits" or "license" for more information. >>> import sys >>> sys.path ['', '/usr/local/lib/python', '/home/dgoodger/lib/python', '/5/tmp/py/Python-2.1a1/Lib', '/5/tmp/py/Python-2.1a1/Lib/plat-qnxJ', '/tmp/py/Python-2.1a1/Modules'] >>> ^D # fullpath . . is //5/tmp/py/Python-2.1a1 The QNX node number prefix '//5' (machine or host number, equivalent to a 'hostname:' prefix for network paths) is being reduced somehow (path normalization?) to '/5', so paths don't resolve. 2 slashes ('//') are required at the head of the path. Is this something that can be fixed? I added a prefix (QNX virtual-to-real path mapping on the filesystem tree) to correct this: # prefix -A /5=//5 Now /5 points to //5, similar to a link. # make 2>&1 | tee ../make.3 ... ./python //5/tmp/py/Python-2.1a1/setup.py build unable to execute ld: No such file or directory running build running build_ext building 'struct' extension creating build creating build/temp.qnx-J-PCI-2.1 cc -O -I. -I/5/tmp/py/Python-2.1a1/./Include -IInclude/ -I/usr/local/include -c /5/tmp/py/Python-2.1a1/Modules/structmodule.c -o build/temp.qnx-J-PCI-2.1/structmodule.o creating build/lib.qnx-J-PCI-2.1 ld build/temp.qnx-J-PCI-2.1/structmodule.o -L/usr/local/lib -o build/lib.qnx-J-PCI-2.1/struct.so error: command 'ld' failed with exit status 1 make: *** [sharedmods] Error 1 QNX doesn't have an 'ld' command. Is configure not getting its info to setup.py? (Is it supposed to?) What should I check? I have logs of each of the configure & make runs. Should I submit this as a bug on SourceForge? Hope to hear from somebody soon. David Goodger Systems Administrator & Programmer, Advanced Systems Automation Tooling Systems Inc., Automation Systems Division direct: (519) 653-4483 ext. 7121 fax: (519) 650-6695 e-mail: dgoodger(a)atsautomation.com

8 12