Mailman 3 April 2001 - Python-Dev

ineffective optimization: method tables
by Neil Schemenauer 23 Apr '01

23 Apr '01

Well, I wasted a fair amount of my time for no apparent gain. The idea was to have function pointer tables indexed by type that could be used for common operations. First, how to do we index things by type? Here's my solution: #define PyType_UNKNOWN 0 #define PyType_NONE 1 #define PyType_INT 2 #define PyType_Ord(t) ((t)->tp_ord) #define PyObject_TypeOrd(o) PyType_Ord((o)->ob_type) Here is an example of methods for PyObject_IsTrue: int int_istrue(PyObject *v) { return ((PyIntObject *)v)->ob_ival != 0; } int none_istrue(PyObject *v) { return 0; } inquiry istrue_table[] = { PyObject_IsTrue, /* PyType_UNKNOWN */ none_istrue, /* PyType_NONE */ int_istrue, /* PyType_INT */ }; #define PyObject_IS_TRUE(v) istrue_table[PyObject_TypeOrd(v)](v) There is a patch at: http://arctrix.com/nas/python/method_table1.diff I did have 2-D tables for binary operations but since they were quite sparse I took them out in favor of arrays and case statements. Unfortunately, all my benchmarks show this patch to be ineffective in terms of speeding up the interpreter. Anyone know why? Neil

2 1

string slicing and method consistency
by Greg Wilson 23 Apr '01

23 Apr '01

One of the students in my class at the Space Telescope Science Institute ("Hubble R Us") last week brought up an interesting point: if "abbc"[-1] is "c", and if "abbc".replace("b", "x", 1) is "axbc", then shouldn't "abbc".replace("b", "x", -1) be "abxc" (i.e. negative numbers replace the *last* occurrences of the value)? Same argument for "split", etc. Turns out that "abbc".replace("b", "x", -1) is "axxc" (i.e. negative arguments are ignored). I would have expected this to raise a ValueError, if anything. Is there a reason for this behavior? Is it worth making replace, split, etc. interpret negative indices in the same way as indexing does? Thanks, Greg

4 4

Importing DLLs on Windows
by M.-A. Lemburg 23 Apr '01

23 Apr '01

Python or Windows seems to have trouble importing DLLs when inside packages. I'm working on an extension which pulls in a DLL gmp31.dll which lives inside a package dir mx/Number/mxNumber along side the Python wrapper extension mxNumber.pyd. Currently, I'm using this code in the subpackage's __init__.py: # On Windows we also distribute the GMP DLL (in the win32 subdir). To # have Win32 find the DLL we need to be explicit about the path in # sys.path though. This trick works for all startup directories # *except* \PythonXX (for some reason this fails), but is not thread # safe... import sys, os if sys.platform[:3] == 'win': abspath = os.path.abspath(__path__[0]) sys.path.insert(0, abspath) from mxNumber import * sys.path.remove(abspath) else: from mxNumber import * I don't have any clue why the import works from everywhere *except* the Python21 install directory... anyone does ? Thanks, -- Marc-Andre Lemburg ______________________________________________________________________ Company & Consulting: http://www.egenix.com/ Python Pages: http://www.lemburg.com/python/

3 6

Suirprise!
by Tim Peters 23 Apr '01

23 Apr '01

I just stared at this a long time: >>> 'a' in 'a' # fine 1 >>> 'a' in 'a' == 1 # what? 0 >>> 'a' in 'b' # fine 0 >>> 'a' in 'b' == 0 # what? 0 >>> It's "correct". I've been using Python longer than Guido <wink>, and I'm amazed this is the first time I got bit by this! Here's a hint: >>> 'a' in 'a' == 'a' 1 >>> thank-god-for-dis.dis-ly y'rs - tim

6 10

Re: [Python-checkins] CVS: python/dist/src/Doc/ref ref3.tex,1.64,1.65 ref5.tex,1.43,1.44
by Jeremy Hylton 23 Apr '01

23 Apr '01

>>>>> "GvR" == Guido van Rossum <gvanrossum(a)users.sourceforge.net> writes: GvR> Log Message: Implement, test and document "key in dict" and GvR> "key not in dict". GvR> I know some people don't like this -- if it's really GvR> controversial, I'll take it out again. (If it's only Alex GvR> Martelli who doesn't like it, that doesn't count as "real GvR> controversial" though. :-) I don't like it either, because it's only clear when it's spelled "for key in dict". It's pretty confusing when it's spelled "for val in dict" or even "for x in dict". But I'm willing to give it a try for a while. Jeremy

2 1

Re: [Pythonmac-SIG] Re: [Python-Dev] Import hook to do end-of-line conversion?
by Jack Jansen 23 Apr '01

23 Apr '01

> [Steven D. Majewski] > > Has anybody looked at sfio ? > > ... > > http://www.research.att.com/sw/tools/sfio/ > > Did just now. Only runs on Unix boxes, so would be a heavyweight way to > solve line-end problems across platforms that don't have any <wink>. But purely by chance I know that Matthias Neeracher, who has written the GUSI unix-emulation package that MacPython uses to do socket IO and select and such, is an SFIO fan, and there's all sorts of hooks in GUSI to allow use of SFIO. So I think that there's a good chance that sfio is okay for MacPython. I've just dug out the 1991 usenix article, I'll read through it shortly. -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen(a)oratrix.com | ++++ if you agree copy these lines to your sig ++++ www.oratrix.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm

1 0

re: string slicing and method consistency
by Greg Wilson 22 Apr '01

22 Apr '01

> > > Greg Wilson: > > > if "abbc"[-1] is "c", and if > > > "abbc".replace("b", "x", 1) is "axbc", then shouldn't > > > "abbc".replace("b", "x", -1) be "abxc" (i.e. negative > > > numbers replace the *last* occurrences of the value)? > > > Same argument for "split", etc. > > >>> path = "/some/long/path/to/file.html" > > >>> main, parent, file = path.split("/", -2) > > >>> main > > "/some/long/path" > > >>> parent > > "to" > > >>> file > > "file.html" > > Guido van Rossum: > OK, that's an example. It's only so-so, because you should be using > os.path.split() anyway. It's done best as follows: > > temp, file = os.path.split(path) > main, parent = os.path.split(temp) Greg Wilson: Or "main, parent, file = os.path.split(path, -2)" :-) > > Greg Wilson again: > > Question still stands --- if these are counts, then shouldn't > > negative values raise exceptions? > > Given that it's documented with the name "maxsplit", it's not > unreasonable that -1 is treated the same as 0. Greg Wilson: But it isn't: >>> print sys.version 2.2a0 (#2, Apr 20 2001, 12:53:03) [GCC 2.95.2 19991024 (release)] >>> "abbc".replace("b", "x", 0) 'abbc' >>> "abbc".replace("b", "x", -1) 'axxc' Thanks, Greg

1 0

[maintenance doc updates]
by fdrake＠beowolf.digicool.com 22 Apr '01

22 Apr '01

The development version of the documentation has been updated: http://python.sourceforge.net/maint-docs/ Third attempt to push maintenance docs to the SourceForge site. Sheesh!

1 0

[maintenance doc updates]
by fdrake＠beowolf.digicool.com 22 Apr '01

22 Apr '01

The development version of the documentation has been updated: http://python.sourceforge.net/maint-docs/ Second attempt to push maintenance docs to the SourceForge site.

1 0

[maintenance doc updates]
by fdrake＠beowolf.digicool.com 22 Apr '01

22 Apr '01

The development version of the documentation has been updated: http://python.sourceforge.net/maint-docs/ First attempt to push maintenance docs to the SourceForge site.

1 0