expensive and overly general PyObject_IsInstance(), call
PyObject_TypeCheck() which is a macro that often avoids a call, and if
it does make a call, calls the much more efficient PyType_IsSubtype().
This saved 6% on a benchmark for slot lookups.
This is a silly workaround for a rather serious bug in MacOSX: if you take
a long filename and convert it to an FSSpec the fsspec gets a magic
cooky (containing a #, indeed). If you then massage the extension of this
fsspec and convert back to a pathname you may end up referring to the
same file. This could destroy your sourcefile. The problem only occcurs
in MacPython-OS9, not MacPython-OSX (I think).
Closes bug #505562.
Jeremy reported that this is not allowed by RFC 2396; however,
other tools support unescaped @'s so we should also.
Apply SF patch 596581 closing bug 581529.
- update DLL version number
- add files required for 2.3 (no changes to modules though)
- restructure build of pgen.exe
NOTE: As I don't have the VACPP compiler, these changes are untested.
Apart from slightly re-ordering some file lists, and matching file name
casing, I believe these changes are the minimum necessary to build 2.3
with VACPP.
- the security fixes to tempfile have lead to test_tempfile wanting
to create 100 temporary files. as the EMX default is only 40,
the number of file handles has been bumped (up to 250).
- changes to pgen have required restructuring its build support.
Also, don't call gettempdir() in the default expression for the 'dir'
argument to various functions; use 'dir=None' for the default and
insert 'if dir is None: dir = gettemptir()' in the bodies. That way
the work done by gettempdir is postponed until needed.
Also, don't handle METH_OLDARGS on the fast path. All the interesting
builtins have been converted to use METH_NOARGS, METH_O, or
METH_VARARGS.
Result is another 1-2% speedup. If I can cobble together 10 of these,
it might make a difference.
This makes the code much easier to ready, because it is at a sane
indentation level. On my box this shows a 1-2% speedup, which means
nothing, except that I'm not going to worry about the performance
effects of the change.
-- replace then with slightly faster PyObject_Call(o,a,NULL). (The
difference is that the latter requires a to be a tuple; the former
allows other values and wraps them in a tuple if necessary; it
involves two more levels of C function calls to accomplish all that.)
nothing special done if keyword arguments were present, so test for
that earlier and fall through to the normal case if there are any.
This ought to slow down CFunction calls with keyword args, but I don't
care; it's a tiny (1%) improvement for pystone.
- Use PyObject_Call() instead of PyEval_CallObject(), saves several
layers of calls and checks.
- Pre-allocate the argument tuple rather than calling Py_BuildValue()
each time round the loop.
- For filter(None, seq), avoid an INCREF and a DECREF.
Loosened the acceptable 'start' and 'stop' arguments so that any
Python (bounded) ints can be used. So, e.g., randrange(-sys.maxint-1,
sys.maxint) no longer blows up.
warning for 'global None', but that's either accompanied by an
assignment to None, which will trigger a warning, or not, in which
case it's harmless. :-)
rigorous instead of hoping for testing not to turn up counterexamples.
Call me heretical, but despite that I'm wholly confident in the proof,
and have done it two different ways now, I still put more faith in
testing ...
[ 587993 ] SET_LINENO killer
Remove SET_LINENO. Tracing is now supported by inspecting co_lnotab.
Many sundry changes to document and adapt to this change.
Push /usr/... further down the list - always check /usr/local/... before
/usr/...
Doubt this will help with http://python.org/sf/589427 or not, but these
changes were prompted by my investigation of that bug report. I wasn't able
to reproduce that problem though
