[pypy-commit] pypy pypy-pyarray: (antocuni, rguillebert): we think it is fine to keep this as a dictionary, it doesn't seem that the lookup occours in performance-critical code
antocuni
noreply at buildbot.pypy.org
Sat Aug 31 16:09:05 CEST 2013
Author: Antonio Cuni <anto.cuni at gmail.com>
Branch: pypy-pyarray
Changeset: r66722:4df4713018c3
Date: 2013-08-31 16:08 +0200
http://bitbucket.org/pypy/pypy/changeset/4df4713018c3/
Log: (antocuni, rguillebert): we think it is fine to keep this as a
dictionary, it doesn't seem that the lookup occours in performance-
critical code
diff --git a/pypy/module/micronumpy/interp_dtype.py b/pypy/module/micronumpy/interp_dtype.py
--- a/pypy/module/micronumpy/interp_dtype.py
+++ b/pypy/module/micronumpy/interp_dtype.py
@@ -804,13 +804,6 @@
self.dtypes_by_name[alias] = dtype
self.dtypes_by_name[dtype.char] = dtype
- # shmuller 2013/07/22: Cannot find complex data types after conversion to
- # list! Solution: Keep in dictionary form.
- #self.dtypes_by_num = [dtype for dtype in
- # sorted(self.dtypes_by_num.values(), key=lambda dtype: dtype.num)
- # if dtype.num <= self.w_float64dtype.num]
- #assert len(self.dtypes_by_num) == self.w_float64dtype.num + 1
-
typeinfo_full = {
'LONGLONG': self.w_int64dtype,
'SHORT': self.w_int16dtype,
More information about the pypy-commit
mailing list