[pypy-commit] pypy numpy-full-fromstring: remaining review notes, almost there
alex_gaynor
noreply at buildbot.pypy.org
Thu Dec 15 08:56:17 CET 2011
Author: Alex Gaynor <alex.gaynor at gmail.com>
Branch: numpy-full-fromstring
Changeset: r50530:69e0e93f244d
Date: 2011-12-15 02:56 -0500
http://bitbucket.org/pypy/pypy/changeset/69e0e93f244d/
Log: remaining review notes, almost there
diff --git a/pypy/module/micronumpy/REVIEW b/pypy/module/micronumpy/REVIEW
--- a/pypy/module/micronumpy/REVIEW
+++ b/pypy/module/micronumpy/REVIEW
@@ -1,18 +1,6 @@
Review items
============
-Fixed
------
-* Rather than `True if len(sep_stripped) == 0 else False` just use:
- `len(sep_stripped) == 0`.
-* Rather than name the variable `A`, name it `items` or somsething like that.
-* Rather than using `ptr`, use `idx`, since it's not really a pointer.
-* Same comment about comparing array equality as before.
-* Rather than name the attribute `char`, name it `format_code`.
-* `default_fromstring` can do `self.box(-1.0)`, instead of the coerce thing.
-* Rather than doing a string format to raise an error (L67), use
- `operationerrfmt`.
-
Notes
-----
* Tests for both bool and long dtypes with this.
@@ -21,5 +9,9 @@
seem to install it to check to make sure my tests are correct.
There are already some tests for int64. Is that what you meant
by long types?)
-* No need for the `self.char == "?"` default.
- (jterrace: Does not translate without it)
\ No newline at end of file
+ * http://bpaste.net/show/20968/ is a log of me playing with bool `fromstring`
+ in 1.6, I can't seem to make it work with text parsing at all (or I'm an
+ idiot, I'm open to both interpretations), however binary parsing works so
+ maybe just do that for now.
+ * by long I mean `dtype(int)`, which is different from either
+ `dtype("int32")` or `dtype("int64")` (for reasons beyond my comprehension)
More information about the pypy-commit
mailing list