I've posted patches for:
#630: If float('123.45') works, so should numpy.float32('123.45')
#581: random.set_state does not reset state of random.standard_normal
James
On Fri, Apr 4, 2008 at 6:49 PM, Anne Archibald
<[EMAIL PROTECTED]> wrote:
> On 04/04/2008, Travis E. Oliphant <[EMAIL PROTECTED]> wrote:
>
> > Hey Anne,
> >
> > Do you currently have SVN access? Would you like it?
> >
> > I think the SciPy/NumPy sprint would be a good time to clean-up the
> > committers list and add new people interested in helping.
>
> I don't have SVN access. I'd be happy (and careful!) to have it, but I
> should warn you that I won't have time to do serious, regular
> development on scipy/numpy; I do hope to be able to write a little
> code here and there, though, and it would be handy to be able to add
> it directly instead of sending patches into the ether.
>
> Anne
>
>
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion@scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion
>
--- mtrand.pyx.old 2008-04-05 19:45:54.000000000 +0100
+++ mtrand.pyx 2008-04-05 19:46:39.000000000 +0100
@@ -37,6 +37,7 @@
ctypedef struct rk_state:
unsigned long key[624]
int pos
+ int has_gauss
ctypedef enum rk_error:
RK_NOERR = 0
@@ -552,6 +553,7 @@
raise ValueError("state must be 624 longs")
memcpy(<void*>(self.internal_state.key), <void*>(obj.data), 624*sizeof(long))
self.internal_state.pos = pos
+ self.internal_state.has_gauss = 0 # Clear out the gaussian value
# Pickling support:
def __getstate__(self):
--- arraytypes.inc.src.old 2008-04-05 19:15:57.000000000 +0100
+++ arraytypes.inc.src 2008-04-05 19:22:12.000000000 +0100
@@ -106,6 +106,19 @@
static double
MyPyFloat_AsDouble(PyObject *obj)
{
+ PyObject *tmp;
+ double d;
+ if (PyString_Check(obj)) {
+ tmp = PyFloat_FromString(obj, 0);
+ if (tmp) {
+ d = PyFloat_AsDouble(tmp);
+ Py_DECREF(tmp);
+ return d;
+ }
+ else {
+ return _getNAN();
+ }
+ }
if (obj == Py_None) return _getNAN();
return PyFloat_AsDouble(obj);
}
_______________________________________________
Numpy-discussion mailing list
Numpy-discussion@scipy.org
http://projects.scipy.org/mailman/listinfo/numpy-discussion
