Hi Michael,
You wrote:
>
> It is, but I forgot to mention it in the other thread. The ticket has
> been closed, but it would be great if you could submit the cleanup
> patch for #217 in the next 36 hours.
I submitted a patch bundle on trac #217, which changed back some substitutions
Py_ssize_t
On Oct 31, 9:55 pm, Jaap Spies <[EMAIL PROTECTED]> wrote:
> Hi Michael,
>
>
>
> > it should also be noted that the bug Carl mention above seems to be
> > the root cause for #973. I am currently updating to 2.8.10 on my local
> > box (which shows the segfault for dance(10)) to see if the problem
Hi Michael,
>
> it should also be noted that the bug Carl mention above seems to be
> the root cause for #973. I am currently updating to 2.8.10 on my local
> box (which shows the segfault for dance(10)) to see if the problem is
> really fixed. I added some additional info to #973 about this.
>
On Oct 30, 2007, at 7:16 PM, mabshoff wrote:
> On Oct 31, 3:11 am, Carl Witty <[EMAIL PROTECTED]> wrote:
>> On Oct 29, 10:11 am, Robert Bradshaw <[EMAIL PROTECTED]>
>> wrote:
>
> Hello,
>
>>
>>> There is another way that Py_ssize_t differs from int, namely that
>>> Cython casting happens via the
On Oct 31, 3:11 am, Carl Witty <[EMAIL PROTECTED]> wrote:
> On Oct 29, 10:11 am, Robert Bradshaw <[EMAIL PROTECTED]>
> wrote:
Hello,
>
> > There is another way that Py_ssize_t differs from int, namely that
> > Cython casting happens via the __index__ rather than __int__
> > function. __int__ i
On Oct 29, 10:11 am, Robert Bradshaw <[EMAIL PROTECTED]>
wrote:
> There is another way that Py_ssize_t differs from int, namely that
> Cython casting happens via the __index__ rather than __int__
> function. __int__ is potentially bad because it tries to hard to cast
> to an int (e.g. truncating,
William Stein wrote:
> On 10/29/07, Robert Bradshaw <[EMAIL PROTECTED]> wrote:
>> On Oct 28, 2007, at 4:29 PM, William Stein wrote:
>>> On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
What is the correct/recommended use of Py_ssize_t in Cython code?
>>> You should use Py_ssize_t *anywhere*
On Oct 29, 2007, at 10:16 AM, William Stein wrote:
>>> Note that it is not just size_t from C, since it is signed,
>>> since in Python list indices can be negative (in which case counting
>>> starts from the right side of the list).
>>
>> I think it is actually signed so that errors can be return
On 10/29/07, Robert Bradshaw <[EMAIL PROTECTED]> wrote:
> On Oct 28, 2007, at 4:29 PM, William Stein wrote:
> > On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
> >> What is the correct/recommended use of Py_ssize_t in Cython code?
> >
> > You should use Py_ssize_t *anywhere* you are indexing in
On Oct 28, 2007, at 4:29 PM, William Stein wrote:
> On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
>> What is the correct/recommended use of Py_ssize_t in Cython code?
>
> You should use Py_ssize_t *anywhere* you are indexing into a Python
> list, tuple, etc. Py_ssize_t is the official Pyth
On 10/29/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
> > Now, instead if the code working like it used to, there will be an overflow,
> > and one will get total nonsense.
> >
>
> In your example the permanent is just the product of the entries.
>
> You deserve to get total nonsense when you try to c
William Stein wrote:
> On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
>
>> I did change some of them in trac #217, but I think a new trac
>> ticket should be created.
>
>
> Are you sure?I just had a look at trac #217, and your changing Py_ssize_t
> into int specifically *introduces* bu
William Stein wrote:
> On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
>>> Py_ssize_t should *never* be used if you really mean to use an
>>> int. But in any situation that you're indexing something, use
>>> it. Note that it is not just size_t from C, since it is signed,
>>> since in Python
On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
> > Py_ssize_t should *never* be used if you really mean to use an
> > int. But in any situation that you're indexing something, use
> > it. Note that it is not just size_t from C, since it is signed,
> > since in Python list indices can be neg
William Stein wrote:
> On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
>> What is the correct/recommended use of Py_ssize_t in Cython code?
>
> Py_ssize_t should *never* be used if you really mean to use an
> int. But in any situation that you're indexing something, use
> it. Note that it
On 10/28/07, Jaap Spies <[EMAIL PROTECTED]> wrote:
> What is the correct/recommended use of Py_ssize_t in Cython code?
You should use Py_ssize_t *anywhere* you are indexing into a Python
list, tuple, etc. Py_ssize_t is the official Python 2.5.x type
for indexing into such structures.
Py_ssize_
16 matches
Mail list logo
