FIAT_NEW + NumPy 64bit problem

[Garth Wells]

When trying to compile an elasticity problem using the FIAT_NEW and experimental FFC, I get

   File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py", line 400, in tabulate
    results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and 'numpy.float64'

Is this a FIAT or a NumPy bug?

[Robert Kirby (robert-c-kirby)]

Not sure. I've been having these bugs crop up.
I have company this weekend and might not get to this.
The quick attempt to work-around is to replace numpy.sqrt with
math.sqrt.

Rob

On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:

> Public bug reported:
>
> When trying to compile an elasticity problem using the FIAT_NEW and
> experimental FFC, I get
>
> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> line 400, in tabulate
> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> 'numpy.float64'
>
> Is this a FIAT or a NumPy bug?
>
> ** Affects: fiat
> Importance: Undecided
> Assignee: Robert Kirby (robert-c-kirby)
> Status: New
>
> ** Changed in: fiat
> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>
> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>
> --
> FIAT_NEW + NumPy 64bit problem
> https://bugs.launchpad.net/bugs/495602
> You received this bug notification because you are a bug assignee.
>
> Status in FInite element Automatic Tabulator: New
>
> Bug description:
> When trying to compile an elasticity problem using the FIAT_NEW and
> experimental FFC, I get
>
> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> line 400, in tabulate
> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> 'numpy.float64'
>
> Is this a FIAT or a NumPy bug?
>
>
>

[Robert Kirby (robert-c-kirby)]

I took a look; I had corrected a similar problem in the triangle expansion
set, so propogating the purported fix was not hard.

I've committed a change.

How do I indicate that the bug should be fixed in Launchpad?

Rob

On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <email address hidden>wrote:

> Not sure. I've been having these bugs crop up.
> I have company this weekend and might not get to this.
> The quick attempt to work-around is to replace numpy.sqrt with
> math.sqrt.
>
> Rob
>
>
> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
>
>> Public bug reported:
>>
>> When trying to compile an elasticity problem using the FIAT_NEW and
>> experimental FFC, I get
>>
>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>> line 400, in tabulate
>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>> 'numpy.float64'
>>
>> Is this a FIAT or a NumPy bug?
>>
>> ** Affects: fiat
>> Importance: Undecided
>> Assignee: Robert Kirby (robert-c-kirby)
>> Status: New
>>
>> ** Changed in: fiat
>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>>
>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>>
>> --
>> FIAT_NEW + NumPy 64bit problem
>> https://bugs.launchpad.net/bugs/495602
>> You received this bug notification because you are a bug assignee.
>>
>> Status in FInite element Automatic Tabulator: New
>>
>> Bug description:
>> When trying to compile an elasticity problem using the FIAT_NEW and
>> experimental FFC, I get
>>
>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>> line 400, in tabulate
>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>> 'numpy.float64'
>>
>> Is this a FIAT or a NumPy bug?
>>
>>
>>
>

[Garth Wells (garth-wells)]

On Dec 12 2009, Robert Kirby wrote:

>I took a look; I had corrected a similar problem in the triangle expansion
>set, so propogating the purported fix was not hard.
>
>I've committed a change.
>

I don't see the change. Where did you push to?

Garth

>How do I indicate that the bug should be fixed in Launchpad?
>
>Rob
>
> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby
> <email address hidden>wrote:
>
>> Not sure. I've been having these bugs crop up.
>> I have company this weekend and might not get to this.
>> The quick attempt to work-around is to replace numpy.sqrt with
>> math.sqrt.
>>
>> Rob
>>
>>
>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
>>
>>> Public bug reported:
>>>
>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>> experimental FFC, I get
>>>
>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>> line 400, in tabulate
>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>> 'numpy.float64'
>>>
>>> Is this a FIAT or a NumPy bug?
>>>
>>> ** Affects: fiat
>>> Importance: Undecided
>>> Assignee: Robert Kirby (robert-c-kirby)
>>> Status: New
>>>
>>> ** Changed in: fiat
>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>>>
>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>>>
>>> --
>>> FIAT_NEW + NumPy 64bit problem
>>> https://bugs.launchpad.net/bugs/495602
>>> You received this bug notification because you are a bug assignee.
>>>
>>> Status in FInite element Automatic Tabulator: New
>>>
>>> Bug description:
>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>> experimental FFC, I get
>>>
>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>> line 400, in tabulate
>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>> 'numpy.float64'
>>>
>>> Is this a FIAT or a NumPy bug?
>>>
>>>
>>>
>>
>

[Robert Kirby (robert-c-kirby)]

oops, commited but not pushed...I've been using cvs on another project...
It's pushed now.

On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden> wrote:

> On Dec 12 2009, Robert Kirby wrote:
>
> I took a look; I had corrected a similar problem in the triangle expansion
>> set, so propogating the purported fix was not hard.
>>
>> I've committed a change.
>>
>>
> I don't see the change. Where did you push to?
>
> Garth
>
>> How do I indicate that the bug should be fixed in Launchpad?
>>
>> Rob
>>
>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <<email address hidden>
>> >wrote:
>>
>> Not sure. I've been having these bugs crop up.
>>> I have company this weekend and might not get to this.
>>> The quick attempt to work-around is to replace numpy.sqrt with
>>> math.sqrt.
>>>
>>> Rob
>>>
>>>
>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
>>>
>>> Public bug reported:
>>>>
>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>> experimental FFC, I get
>>>>
>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>> line 400, in tabulate
>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>> 'numpy.float64'
>>>>
>>>> Is this a FIAT or a NumPy bug?
>>>>
>>>> ** Affects: fiat
>>>> Importance: Undecided
>>>> Assignee: Robert Kirby (robert-c-kirby)
>>>> Status: New
>>>>
>>>> ** Changed in: fiat
>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>>>>
>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>>>>
>>>> --
>>>> FIAT_NEW + NumPy 64bit problem
>>>> https://bugs.launchpad.net/bugs/495602
>>>> You received this bug notification because you are a bug assignee.
>>>>
>>>> Status in FInite element Automatic Tabulator: New
>>>>
>>>> Bug description:
>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>> experimental FFC, I get
>>>>
>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>> line 400, in tabulate
>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>> 'numpy.float64'
>>>>
>>>> Is this a FIAT or a NumPy bug?
>>>>
>>>>
>>>>
>>>>
>>>
>>
>

[Garth Wells (garth-wells)]

Robert Kirby wrote:
> oops, commited but not pushed...I've been using cvs on another project...
> It's pushed now.
>
>
> On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden> wrote:
>
>> On Dec 12 2009, Robert Kirby wrote:
>>
>> I took a look; I had corrected a similar problem in the triangle expansion
>>> set, so propogating the purported fix was not hard.
>>>
>>> I've committed a change.
>>>
>>>
>> I don't see the change. Where did you push to?
>>
>> Garth
>>
>>> How do I indicate that the bug should be fixed in Launchpad?
>>>

Go to

   https://bugs.launchpad.net/fiat/+bug/495602

and click on 'New' which is under 'Status'.

Garth

>>> Rob
>>>
>>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <<email address hidden>
>>>> wrote:
>>> Not sure. I've been having these bugs crop up.
>>>> I have company this weekend and might not get to this.
>>>> The quick attempt to work-around is to replace numpy.sqrt with
>>>> math.sqrt.
>>>>
>>>> Rob
>>>>
>>>>
>>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
>>>>
>>>> Public bug reported:
>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>>> experimental FFC, I get
>>>>>
>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>>> line 400, in tabulate
>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>>> 'numpy.float64'
>>>>>
>>>>> Is this a FIAT or a NumPy bug?
>>>>>
>>>>> ** Affects: fiat
>>>>> Importance: Undecided
>>>>> Assignee: Robert Kirby (robert-c-kirby)
>>>>> Status: New
>>>>>
>>>>> ** Changed in: fiat
>>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>>>>>
>>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>>>>>
>>>>> --
>>>>> FIAT_NEW + NumPy 64bit problem
>>>>> https://bugs.launchpad.net/bugs/495602
>>>>> You received this bug notification because you are a bug assignee.
>>>>>
>>>>> Status in FInite element Automatic Tabulator: New
>>>>>
>>>>> Bug description:
>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>>> experimental FFC, I get
>>>>>
>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>>> line 400, in tabulate
>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>>> 'numpy.float64'
>>>>>
>>>>> Is this a FIAT or a NumPy bug?
>>>>>
>>>>>
>>>>>
>>>>>
>

[Anders Logg (logg)]

On Sun, Dec 13, 2009 at 08:30:43PM -0000, Garth Wells wrote:
>
> Robert Kirby wrote:
> > oops, commited but not pushed...I've been using cvs on another project...
> > It's pushed now.
> >
> >
> > On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden> wrote:
> >
> >> On Dec 12 2009, Robert Kirby wrote:
> >>
> >> I took a look; I had corrected a similar problem in the triangle expansion
> >>> set, so propogating the purported fix was not hard.
> >>>
> >>> I've committed a change.
> >>>
> >>>
> >> I don't see the change. Where did you push to?
> >>
> >> Garth
> >>
> >>> How do I indicate that the bug should be fixed in Launchpad?
> >>>
>
> Go to
>
> https://bugs.launchpad.net/fiat/+bug/495602
>
> and click on 'New' which is under 'Status'.
>
> Garth

Shouldn't he click "Fix committed"?

--
Anders

>
> >>> Rob
> >>>
> >>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <<email address hidden>
> >>>> wrote:
> >>> Not sure. I've been having these bugs crop up.
> >>>> I have company this weekend and might not get to this.
> >>>> The quick attempt to work-around is to replace numpy.sqrt with
> >>>> math.sqrt.
> >>>>
> >>>> Rob
> >>>>
> >>>>
> >>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
> >>>>
> >>>> Public bug reported:
> >>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>> experimental FFC, I get
> >>>>>
> >>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>> line 400, in tabulate
> >>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>> 'numpy.float64'
> >>>>>
> >>>>> Is this a FIAT or a NumPy bug?
> >>>>>
> >>>>> ** Affects: fiat
> >>>>> Importance: Undecided
> >>>>> Assignee: Robert Kirby (robert-c-kirby)
> >>>>> Status: New
> >>>>>
> >>>>> ** Changed in: fiat
> >>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
> >>>>>
> >>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
> >>>>>
> >>>>>
> >>>>> Status in FInite element Automatic Tabulator: New
> >>>>>
> >>>>> Bug description:
> >>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>> experimental FFC, I get
> >>>>>
> >>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>> line 400, in tabulate
> >>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>> 'numpy.float64'
> >>>>>
> >>>>> Is this a FIAT or a NumPy bug?
> >>>>>
> >>>>>
> >>>>>
> >>>>>
> >
>

[Garth Wells (garth-wells)]

Anders Logg wrote:
> On Sun, Dec 13, 2009 at 08:30:43PM -0000, Garth Wells wrote:
>> Robert Kirby wrote:
>>> oops, commited but not pushed...I've been using cvs on another project...
>>> It's pushed now.
>>>
>>>
>>> On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden> wrote:
>>>
>>>> On Dec 12 2009, Robert Kirby wrote:
>>>>
>>>> I took a look; I had corrected a similar problem in the triangle expansion
>>>>> set, so propogating the purported fix was not hard.
>>>>>
>>>>> I've committed a change.
>>>>>
>>>>>
>>>> I don't see the change. Where did you push to?
>>>>
>>>> Garth
>>>>
>>>>> How do I indicate that the bug should be fixed in Launchpad?
>>>>>
>> Go to
>>
>> https://bugs.launchpad.net/fiat/+bug/495602
>>
>> and click on 'New' which is under 'Status'.
>>
>> Garth
>
> Shouldn't he click "Fix committed"?
>

Doesn't clicking on the status (which is now 'New') pop up a window from
which the new status can be chosen?

Garth

> --
> Anders
>
>
>>>>> Rob
>>>>>
>>>>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <<email address hidden>
>>>>>> wrote:
>>>>> Not sure. I've been having these bugs crop up.
>>>>>> I have company this weekend and might not get to this.
>>>>>> The quick attempt to work-around is to replace numpy.sqrt with
>>>>>> math.sqrt.
>>>>>>
>>>>>> Rob
>>>>>>
>>>>>>
>>>>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
>>>>>>
>>>>>> Public bug reported:
>>>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>>>>> experimental FFC, I get
>>>>>>>
>>>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>>>>> line 400, in tabulate
>>>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>>>>> 'numpy.float64'
>>>>>>>
>>>>>>> Is this a FIAT or a NumPy bug?
>>>>>>>
>>>>>>> ** Affects: fiat
>>>>>>> Importance: Undecided
>>>>>>> Assignee: Robert Kirby (robert-c-kirby)
>>>>>>> Status: New
>>>>>>>
>>>>>>> ** Changed in: fiat
>>>>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
>>>>>>>
>>>>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
>>>>>>>
>>>>>>>
>>>>>>> Status in FInite element Automatic Tabulator: New
>>>>>>>
>>>>>>> Bug description:
>>>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
>>>>>>> experimental FFC, I get
>>>>>>>
>>>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
>>>>>>> line 400, in tabulate
>>>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
>>>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
>>>>>>> 'numpy.float64'
>>>>>>>
>>>>>>> Is this a FIAT or a NumPy bug?
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>

[Anders Logg (logg)]

On Mon, Dec 14, 2009 at 08:08:03AM -0000, Garth Wells wrote:
>
> Anders Logg wrote:
> > On Sun, Dec 13, 2009 at 08:30:43PM -0000, Garth Wells wrote:
> >> Robert Kirby wrote:
> >>> oops, commited but not pushed...I've been using cvs on another project...
> >>> It's pushed now.
> >>>
> >>>
> >>> On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden> wrote:
> >>>
> >>>> On Dec 12 2009, Robert Kirby wrote:
> >>>>
> >>>> I took a look; I had corrected a similar problem in the triangle expansion
> >>>>> set, so propogating the purported fix was not hard.
> >>>>>
> >>>>> I've committed a change.
> >>>>>
> >>>>>
> >>>> I don't see the change. Where did you push to?
> >>>>
> >>>> Garth
> >>>>
> >>>>> How do I indicate that the bug should be fixed in Launchpad?
> >>>>>
> >> Go to
> >>
> >> https://bugs.launchpad.net/fiat/+bug/495602
> >>
> >> and click on 'New' which is under 'Status'.
> >>
> >> Garth
> >
> > Shouldn't he click "Fix committed"?
> >
>
> Doesn't clicking on the status (which is now 'New') pop up a window from
> which the new status can be chosen?

Correct. I missed that. So click the status which is now New and
change it to "Fix committed". Then when a new release is made of FIAT
change it to "Fix released".

--
Anders

> Garth
>
> >>>>> Rob
> >>>>>
> >>>>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <<email address hidden>
> >>>>>> wrote:
> >>>>> Not sure. I've been having these bugs crop up.
> >>>>>> I have company this weekend and might not get to this.
> >>>>>> The quick attempt to work-around is to replace numpy.sqrt with
> >>>>>> math.sqrt.
> >>>>>>
> >>>>>> Rob
> >>>>>>
> >>>>>>
> >>>>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden> wrote:
> >>>>>>
> >>>>>> Public bug reported:
> >>>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>>>> experimental FFC, I get
> >>>>>>>
> >>>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>>>> line 400, in tabulate
> >>>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>>>> 'numpy.float64'
> >>>>>>>
> >>>>>>> Is this a FIAT or a NumPy bug?
> >>>>>>>
> >>>>>>> ** Affects: fiat
> >>>>>>> Importance: Undecided
> >>>>>>> Assignee: Robert Kirby (robert-c-kirby)
> >>>>>>> Status: New
> >>>>>>>
> >>>>>>> ** Changed in: fiat
> >>>>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
> >>>>>>>
> >>>>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
> >>>>>>>
> >>>>>>>
> >>>>>>> Status in FInite element Automatic Tabulator: New
> >>>>>>>
> >>>>>>> Bug description:
> >>>>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>>>> experimental FFC, I get
> >>>>>>>
> >>>>>>> File "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>>>> line 400, in tabulate
> >>>>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>>>> 'numpy.float64'
> >>>>>>>
> >>>>>>> Is this a FIAT or a NumPy bug?
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >
>

[Robert Kirby (robert-c-kirby)]

Fix is now "committed"

On Sun, Dec 13, 2009 at 2:30 PM, Garth N. Wells <email address hidden> wrote:

>
>
> Robert Kirby wrote:
> > oops, commited but not pushed...I've been using cvs on another project...
> > It's pushed now.
> >
> >
> > On Sat, Dec 12, 2009 at 11:42 AM, Garth N. Wells <email address hidden>
> wrote:
> >
> >> On Dec 12 2009, Robert Kirby wrote:
> >>
> >> I took a look; I had corrected a similar problem in the triangle
> expansion
> >>> set, so propogating the purported fix was not hard.
> >>>
> >>> I've committed a change.
> >>>
> >>>
> >> I don't see the change. Where did you push to?
> >>
> >> Garth
> >>
> >>> How do I indicate that the bug should be fixed in Launchpad?
> >>>
>
> Go to
>
> https://bugs.launchpad.net/fiat/+bug/495602
>
> and click on 'New' which is under 'Status'.
>
> Garth
>
>
>
> >>> Rob
> >>>
> >>> On Fri, Dec 11, 2009 at 3:43 PM, Robert Kirby <
> <email address hidden>
> >>>> wrote:
> >>> Not sure. I've been having these bugs crop up.
> >>>> I have company this weekend and might not get to this.
> >>>> The quick attempt to work-around is to replace numpy.sqrt with
> >>>> math.sqrt.
> >>>>
> >>>> Rob
> >>>>
> >>>>
> >>>> On Fri, Dec 11, 2009 at 12:58 PM, Garth Wells <email address hidden>
> wrote:
> >>>>
> >>>> Public bug reported:
> >>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>> experimental FFC, I get
> >>>>>
> >>>>> File
> "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>> line 400, in tabulate
> >>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>> 'numpy.float64'
> >>>>>
> >>>>> Is this a FIAT or a NumPy bug?
> >>>>>
> >>>>> ** Affects: fiat
> >>>>> Importance: Undecided
> >>>>> Assignee: Robert Kirby (robert-c-kirby)
> >>>>> Status: New
> >>>>>
> >>>>> ** Changed in: fiat
> >>>>> Assignee: (unassigned) => Robert Kirby (robert-c-kirby)
> >>>>>
> >>>>> ** Branch linked: lp:~fiat-core/fiat/fiat-new
> >>>>>
> >>>>> --
> >>>>> FIAT_NEW + NumPy 64bit problem
> >>>>> https://bugs.launchpad.net/bugs/495602
> >>>>> You received this bug notification because you are a bug assignee.
> >>>>>
> >>>>> Status in FInite element Automatic Tabulator: New
> >>>>>
> >>>>> Bug description:
> >>>>> When trying to compile an elasticity problem using the FIAT_NEW and
> >>>>> experimental FFC, I get
> >>>>>
> >>>>> File
> "/usr/local/lib/python2.6/dist-packages/FIAT_NEW/expansions.py",
> >>>>> line 400, in tabulate
> >>>>> results[idx(p,q,r)] *= numpy.sqrt((p+0.5)*(p+q+1.0)*(p+q+r+1.5))
> >>>>> TypeError: unsupported operand type(s) for *=: 'numpy.ndarray' and
> >>>>> 'numpy.float64'
> >>>>>
> >>>>> Is this a FIAT or a NumPy bug?
> >>>>>
> >>>>>
> >>>>>
> >>>>>
> >
>
>
>