Pierre, The attached script shows how one can make a masked array with different dimensions for the mask than for the data. I presume this is a bug. It is triggered by the present code in the matplotlib quiver function. Eric
On Sunday 20 January 2008 16:32:40 you wrote:
Pierre,
The attached script shows how one can make a masked array with different dimensions for the mask than for the data. I presume this is a bug. It is triggered by the present code in the matplotlib quiver function.
Yep, bug indeed. Thanks for pointing that out ! The following patch should take care of the problem. (In short, a getmask function was used instead of getmaskarray). Note that the patch takes also into account elements I had sent to Stefan 2 weeks ago, that were not ported yet to the SVN: I still can't commit to the numpy/branches/maskedarray. Let me know how it works, and thanks again for your time. P.
Pierre GM wrote:
On Sunday 20 January 2008 16:32:40 you wrote:
Pierre,
The attached script shows how one can make a masked array with different dimensions for the mask than for the data. I presume this is a bug. It is triggered by the present code in the matplotlib quiver function.
Yep, bug indeed. Thanks for pointing that out ! The following patch should take care of the problem. (In short, a getmask function was used instead of getmaskarray). Note that the patch takes also into account elements I had sent to Stefan 2 weeks ago, that were not ported yet to the SVN: I still can't commit to the numpy/branches/maskedarray. Let me know how it works, and thanks again for your time.
Pierre, Thank you, it works fine. Eric
Hi Pierre, Eric On Mon, Jan 21, 2008 at 03:35:37PM -0500, Pierre GM wrote:
Yep, bug indeed. Thanks for pointing that out ! The following patch should take care of the problem. (In short, a getmask function was used instead of getmaskarray). Note that the patch takes also into account elements I had sent to Stefan 2 weeks ago, that were not ported yet to the SVN: I still can't commit to the numpy/branches/maskedarray.
The patch you refer to was applied in r4718 on 10 January: http://projects.scipy.org/scipy/numpy/changeset/4718 I hope there weren't others I missed. I shall work in your latest patch as soon as I can access the repo again. Regards Stéfan
participants (3)
-
Eric Firing -
Pierre GM -
Stefan van der Walt