Hi all, Just the latest update, and got these errors when trying to output a solution that is in CG1 space: Traceback (most recent call last): File "2D_OS_point_ex1.py", line 309, in <module> outFile_psiA << u0_A File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__ self._file << data File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__ self._update_PVD(data) File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD new_vtk << function File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__ coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q') File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1 if isinstance(e, (hc.HDiv, hc.HCurl)): AttributeError: 'module' object has no attribute 'HDiv' I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it? On a side note, I also got these warnings: *Mesh.cell() is deprecated, please use .ufl_cell() instead.* Thanks, Justin
Hi, if isinstance(e, (hc.HDiv, hc.HCurl)): I believe this line should be: if isinstance(e, (hc.HDivElement, hc.HCurlElement)): (We should fix this.) Thanks, Miklos ________________________________ From: firedrake-bounces@imperial.ac.uk [firedrake-bounces@imperial.ac.uk] on behalf of Justin Chang [jychang48@gmail.com] Sent: 20 November 2015 17:15 To: firedrake Subject: [firedrake] Errors Hi all, Just the latest update, and got these errors when trying to output a solution that is in CG1 space: Traceback (most recent call last): File "2D_OS_point_ex1.py", line 309, in <module> outFile_psiA << u0_A File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__ self._file << data File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__ self._update_PVD(data) File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD new_vtk << function File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__ coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q') File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1 if isinstance(e, (hc.HDiv, hc.HCurl)): AttributeError: 'module' object has no attribute 'HDiv' I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it? On a side note, I also got these warnings: Mesh.cell() is deprecated, please use .ufl_cell() instead. Thanks, Justin
On 20 Nov 2015, at 17:15, Justin Chang <jychang48@gmail.com> wrote:
Hi all,
Just the latest update, and got these errors when trying to output a solution that is in CG1 space:
Traceback (most recent call last): File "2D_OS_point_ex1.py", line 309, in <module> outFile_psiA << u0_A File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__ self._file << data File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__ self._update_PVD(data) File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD new_vtk << function File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__ coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q') File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1 if isinstance(e, (hc.HDiv, hc.HCurl)): AttributeError: 'module' object has no attribute 'HDiv'
Oops, in firedrake/io.py change references to HDiv and HCurl into HDivElement and HCurlElement respectively. If you do this can you prepare a pull request?
I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it?
On a side note, I also got these warnings:
Mesh.cell() is deprecated, please use .ufl_cell() instead.
Ah, this is harmless, I thought we'd changed all of these places but clearly there are a few left. Cheers, Lawrence
okay just did. On Fri, Nov 20, 2015 at 11:20 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 20 Nov 2015, at 17:15, Justin Chang <jychang48@gmail.com> wrote:
Hi all,
Just the latest update, and got these errors when trying to output a solution that is in CG1 space:
Traceback (most recent call last):
File "2D_OS_point_ex1.py", line 309, in <module>
outFile_psiA << u0_A
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__
self._file << data
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__
self._update_PVD(data)
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD
new_vtk << function
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__
coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q')
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1
if isinstance(e, (hc.HDiv, hc.HCurl)):
AttributeError: 'module' object has no attribute 'HDiv'
Oops, in firedrake/io.py change references to HDiv and HCurl into HDivElement and HCurlElement respectively. If you do this can you prepare a pull request?
I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it?
On a side note, I also got these warnings:
*Mesh.cell() is deprecated, please use .ufl_cell() instead.*
Ah, this is harmless, I thought we'd changed all of these places but clearly there are a few left.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Hi all, So fixing HDivElement and HCurlElement works. Now, unrelated, I am getting some installation errors when installing firedrake on my University iMac (still on Yosemite). Attached is the output from --log. It seems this error is PETSc/HDF5/h5py related, and I can't seem to dissect what the error means. Can y'all understand what's going on or should I defer this to the PETSc folks? Thanks, Justin On Fri, Nov 20, 2015 at 11:44 AM, Justin Chang <jychang48@gmail.com> wrote:
okay just did.
On Fri, Nov 20, 2015 at 11:20 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 20 Nov 2015, at 17:15, Justin Chang <jychang48@gmail.com> wrote:
Hi all,
Just the latest update, and got these errors when trying to output a solution that is in CG1 space:
Traceback (most recent call last):
File "2D_OS_point_ex1.py", line 309, in <module>
outFile_psiA << u0_A
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__
self._file << data
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__
self._update_PVD(data)
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD
new_vtk << function
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__
coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q')
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1
if isinstance(e, (hc.HDiv, hc.HCurl)):
AttributeError: 'module' object has no attribute 'HDiv'
Oops, in firedrake/io.py change references to HDiv and HCurl into HDivElement and HCurlElement respectively. If you do this can you prepare a pull request?
I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it?
On a side note, I also got these warnings:
*Mesh.cell() is deprecated, please use .ufl_cell() instead.*
Ah, this is harmless, I thought we'd changed all of these places but clearly there are a few left.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Sorry wrong log file. I also had to clip the other stuff because the file was too big. Do you need to see more? Thanks, Justin On Fri, Nov 20, 2015 at 12:37 PM, Justin Chang <jychang48@gmail.com> wrote:
Hi all,
So fixing HDivElement and HCurlElement works.
Now, unrelated, I am getting some installation errors when installing firedrake on my University iMac (still on Yosemite). Attached is the output from --log.
It seems this error is PETSc/HDF5/h5py related, and I can't seem to dissect what the error means. Can y'all understand what's going on or should I defer this to the PETSc folks?
Thanks, Justin
On Fri, Nov 20, 2015 at 11:44 AM, Justin Chang <jychang48@gmail.com> wrote:
okay just did.
On Fri, Nov 20, 2015 at 11:20 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 20 Nov 2015, at 17:15, Justin Chang <jychang48@gmail.com> wrote:
Hi all,
Just the latest update, and got these errors when trying to output a solution that is in CG1 space:
Traceback (most recent call last):
File "2D_OS_point_ex1.py", line 309, in <module>
outFile_psiA << u0_A
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 110, in __lshift__
self._file << data
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 524, in __lshift__
self._update_PVD(data)
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 542, in _update_PVD
new_vtk << function
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 188, in __lshift__
coords_p1 = is_family1(ce, 'Lagrange') or is_family1(ce, 'Q')
File "/Users/justin/Software/firedrake/src/firedrake/firedrake/io.py", line 157, in is_family1
if isinstance(e, (hc.HDiv, hc.HCurl)):
AttributeError: 'module' object has no attribute 'HDiv'
Oops, in firedrake/io.py change references to HDiv and HCurl into HDivElement and HCurlElement respectively. If you do this can you prepare a pull request?
I noticed that there have been many updates/changes regarding UFL, so could this have something to do with it?
On a side note, I also got these warnings:
*Mesh.cell() is deprecated, please use .ufl_cell() instead.*
Ah, this is harmless, I thought we'd changed all of these places but clearly there are a few left.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Hi Justin,
On 20 Nov 2015, at 18:40, Justin Chang <jychang48@gmail.com> wrote:
Sorry wrong log file.
I also had to clip the other stuff because the file was too big. Do you need to see more?
This /looks/ like a pure PETSc problem. If you try and build PETSc on its own with the firedrake-install "minimal" configure flags: --download-ctetgen --download-triangle --download-chaco --download-hdf5 Do you obtain the same error? If so, maybe punt to petsc-maint. If not, and that works, we'll need to figure out what we're doing wrong. It looks vaguely like the compiler is spewing out object code that the assembler doesn't like. Lawrence
Yup same error, because --download-hdf5 is still there Okay I will mail petsc-maint, but is there a way to extract the petsc configure log from firedrake? Or should I pull my own mapdes petsc branch and send that configure log? Thanks, Justin On Fri, Nov 20, 2015 at 12:45 PM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
Hi Justin,
On 20 Nov 2015, at 18:40, Justin Chang <jychang48@gmail.com> wrote:
Sorry wrong log file.
I also had to clip the other stuff because the file was too big. Do you need to see more?
This /looks/ like a pure PETSc problem. If you try and build PETSc on its own with the firedrake-install "minimal" configure flags:
--download-ctetgen --download-triangle --download-chaco --download-hdf5
Do you obtain the same error?
If so, maybe punt to petsc-maint. If not, and that works, we'll need to figure out what we're doing wrong. It looks vaguely like the compiler is spewing out object code that the assembler doesn't like.
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
On 20 Nov 2015, at 18:51, Justin Chang <jychang48@gmail.com> wrote:
Yup same error, because --download-hdf5 is still there
Okay I will mail petsc-maint, but is there a way to extract the petsc configure log from firedrake? Or should I pull my own mapdes petsc branch and send that configure log?
I think that it gets lost in the pip install process, sorry. Lawrence
participants (3)
-
Homolya, Miklós
-
Justin Chang
-
Lawrence Mitchell