Conda-based testing

[lmoresi]

The updated petsc etc (#10) has broken the old conda-based build and test workflow. Should be simple to update but does need to be done.

[rbeucher]

I am working on it

[brmather]

The tests are still failing, and from the error messages it looks like conda must be installing some version of quagmire without QuagMesh. I think we should have separate actions for testing with pytest and building a Quagmire binary to upload to the cloud. The former should run on every push, the latter on every push to the master branch.

[lmoresi]

Agreed - the tests can’t be based on an uploaded version or there will be a race condition. Binaries really should update on a release rather than a push. Prof Louis Moresi louis.moresi@anu.edu.au (m) +61 4 0333 1413 (us) +1 505 349 4425 www.moresi.info<http://www.moresi.info/> www.facebook.com/underworldcode<http://www.facebook.com/underworldcode> @LouisMoresi<https://twitter.com/LouisMoresi> On 28 May 2020, 4:17 PM +1000, Ben Mather <notifications@github.com>, wrote: The tests are still failing on the test stage, and from the error messages it looks like conda must be installing some version of quagmire without QuagMesh. I think we should have separate actions for testing with pytest and building a Quagmire binary to upload to the cloud. The former should run on every push, the latter on every push to the master branch. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADABPIZ2B77KVXNFL2CLGHDRTX6WRANCNFSM4NF2FYEA>.

[rbeucher]

All right, my bad, it was still getting quagmire from my branch.
I have made the changes.

[brmather]

I can't seem to install petsc binaries with a fresh conda install (using a miniconda3 docker container).

conda install -c underworldcode petsc
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: - 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
failed

[lmoresi]

Unclear why that should be - a completely clean environment ? Prof Louis Moresi louis.moresi@anu.edu.au (m) +61 4 0333 1413 (us) +1 505 349 4425 www.moresi.info<http://www.moresi.info/> www.facebook.com/underworldcode<http://www.facebook.com/underworldcode> @LouisMoresi<https://twitter.com/LouisMoresi> On 29 May 2020, 2:13 PM +1000, Ben Mather <notifications@github.com>, wrote: I can't seem to install petsc binaries with a fresh conda install (using a miniconda3 docker container<https://hub.docker.com/r/continuumio/miniconda3>). conda install -c underworldcode petsc Collecting package metadata (current_repodata.json): done Solving environment: failed with initial frozen solve. Retrying with flexible solve. Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source. Collecting package metadata (repodata.json): done Solving environment: failed with initial frozen solve. Retrying with flexible solve. Solving environment: - Found conflicts! Looking for incompatible packages. This can take several minutes. Press CTRL-C to abort. failed — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADABPIY5BOKMFNOASGKHRPTRT4Y6FANCNFSM4NF2FYEA>.

[brmather]

That's right. More output:

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions

Package _libgcc_mutex conflicts for:
petsc -> libgcc-ng[version='>=7.5.0'] -> _libgcc_mutex=[build=main]
python=3.7 -> libgcc-ng[version='>=7.3.0'] -> _libgcc_mutex=[build=main]

Package libstdcxx-ng conflicts for:
petsc -> metis[version='>=5.1.0,<5.2.0a0'] -> libstdcxx-ng[version='>=7.2.0|>=7.3.0']
petsc -> libstdcxx-ng[version='>=7.5.0']

Package libgcc-ng conflicts for:
petsc -> libgcc-ng[version='>=7.5.0']
petsc -> metis[version='>=5.1.0,<5.2.0a0'] -> libgcc-ng[version='>=7.2.0|>=7.3.0']

[rbeucher]

Yeah... I can't either. Well I can install petsc fine. But I struggle with quagmire. !!!!!

[rbeucher]

I think I need to build it with more flexible settings.

[lmoresi]

It would probably work just fine if you removed the hdf5 options On 29 May 2020, 2:33 PM +1000, Romain Beucher <notifications@github.com>, wrote: I think I need to build it with more flexible settings. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#25 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADABPI7BNBBEB5OVGSWW3GDRT43JPANCNFSM4NF2FYEA>.

[brmather]

What I don't understand is that installing the conda-forge petsc pulls in hdf5, but petsc doesn't compile against it.

[rbeucher]

Looking at it now. QUAG-3 QUAG-4

[rbeucher]

OK comes from h5py that does not have a build with HDF5 1.10.6.
They seem to have some issues with windows. We can build our own h5py in our channel for now.

[brmather]

Fingers crossed PR conda-forge/h5py-feedstock#72 is merged soon!

[brmather]

#28 has been merged into dev. Thanks @rbeucher for all your hard work getting conda to work with Quagmire! I'm closing this issue...