Closed olupton closed 1 year ago
The latest run only has one model showing up in the diff
when comparing 8.2.2 with neuron-nightly. Note that this does not include the gout comparison, but I believe that this also shows no diffs when I run it locally.
On the face of it this is a more substantive issue that needs further investigation. I suggest we can move ahead and merge this PR as soon as the dependent updates to models linked in the description are merged.
The latest run only has one model showing up in the diff when comparing 8.2.2 with neuron-nightly. Note that this does not include the gout comparison, but I believe that this also shows no diffs when I run it locally.
On the face of it this is a more substantive issue that needs further investigation. I suggest we can move ahead and merge this PR as soon as the dependent updates to models linked in the description are merged.
This diff is removed by https://github.com/ModelDBRepository/150245/pull/3/commits/9e033d44df8a77458f4afed1fb84b290d9130ca7. In my local setup, the uninitialised value in 8.2.2 was 25, while in neuron-nightly it was zero.
Many of the diffs are due to error messages that are, in essence, the same on both sides of the comparison. Some error message formatting has changed since NEURON 8.2.2, so "the same error" can still result in a diff. This is an effort to avoid the problem by fixing the errors, with obvious benefits.
To be merged first +
github:
keys removed from YAML in this PR:
- [ ] C++ compatibility ModelDBRepository/83344#1
- [ ] Remove new_seed RANGE variables ModelDBRepository/118631#1
- [ ] Fix: array dimension of o now 1 (at compile time it was 0) ModelDBRepository/146949#3
- [ ] CNAME: don't assume strings live long enough ModelDBRepository/150245#3
- [ ] C++ compatibility ModelDBRepository/267067#3
All the approved PRs above have been merged and updated to ModelDB.
However, the drat pull request of the three models - 146949, 150245, and 267067 have very similar changes as the approved PR. Do you plan to continue working on the draft PR, or do you like to delete them? @olupton
Many of the diffs are due to error messages that are, in essence, the same on both sides of the comparison. Some error message formatting has changed since NEURON 8.2.2, so "the same error" can still result in a diff. This is an effort to avoid the problem by fixing the errors, with obvious benefits. To be merged first +
github:
keys removed from YAML in this PR:
- [ ] C++ compatibility ModelDBRepository/83344#1
- [ ] Remove new_seed RANGE variables ModelDBRepository/118631#1
- [ ] Fix: array dimension of o now 1 (at compile time it was 0) ModelDBRepository/146949#3
- [ ] CNAME: don't assume strings live long enough ModelDBRepository/150245#3
- [ ] C++ compatibility ModelDBRepository/267067#3
All the approved PRs above have been merged and updated to ModelDB.
However, the drat pull request of the three models - 146949, 150245, and 267067 have very similar changes as the approved PR. Do you plan to continue working on the draft PR, or do you like to delete them? @olupton
Thanks for merging the updates! Those other draft PRs are connected to neuronsimulator/nrn#2027; please leave them open for the moment, they will be updated soon.
Launched https://github.com/neuronsimulator/nrn-modeldb-ci/actions/runs/4325619099 with the full set of models.
Many of the diffs are due to error messages that are, in essence, the same on both sides of the comparison. Some error message formatting has changed since NEURON 8.2.2, so "the same error" can still result in a diff. This is an effort to avoid the problem by fixing the errors, with obvious benefits.
To be merged first +
github:
keys removed from YAML in this PR: