You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A subset of the same tests from test_tutorials.py::test_geoFiles fails on non-Intel architectures such as arm64. It looks like a simple matter of floating point tolerance, though a couple are integer tests.
The specific set is
The "integer" errors seem to be vertex/face counts. The difference is probably caused by slightly different meshing decisions and tolerance calculations, i.e. numerical differences again.
I can see the same on openSUSE Tumbleweed for aarch64 and ppcle64.
A subset of the same tests from
test_tutorials.py::test_geoFiles
fails on non-Intel architectures such as arm64. It looks like a simple matter of floating point tolerance, though a couple are integer tests.The specific set is
Build logs can be found at https://buildd.debian.org/status/package.php?p=netgen&suite=experimental (past builds will be found under the "old" links at https://buildd.debian.org/status/package.php?p=netgen)
An example is the boxcyl error from arm64 at https://buildd.debian.org/status/fetch.php?pkg=netgen&arch=arm64&ver=6.2.2305%2Bdfsg1-1exp1&stamp=1699664433&raw=0
The text was updated successfully, but these errors were encountered: