Test fails on particular system
Issue #2
on hold
Comments (6)
-
reporter
-
reporter
Looks like this is specific to 5.22.1 compiled with threads – 5.24 seems to fix this – and Moose.
-
reporter
- changed status to wontfix
Since this doesn't look like a bug in our code – more likely a an issue with threads in 5.22.1 and Moose or something – this failed test just does what it is supposed to do: Prevent installation on a broken system.
-
reporter
- changed status to open
Looks like this also a problem on Raspberry Pi, so we investigate that.
-
reporter
It’s related to Class::Measure::Scientific_fx992 failing when an object is created with an undefined amount.
-
reporter
- changed status to on hold
- Log in to comment
It’s odd this fails only on this particular system. What we see is that only the first change in metar breaks this, the other METARs test the same but pass. Looks like it has something to do with how Geo::METAR is initialised and the new Deduced object doesn’t properly inherit this property.