Closed theseion closed 2 months ago
@estebanlm, @guillep Here's such an instance where the behavior of Pharo64-13 differs from Pharo64-alpha: https://github.com/hpi-swa/smalltalkCI/actions/runs/9525945437/job/26260957214?pr=646#step:4:236.
The method simply checks wether generating two random values twice produces a different identifier, equivalent of:
(ByteArray with: 255 atRandom with: 255 atRandom) hex ~= (ByteArray with: 255 atRandom with: 255 atRandom)
For some very weird reason, Random>>#nextInteger:
produces the same value multiple times. At first, I thought it was a timing issue, so I fixed the use of Random
in #newTravisID
, but the test still fails. Interestingly, the test only fails on the CI, not on my machine. I suspect, that it's an issue that depends on OS and VM combination (can't test that on my machine).
The Random problem in Pharo is now fixed: https://github.com/pharo-project/pharo/pull/16780. I also proposed a PR to fi fix the problem of the system reporter in P13: https://github.com/hpi-swa/smalltalkCI/pull/649/
@fniephaus good to go
StSystemReporter
was removed in Pharo 13.