Closed ed-xmos closed 8 months ago
We need to fix these lib tests (or ignore the fails) - we don't want to have version numbers in the git history that don't exist.
We need to fix these lib tests (or ignore the fails) - we don't want to have version numbers in the git history that don't exist.
OK - this is the way the software team have always done it so this is a different practice to what I am used to. Does it really matter as long as the changelog is correct (semantic version-wise) and no tagged release has been made? The user should always expect some changes (although they should build and work) on develop? We need a green Jenkins run before we can merge and disabling tests doesn't feel right. This would point to changing the xmos_changelog script rules which I think would need broader circulation to approve.
Understand. Ive made (another) note to fix the process/scripts
Looks like fptr groups were not fully implemented, causing stack corruption in my ASRC implementation which uses multiple threads. Also, the head of develop didn't pass lib tests so fixed that too. I have temporarily given this version 2.5.0 since there is a fix (patch) plus a new feature (minor)