-
Notifications
You must be signed in to change notification settings - Fork 50
use stable hash for labels and name suffixes (aligning with OLMv0 changes) #764
use stable hash for labels and name suffixes (aligning with OLMv0 changes) #764
Conversation
Signed-off-by: Joe Lanford <[email protected]>
8cc3333
to
84181f2
Compare
84181f2
to
a615bd8
Compare
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #764 +/- ##
==========================================
- Coverage 21.14% 20.62% -0.52%
==========================================
Files 14 14
Lines 1069 1086 +17
==========================================
- Hits 226 224 -2
- Misses 795 812 +17
- Partials 48 50 +2 ☔ View full report in Codecov by Sentry. |
@joelanford e2e looks very borked |
Looks like the bundle validation is rejecting the newly generated bundle name. Seems like this is a built-in validation for most resource types...
|
That error might be misleading, I think you pass the regex fine, but may be running afoul of the length requirement. Suggest not adding the prefixes unless you really need to? |
It definitely looks like the regex rejects uppercase letters. I changed to base36 and now seeing the length validation failure.
|
I'm kinda leaning toward base36 + truncate the base name, kinda like we do with RBAC. But that leaves only 8 characters for the base name if we don't also truncate the hash. 🤔 |
length of 52 is coming from here:
I think that was because we knew that we wanted to leave some extra characters available for other objects that might need to exist for the bundle (right now, that's at least the unpack pod, can't remember if there are others) |
…nges) Signed-off-by: Joe Lanford <[email protected]>
a615bd8
to
0b87aa6
Compare
b4e9976
See operator-framework/operator-lifecycle-manager#3115