Closed stefan-lindstrom closed 4 months ago
Tagging @jamesagnew
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 83.45%. Comparing base (
497b9f2
) to head (bef011d
). Report is 74 commits behind head on master.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Fixes #5738
Thanks for the contribution!
HAPI to help. (Sorry, I'll see my self out). Is it OK to merge (once I managed to sort out authorization)?
Yup, I've merged. Don't worry about the 2 failed CI checks.. Those fail on external contributions because of some permission issue we haven't figured out yet. Thanks again!
Hi,
first time through the code trying to fix an issue, so I hope I haven't violated too many architectural rules. :)
We saw this problem when using extensions on datatypes (using data-absent-reason, with code masked and we had an object with both a primitive and an array-property with masked values. We got this output (example from practitioner):
My understanding is that when the object is normally handled, the parser counts alternateNamesSeen, and alternateNames Handled, and when seen > handled, a second pass is done. But instead of only doing un-handled, the second pass will go over all alternates including non-primitive ones where the alternate json.-type need not be an object but could be an array for (any non-primitive property).
Som y idea is to check, and only do the secondary handling if the alternate property in an object does not have a corresponding "real" property.