We have a build of the mysql-router charm that was released 1 hour[0] before the data samples shared here, this upload was marked as revision 30 by the store in the 8.0/edge track, where checking "charmcraft status mysql-router" is there (see https://pastebin.canonical.com/p/GXWbt5J7j6/ ), while the command "juju info mysql-router --channel 8.0 --series jammy --arch amd64" is still reporting the revision 26 for the 8.0/edge track (see https://pastebin.canonical.com/p/VPvCsZTcp9/ )
Querying the charmhub's api directly returns the expected revision 30
~ $ curl -s https://api.charmhub.io/v2/charms/info/mysql-router?fields=channel-map,channel-map.revision | jq '."channel-map" |. [] | select(.channel.risk == "edge" and .channel.track == "8.0" and .channel.base.architecture == "amd64" and .channel.base.channel == "22.04") |.revision.revision'
30
Running juju deploy does find revision 30, so this seems to be constrained to the code behind "juju info".
~ $ juju deploy --series jammy --channel 8.0/edge mysql-router
Located charm "mysql-router" in charm-hub, revision 30
Deploying "mysql-router" from charm-hub charm "mysql-router", revision 30 in channel 8.0/edge on jammy
[0] https://launchpad.net/~openstack-charmers/charm-mysql-router/+charm/charm-mysql-router.stable-jammy.8.0.19/+build/2473
The problem appears when using the --channel selector with the query. When the selector is not used the retrieved information is correct. See https:/ /pastebin. canonical. com/p/CBpCnRKYN p/
This could be related with the fact that charmhub returns the list of found entries in incremental order
curl -s https:/ /api.charmhub. io/v2/charms/ info/mysql- router? fields= channel- map,channel- map.revision | jq '."channel-map" |. [] | select( .channel.track == "8.0" ) |.revision. revision'