In general, Subversion updates are controlled via Subversion AuthZ and/or Subversion one or more of the "pre" hooks (pre-commit, pre-revprop-change, pre-lock, pre-unlock).
I know of zero publicly available hook implementations that enable such a policy out of the box. That said, there are any number of ways that such a policy could be implemented, depending on how you want to provide your approval (automated or manual).
Thanks DougR. I'm looking at doing this automatically in the background everytime a user runs the sanity. So, if the sanity isn't run or sanity failed then the user wouldn't get permissions to commit
There's going to need to be some communication between the "sanity" processor (SP) and the "pre-commit" hook (PCH). The simplest thing would be for the SP to run as part of the PCH. That's great if the processing is fast, less so if slower. But if that can be done then everything is automatic: the end-user does their testing, perhaps using some desktop version of the SP to validate that it will be happy when the checkin occurs. Then they checkin, the PCH runs the SP and it's happy so it returns success and the checkin completes. If the SP is unhappy it returns non-zero and the checkin fails.