On 2019/05/26 17:18, Antoine Jacoutot wrote: > On Sun, May 26, 2019 at 05:05:19AM -0600, Theo de Raadt wrote: > > That is weird. > > > > Why did syspatch believe this is a system it can operate on? > > Because it's that time of the year where there's no way to know on what we run > (release or current). > If you have a better idea than parsing the output of sysctl kern.version, I am > all ears.
I proposed this before for another reason and it didn't get anywhere, but maybe it's worth mentioning again. Include a build_id file in base*.tgz containing a uuid created during mkr. When a version release is made, add the uuids for actual releases to a file on ftp, anything that needs to know if it's on a release build can check against that file. There are things people might do that make it "not release" (local builds of base or kernel), if we care about that we could just have them invalidate the build_id.
