Hi Paul, > > However, why the slight change to security-related overflow handling > > in bitfield fields *on i386 systems* should result in this failure > > eludes me... :/ > > The changelog mentions some other bug fixes, the first one looks > potentially related (new failure): > Fail EXEC command in case a watched key is expired (#9194) > and so does the third (WRONGTYPE error): > Fix SINTERSTORE not to delete dest key when getting a wrong type error > (#9032)
Good news. I've tried reverting a bunch of commits from the changelog, and I can narrow it down to: https://github.com/redis/redis/pull/9032 As in, reverting the commit associated with this pull request: https://github.com/redis/redis/commit/1655576e23c41ea9c12a42699651d207656a0e83 ... results in the test passing again. § I'd be happy to report this to Redis upstream, but I have no evidence that this indicates an actual bug in Redis itself or any kind of "When I see X we see Y but we should see Z". I lack knowledge about what python-fakeredis is actually testing here (as well as how Hypothesis works!) to determine which package is buggy. Could the fakeredis maintainer chime in perhaps? Regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org 🍥 chris-lamb.co.uk `-