Hi Dave, Yes, I'd be happy to drive an RC release. It'll be a good test of the ATR skill with someone who didn't write it.
Where things stand: - This week's 7.0 PRs are merged, including #251, which renames the source archive's root folder to apache-roller-<version>-source/ so it matches the archive name and ATR's archive-structure check. - I created a new RSA-4096 code signing key for [email protected] and added it to KEYS in dist (r89369). Fingerprint: 7EBB A59D 8240 A40D F8C4 C855 42D9 2150 BCF2 7404 - I've been running master against a copy of raibledesigns.com's data (about 3,500 entries and 15,000 comments, migrated from 5.0.3). It starts cleanly, and I haven't found any blockers. Reordering categories shows up on the public site right away, without a restart. The upgrade turned up a few smaller issues, and I'd like to get them in before the RC. Each is a small PR: - #252: A database password made only of hex digits stopped Roller from starting, because EclipseLink thought it was encrypted with its old algorithm. - #253: The weblog's "about" text and an entry's search description weren't encoded in the meta description tag, so a quote cut the tag short. It's not exploitable under our trust model, but it's the kind of thing that gets reported as XSS. - #254: An older theme that used a removed message key as a date pattern lost its whole entry list. formatDate now falls back to a default format, and CHANGES.md notes the macros 5.1 removed. - #255: Docs. The install guide now explains unattended upgrades (installation.type=manual plus the migration scripts), and the user guide no longer says the title and tagline can include HTML. Your 2023 change strips it on purpose, so I left that alone. Could you review those? I'll start the RC once they're in and ATR picks up my key. I'll let you know if I get stuck. Matt > On Nov 14, 2026, at 10:28, Dave <[email protected]> wrote: > > Hi Matt, > > Do you want to try out the new Apache Trusted Releases skill and spin up a > Roller 7 RC once everything is merged today? I'm happy to do it, but maybe > you want to drive this time? > > Dave > > > On Fri, Nov 13, 2026 at 6:43 PM Dave <[email protected]> wrote: > >> concur. I think we should have a one other approver rule as we are a small >> project. >> >> >> >> On Fri, Nov 13, 2026 at 6:07 PM Matt Raible <[email protected]> >> wrote: >> >>> I’m OK with it. It always seems strange to work on an OSS project and >>> have the ability to merge right after creating a PR. >>> >>> I’ve done a lot of professional OSS and companies seem to always require >>> a 2nd approver. >>> >>> That being said, we still don’t require it on JHipster, but we do have a >>> convention that every PR gets an approval from a committer. >>> >>>> On Nov 13, 2026, at 15:33, Dave <[email protected]> wrote: >>>> >>>> On second thought, I do think approvals are more important now than they >>>> were before because people are submitting AI generated PRs that they may >>>> not 100% understand and these AI generated PRs are generally larger and >>>> harder to review. >>>> >>>> Dave >>>> >>>> >>>> On Fri, Nov 13, 2026 at 5:16 PM Dave <[email protected]> wrote: >>>> >>>>> It's awesome to see so many PRs open against Roller! Thanks, folks. I >>>>> will do my best to review all things this weekend and drive towards a >>>>> Roller 7 RC soon. >>>>> >>>>> Reviews are not required before merge based on our current policy. Does >>>>> anybody think we should have branch protection and require 1 or 2 >>> approvals >>>>> on PRs? I am undecided about this. >>>>> >>>>> Dave >>>>> >>> >>>
