Previous runs failed because github.com/simolus3/sqlite3.dart release
downloads were returning 504 Gateway Timeout (the sqlite3 hook hashed
the error body to bef140a1…). Upstream is now returning the binary
with the expected hash 9b1c8c24…, so retrigger CI.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previous CI run failed during integration tests with a libsqlite3 download
hash mismatch — an upstream/CDN flake unrelated to this PR's Dagger pin
alignment.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
CI on the previous commit failed because the runner CLI and remote engine
are both still v0.20.8 ("module requires dagger v0.21.4, but you have
v0.20.8"). The issue offered two paths — bump everything to v0.21.4, or
re-pin the CLI — so take the second: align ci/dagger.json, flake.nix,
.forgejo/Dockerfile and DAGGER.md back to v0.20.8. The check script
remains and now enforces lockstep at the new baseline; once the host
engine and runner image are upgraded, a single follow-up bumps all four
references together.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Bump ci/dagger.json engineVersion, the .forgejo runner Dockerfile, and the
example systemd unit in DAGGER.md from 0.20.8 to 0.21.4 so they match the
running engine and the CLI already pinned by flake.nix.
Add scripts/check_dagger_versions.sh, wired into Taskfile (task
check-dagger-versions) and pre-commit, which fails if the four version
references drift.
Closes#542
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>