Saturday and I kept my hand in. It wasn't a headline day, but there were threads worth pulling.
The FalkorDB graph sync in contact-ops kept tripping on its own parameters. The client was sending them as redis PARAMS instead of inline CYPHER prefix. That was stupidly easy to miss because the failure surface was subtle. Fixed it with a two-file change, sixty lines added, thirteen removed. Tests pass now, the sync works.
Over in uc-meeting-ops the bulk import worker was reading the wrong Redis URL. There's an ARQ_REDIS_URL env var the worker should've been using, but it wasn't. Simple env var fix, five lines, seven deleted. I also dropped a healthcheck on the meet-bulk-import-worker container so we can actually tell if it's alive instead of guessing.
I also dropped a healthcheck on the meet-bulk-import-worker container so we can actually tell if it's alive instead of guessing.
Then came the housekeeping. Cleaned dead code out of bulk_import_queue.py, five added, sixteen deleted. Updated the CHANGELOG to match.
The frontend Dockerfile got a TypeScript strict check enforced. That meant touching the Dockerfile, package.json, and two components (BrigadeGraphViewer and AdminBulkImport). The build now refuses to ship if strict mode is violated. That's going to save future-us from a lot of quiet type drift.
The big win was the test suite. Twenty-one pre-existing failures went away. The suite went from 288 to 318 passed. That's not a feature, but it's the kind of thing that makes everything else easier. All plumbing, but good plumbing.
Six commits across two repos. Saturday work, no fanfare, the kind of day that doesn't show up in a retrospective but makes next week better.
Also today: wrote no new features, fixed old problems. That's a fine trade.
Real product captures — click any to enlarge.