Issues 38 open · 61 closed

● #101 flow deps check: is every crate in Cargo.lock vendored in the image? featuregap

opened 1 day ago by andymccall

● #98 flow runner register keeps a relative --ca path, so the service cannot find it bug

opened 8 days ago by andymccall

● #97 The swarm moved an image at 2 MB/s: scattered reads on a spinning disk and a quadratic picker performance

opened 8 days ago by andymccall

● #96 The cheap check jobs as WASM, so the phone, Windows and browser runners can take them feature

opened 8 days ago by andymccall

● #94 Runners send the manifests of images they hold, so the swarm moves an image nobody pushed feature

opened 8 days ago by andymccall · 1 comment

● #93 A WASM job stopped while its module is still compiling never stops bug

opened 8 days ago by andymccall

● #92 An idle runner is offered sandboxes its host cannot start, fetches their images, then declines them bug

opened 8 days ago by andymccall

● #85 Runner updates, runner side: flow runner update --from FILE feature

opened 9 days ago by andymccall

● #82 Test report and coverage for OxygenFlow on every push feature

opened 10 days ago by andymccall · 1 comment

● #81 cargo-nextest in the Rust image, so Rust jobs write JUnit test reports feature

opened 10 days ago by andymccall

● #75 The WASM tier: WASI jobs on any runner, phones and browsers included feature

opened 10 days ago by andymccall

● #74 Community runners: lend your idle runners to open source projects feature

opened 11 days ago by andymccall · 1 comment

● #73 syscall tier: network "loopback" brings up lo, but seccomp refuses AF_INET sockets

opened 14 days ago by andymccall

● #72 Kept build caches grow until the job's disk is full flow-next

opened 15 days ago by andymccall

● #70 Build caches in the syscall tier (narvi-1's riscv64 builds are always cold) flow-next

opened 16 days ago by andymccall

● #69 narvi-1: a RISC-V runner (syscall tier) building the riscv64 releases flow-next

opened 16 days ago by andymccall · 3 comments

● #63 Runner: one job's slow fetch holds every job on the runner bug

opened 19 days ago by andymccall

● #56 A tofu image: OpenTofu for deploy jobs

opened 20 days ago by andymccall

● #54 Fresher cache placement feature

opened 21 days ago by andymccall

● #53 Runner architecture: detect and learn feature

opened 21 days ago by andymccall

● #52 Cached builds: unchanged sources keep their timestamps

opened 22 days ago by andymccall

● #51 Cache-aware placement: jobs go where their cache is

opened 22 days ago by andymccall

● #50 Drain on restart: finish the job in hand, then restart

opened 22 days ago by andymccall

● #49 Runners ride out a hub restart

opened 22 days ago by andymccall

● #48 Swarm: TLS between peers, pinned by the hub feature

opened 23 days ago by andymccall