Samuel Cui
a420324adb
Both were invisible on Darwin: one needs a real Linux ENOSPC, the other needs the command's own stderr. - A non-linear target pre-allocates with fallocate, and a failure there was reported as a bare error: it never ran through mappingError, never aborted the device and never ended a linear target, unlike every neighbouring path. A volume that fills up after the disk-usage snapshot therefore reported a target that was "not written" instead of ErrTargetNoSpace. Classify it like the rest. TestFullVolumePreallocationReportsNoSpace pins it with a real ENOSPC on a small tmpfs (it fails on the previous revision), and TestNonLinearTargetOnRefusingDeviceReportsATargetFailure records that a device refusing fallocate reports its own refusal rather than claiming no space. The drain test now uses a linear target, which is the shape that actually queues buffers before the write fails. - cmd/acp registered the progress bar and then the report collector, and a run keeps one event handler, so the bar received nothing: the command printed its construction frame and no update. Compose the two handlers in the command and pin it end to end (the new test fails on the previous revision). README now says a command that wants both composes them. - item.go still promised that an empty Targets slice hashes the source; the default HashOff reads nothing. TESTING.md and TODO.md now describe the Linux suite that actually exists.
1.4 KiB
Open work
Review cmd/acp-rewrite on its own terms
The tool exists for a real operational need: after a ZFS pool gains vdevs, existing files keep their old block layout, so rewriting them spreads the data across the new topology. The v0.2.1 API reshape moved it onto the new surface as-is; it has never been reviewed on its own terms.
To settle: whether it stays a separate binary or becomes a subcommand of acp; its flag set
(-dryrun, -state, -report, -report-indent, -ignore); the resumable state and report
formats, now that a report must round-trip through encoding/json as a public document;
hardlink-group handling and duplicate reporting; the rewrite order it uses for a linear or
object-backed target; and whether the existing suite — rewriteFile, the state and report round
trips, temporary-file cleanup and the resume run are covered — needs a real-device end-to-end
case. It also needs a README section of its own: today README.md documents its report behaviour
and AGENTS.md describes what it is for.
Platform verification still owed
Linux-specific runtime paths run on the isolated Linux acceptance host (go test for the
Linux-only file, including the /dev/full and filled-tmpfs ENOSPC cases, O_NOATIME and its
permission fallback, fallocate, madvise). Windows and FreeBSD stay compile-verified only until
a runtime is available (mmap_other.go, signature_xattr_freebsd.go, Windows path ordering).