xtc/docs/MAINTENANCE.md

# Release and maintenance policy

## Supported development target

Initial target: x86_64 Artix/Arch-family Linux with glibc, GNU Make, GCC/Clang,
FLTK 1.4 and a JACK-compatible client/server. The inspected host uses GCC 16.2.1,
FLTK 1.4.5 and the dependencies in LICENSES.md. Core-only/no-JACK/no-LADSPA builds
are checked too. Build on the destination distribution for different library ABIs;
there is no cross-distribution static-binary promise. Python 3.12+ and binutils
are build/package/test tools only. No dependency is fetched automatically.

## Release checklist

1. Work from a clean reviewed commit. Record `git rev-parse HEAD`, compiler,
   libc/FLTK/JACK versions and `xtc-check --diagnose`. No automatic version bump,
   Git tag, public upload or external release is performed by local targets.
2. Run warning-enabled `make all test gui test-plugins test-plugins-real
   test-gui-plugins test-workflow test-gui-jack test-jack-midi test-recording
   test-recovery test-reliability test-scale analyze` where dependencies exist.
3. Use a separate BUILD_DIR for ASan/UBSan/leak runs and another for optional-
   dependency builds. GUI leak checks use only the documented fontconfig
   suppression; core tests use none. Run publication/recording race tests as
   described in RELIABILITY.md. Sanitizers are not latency qualification.
4. Run `make benchmark test-performance test-stress`; compare on matching
   host/compiler/flags against a reviewed baseline. Investigate regressions;
   do not weaken thresholds or replace evidence without explanation.
5. Run `make test-install test-release test-package`. Inspect archive manifests,
   hashes, dynamic dependency names and licenses. Source exports rebuild twice
   with identical artifacts on this host. Repeat binary packaging compares bytes.
6. Complete physical listening, recording/device/endurance, and independent human
   install/workflow checks before claiming production qualification. They remain
   pending for this prototype, even when all automated checks pass.
7. Save checksums/test logs/conditions with the proposed release. Review known
   limitations and format compatibility and update user/developer docs. Publishing
   or distro repository submission is a separate action.

## Regressions and compatibility

Report the exact version/commit, OS/toolchain, compiled features, reproduction,
expected/actual output, GUI errors/xruns and whether plugins are involved. Keep a
minimal project/media fixture when redistribution is permitted; do not attach
private recordings or full memory dumps inadvertently. No telemetry/upload occurs.

A confirmed defect gets a focused deterministic regression or repeatable runtime
check before its fix. Memory/thread/file-loss defects take priority; disable the
unsafe path rather than silently dropping media or unknown project records. Keep
candidate preparation/plan publication transactional and callbacks bounded.

Format changes increment XTC_PROJECT_FORMAT_VERSION, retain explicit bounded
migrations/fixtures for all supported older versions, and add malformed/roundtrip
checks. Unknown versions/types/records fail with actionable errors. Never rewrite
an unsupported source file during failed Open. Keep the original when diagnosing
post-publication sync errors. Native descriptor ABI compatibility is distinct from
project format and is documented in NATIVE_API.md; opaque DSP state is not promised.
No public stability guarantee or support SLA is implied by development version 0.1.0.

## Phase 30 support and triage record

Verified compilers are GCC/G++ 16.2.1 and Clang/Clang++ 23.1.3 on x86_64 Linux;
these are tested versions, not a promise that arbitrary older/newer versions work.
The code requires C99 plus GCC/Clang lock-free atomics, C++17 for the GUI, FLTK
1.4 and POSIX/glibc APIs. JACK (including PipeWire's JACK compatibility) is the
only live backend. Native Windows/macOS, other architectures and direct ALSA
backends are not qualified. Core builds can disable JACK, FLTK and LADSPA.
See BUILD.md and LICENSES.md for dependency roles and versions.

Release qualification adds `test-install-layout`, installed example inspection,
all performance report resources, source checksum verification and the current
RELEASE_NOTES.md. Retain the original graph benchmark alert in release records;
a successful compile does not clear timing qualification gaps.

File reproducible bugs with the repository maintainers using the source commit
and archive SHA-256, `xtc-check --version`/`--diagnose`, distro/kernel and toolchain,
exact commands or GUI steps, expected versus observed behavior and frequency.
Reduce to a redistributable project and state whether it reproduces without
external plugins and with the offline/core tools. Include relevant stderr and
sanitizer traces; preserve original recovery files before experimenting.

Maintainers should first confirm version/features and reproduce in an isolated
build. Prioritize data loss/security/memory/thread defects, then crashes and
callback/deadline failures, then functional bugs and UI issues. Record severity,
reproduction conditions, affected versions, regression test and verification
command with the fix. Route timing-only alerts through matching-host reports and
repeat measurements; do not treat sanitized timings as realtime evidence.