27 commits since v0.31.0 (17 non-merge). Cut from origin/main 2be444b3
(#361 merged).
THE NUMBER: a patch, on the user's call, and the version table does not
argue with it. Nothing versioned moved — WIRE_VERSION 2, C ABI 25 with
include/punktfunk_core.h differing from the v0.31.0 tag by one `#define`,
driver protocol 6 / min 3 (pf-driver-proto has no diff against the tag),
gamepad channel 3, plugin index schema 1, host event schema 1, gamescope
+pfhdr8 with no new patch files, SDK 0.1.5 untouched. No `!` commit and
no breaking change of any kind.
Two surfaces grow, both by pure addition, and neither forces the minor:
the management API gains PUT /library/provider/{provider}/running (#361,
already regenerated into both openapi copies there — this commit moves
only the stamp), and plugin-kit goes 0.4.3 -> 0.4.4 to export
ProviderClient.reportRunning from the package root. plugin-kit-v0.4.4 is
tagged AND published — it is the registry's `latest`; 0.4.2 is still
skipped there. A plugin that adopts neither is unaffected, and a 404 from
an older host means "this host tracks games by scanning".
The cycle is otherwise fix-shaped: fifteen fixes and two small features
— the console's own switch for turning itself off (which exists because
0.31.0 made the console the only interface Android shows with a pad
attached) and the provider running-report above.
The new control message (DeliveryReport, type byte 0x0B) is a `#define`,
not an ABI step — the same shape PipelineGap (0x0A) had at v0.30.0, and
it takes a free byte in its block rather than lengthening LossReport,
which is length-checked exactly.
Two behaviour changes are called out in the CHANGELOG rather than left
to be discovered: Windows `service install` now provisions a
program-scoped inbound UDP rule for the host exe (the black-picture
fix — no `localport=` rule can ever cover an ephemeral data port), and
the GameStream plane's default virtual-Xbox backend flips from the XUSB
companion to the UMDF HID pad, which is the correction of a default that
was only ever a module-name coincidence. Both revert with
PUNKTFUNK_XBOX_BACKEND=xusb / by removing the rule.
Also in this commit, because a cut is when it bites:
packaging/gamescope/PKGBUILD moves pfhdr7 -> pfhdr8. The banner has said
+pfhdr8 since patch 0010 and the host probes it for >= 8, but pacman
compares pkgver-pkgrel, read 3.16.25.pfhdr7-1 on both v0.30.0 and
v0.31.0, and offered Arch/SteamOS hosts no upgrade at all — so the
keyboard-layout fix never reached them. deb and rpm derive the version
from the binary banner and moved by themselves; Arch is the only channel
that hardcodes it. The v0.31.0 table flagged this as pre-existing.
api/openapi.json is RE-STAMPED here, not regenerated: punktfunk-host
does not build on macOS. #361 regenerated the document itself, with the
new route and its three schemas, on a runner where
openapi_document_is_complete_and_checked_in actually executes. This
commit moves one line (info.version, which utoipa fills from
CARGO_PKG_VERSION), and "0.31.0" appears nowhere else in either copy.
Gates: cargo fmt --all --check clean; cargo metadata --offline ok with
the Cargo.lock diff versions-only (36/36); cargo test -p punktfunk-core
--lib 273 passed; the C ABI harness passed reporting abi_version=25 —
it did NOT run on the v0.31.0 cut, so this is the first cut since ABI 25
where a C compiler has built the generated header;
scripts/ci/check-docs-drift.sh clean; the android.yml Play notes gate
run verbatim, 442/500 characters and unique; both openapi copies cmp
identical and stamped 0.31.1; notes voice scan clean.
enricobuehler