ffmpeg-v4l2-request-fourier: PKGREL 3 → 5 (force rebuild past orphan -4 .deb)

PR #76 (H.264 IDCT 4×4 daedalus-fourier substitution) was merged but
the resulting .deb was not actually built: an orphan
ffmpeg-v4l2-request-fourier_8.1+rfourier+gb57fbbe-4_arm64.deb (dated
2026-05-19, no matching source commit in main) sat in the apt pool.
.gitea/scripts/check-already-published.sh's debian branch compares
`dpkg --compare-versions $pool_ver ge $source_full` — pool -4
≥ source -3, so CI's skip-check emitted skip=1 and short-circuited
the build.  The ffmpeg-v4l2-request-debian Action reported success
without actually publishing.

Bump source PKGREL past -4 so the next CI run sees source >= pool
and proceeds to build + publish.

No source code change beyond PKGREL + changelog.  Arch side
unaffected (its skip-check is exact-URL-match, not pool-head-ge).

debian/ffmpeg-v4l2-request-fourier/build-deb.sh

@@ -33,7 +33,10 @@ FFMPEG_VERSION=8.1
 # epoch 2 matches Debian's stock ffmpeg (currently 7:7.1.x in trixie);
 # +rfourier suffix to avoid colliding with upstream/Debian rebuilds.
 PKGVER=2:${FFMPEG_VERSION}+rfourier+gb57fbbe
-PKGREL=3  # pkgrel=3 — H.264 IDCT 4x4 daedalus-fourier substitution (2026-05-21)
+PKGREL=5  # pkgrel=5 — H.264 IDCT 4x4 daedalus-fourier substitution; skip past
+          # an orphan -4 .deb sitting in the apt pool that made
+          # check-already-published.sh's `pool_ver ge source_full` short-
+          # circuit the previous -3 build (PR #76).  (2026-05-21)
 
 # daedalus-fourier pin — first kernel substitution in libavcodec (cycle 6
 # H.264 IDCT 4x4).  Same SHA as the daedalus-v4l2 daemon already ships