libdrm, libpciaccess, openexr, pngquant3: state compiler floors, fix the meson c_std overrides

[th0ma7]

Rebased on master, and split: the cairo and glib half moved to #7427, which carries the larger dependency fan-out. This PR keeps the meson c_std overrides, the two remaining floors, and the tvheadend ffmpeg bump. It will rebase onto #7427 once that lands.

Groundwork for #7391.

libdrm needed no floor at all

Its meson.build asks for c_std=c11 and gcc < 4.7 only knows -std=c1x, so it died on "unrecognized command line option '-std=c11'" — an option to override, not a compiler to outgrow. With -Dc_std=gnu99 it builds on gcc 4.6.4 and on ppc853x's 4.3.7, so UNSUPPORTED_ARCHS = $(ARMv5_ARCHS) $(OLD_PPC_ARCHS) is dropped outright rather than restated as a floor.

This is exactly the failure mode a version floor is supposed to avoid. That Makefile stated no reason, and an earlier revision of this branch inferred 4.8 from its exclusion set; the inference was wrong.

libpciaccess had the same override already, but written in two broken ways: the condition read version_ge 4.9, handing gnu99 to the compilers that do not need it, and it sat above the include where TC_GCC does not exist yet, so it never fired either way. Now version_lt 5, between spksrc.common.mk and spksrc.cross-meson.mk.

The two packages that do keep a floor

package	was	now
openexr	$(ARMv5_ARCHS) $(OLD_PPC_ARCHS) + UNSUPPORTED_ARCHS_TCVERSION = x86-5.2	MIN_GCC_VERSION = 4.8
pngquant3	$(ARMv5_ARCHS) $(OLD_PPC_ARCHS)	MIN_GCC_VERSION = 4.8

Both said so themselves — "Toolchains lacking c++11 compiler", "Fails on older gcc for its internal dependencies". 4.8 is measured on the Synology toolchains, not taken from upstream release notes. Core C++11 — unique_ptr brace-init, std::mutex, lambdas, -std=c++11:

gcc 4.3.7  no      gcc 4.6.4  no      gcc 4.7.3  no      gcc 4.8.3  yes      gcc 4.9.3  yes

openexr also loses its x86-5.2 exception — that arch runs gcc 4.7.3 and fails the same check, so the floor already covers it.

Blast radius

No architecture is freed except on libdrm, which gains ARMv5 and old PowerPC. Every newly excluded one is a DSM 5.2 arch below the floor:

alpine  armada370  armada375  armadaxp  avoton  braswell  bromolow
cedarview  comcerto2k  evansport  qoriq  x64  x86         (all -5.2, gcc 4.3.7 … 4.7.3)

They were never named because DSM 5.2 is not built in CI, so nobody hit them. DSM 6.x and 7.x are untouched.

build.yml turns DSM 5.2 on temporarily so CI exercises exactly those archs. That commit reverts before merge.

The DSM 5.2 toolchain fixes (cross/m4, cross/libmaxminddb, the videodriver meta selector) are the same commit as in #7427 — they are needed under both halves while 5.2 is switched on, and the rebase will absorb it once #7427 lands. Details there.

tvheadend

OPTIONAL_DEPENDS and the DEPENDS fallback follow cross/ffmpeg7 → cross/ffmpeg8.

Deliberately left as architecture lists

• llvm-140, Khronos-SPIRV-LLVM-Translator-140 — x86_64 host tooling, a genuine architecture restriction.
• openh264 — no PowerPC support upstream.
• c-ares-latest — needs IN_NONBLOCK, which no glibc floor can express: x86-5.2 ships glibc 2.17 without it while 88f6281 ships 2.15 with it. The Synology toolchains are not uniform at equal version, so a floor would wrongly admit x86-5.2. Worth noting that x86-5.2 is therefore already broken for this package today, silently.

Follow-on

Once #7391 lands, five archs regain these packages through the overlay: 88f6281-5.2, 88f6281-6.2.4, ppc853x-5.2, x64-5.2, x86-5.2.

nghttp3 is not part of this set. It reaches tvheadend only through cross/curl's OPTIONAL_DEPENDS, and cross/tvheadend sets BUILD_CURL_WITH="gnutls libssh2 zstd" without it; it appeared here because the dependency walk was run without ARCH, which returns the optional superset. Its floor also turned out to be 4.9, not 4.8 — nghttp3 1.x builds its min/max helpers on _Generic, which gcc only implements from 4.9. That landed with the version bump in #7399.

ffmpeg5 and ffmpeg6

Both cross/ and spk/ Makefiles carried

# requires c11 and proper atomic using gcc-4.9+ support
UNSUPPORTED_ARCHS = $(ARMv5_ARCHS) $(ARMv7L_ARCHS) $(OLD_PPC_ARCHS)

— a comment naming 4.9 above a list that does not say 4.9. ffmpeg7 and ffmpeg8 already declare MIN_GCC_VERSION = 4.9; all four versions now express it the same way.

Nothing is freed: every arch the list named is below 4.9 at every DSM it has a toolchain for (88f6281 4.6.4, hi3535 4.8.3, ppc853x 4.3.7), and powerpc, ppc824x, ppc854x have no toolchain at all. What the floor adds is the fourteen toolchains the list silently admitted — qoriq-5.2 4.3.7, evansport-5.2 4.6.3, alpine/armada370/armada375/armadaxp/comcerto2k-5.2 4.6.4, avoton/braswell/bromolow/cedarview/x64/x86-5.2 4.7.3, monaco-5.2 4.8.3. All DSM 5.2, all below the stated requirement, none ever built. qoriq is the one that could have been hit outside 5.2: it is a PowerPC arch but not in OLD_PPC_ARCHS, so the list never covered it.

This lands here rather than separately because the videodriver-meta fix lets ffmpeg6 get past spk-stage1 on x86-5.2, where it would then reach the compiler its own comment says it cannot use.

🤖 Generated with Claude Code

https://claude.ai/code/session_01JiBhJxRTbjD6HWSEjzE41g