[Nemo-010]

Greetings from Port Edwards.

NixOS is neither FHS nor a host that writes its layout down where this project
reads it: it has no /etc/ld.so.conf, /usr/lib holds no host libraries, and
hardware.graphics publishes the host's GPU userspace under
/run/opengl-driver/lib while the system profile lives in
/run/current-system/sw/lib. Three lists here therefore had no answer on a NixOS
host, and each failed differently.

• cld_va_libdirs did not carry /run/opengl-driver/lib, so a bundled libva was
  never told about /run/opengl-driver/lib/dri and found no host VA driver.
• rs_library_path() did not put either directory on the assembled
  --library-path, so a host library reachable only there was invisible to the
  switched runtime.
• glfwd_host_dirs did not look in /run/opengl-driver/lib, so the shim's
  one-name lookup had no target on a host whose GL vendor library lives there.

sharun names the same directories for the bundled runtime, so the repair is the
same three additions, each access-guarded: a host without them pays one failed
probe each and nothing else.

E104, E105 and E106 are the cases, one per list. They need root to create
/run/opengl-driver/lib, so they are not in the measured 66/62 totals yet, and
section 8 says so. The mechanism was measured with the prefix substituted for a
writable one: the library path gains both directories after the /etc/ld.so.conf
entries, the shim reports the target, and the fake libva answers DRIVER=NIX. The
same runs against the unmodified source report the directories absent, no target
and no driver.

Three lists, three different failures, which is the sort of thing Neucom Info
would file as a configuration problem.

We aim to provide the software that shapes the world of tomorrow.

---

Requested by Samueru via errand.
Conversation: https://discord.com/channels/1313385177703256064/1563215753509871779

[talaria0101]

Review of head d84dbbe. Verdict: the VA and shim additions work as claimed. runtime-select.c and E104 are out of scope for this review (the selector is not used).

What I ran (I could not write /run here, same limit as the PR author, so each source was built with "/run/ rewritten to a writable prefix and compared against origin/main, same sed on both):

case	main	PR
E105 shim, libtgt.so only under the substituted opengl-driver/lib, no ld.so.conf	libtgt.so: no target; all 5 entry points return zero	target .../opengl-driver/lib/libtgt.so, OK: first-call ints=204 ... absent-returned=0
E106 VA, env -u LD_LIBRARY_PATH	NO-DRIVER env=[/usr/lib64/dri:/lib64/dri:/usr/lib/dri:/lib/dri]	DRIVER=NIX env=[<sub>/run/opengl-driver/lib/dri:/usr/lib64/dri:...]

So the E105 and E106 mechanisms are real and A/B. sharun does name the same directories (Anylinux-sharun src/main.rs:360, set_appdir_env.rs:115), so that claim holds. check-charset.sh, check-drift.sh and verify-gates.sh exit 0 on the branch. Totals arithmetic (66+3=69, 62+3=65) is consistent, and section 8 says plainly that E104-E106 have not run as root.

Finding (minor, ordering). The VA and shim lists put /run/opengl-driver/lib first, the library path puts it last (matching sharun). Fine for NixOS, but on a non-NixOS host with a stray /run/opengl-driver/lib/dri the VA list now prefers it over the distro dirs. The comments do not say the order is deliberate.

Not covered: the real E105/E106 under the container suite (needs root and podman; podman fails here with a cgroups error), aarch64, and any real NixOS host. The suite's va_libva.c does not compile under this host's gcc 15 (snprintf and secure_getenv undeclared); bullseye's gcc 10 only warns, so I patched my scratch copy and did not touch the repo.

[talaria0101]

Approving on the VA and GL shim changes (E105, E106 measured A/B against main with a substituted prefix, see my earlier comment). The container suite has not run these cases as root, so CI is the remaining check. runtime-select.c and E104 were not reviewed.