Skip to content

ECO-CANON-2: RED witness for semantic callable identity - #93

Closed
juv4uk wants to merge 29 commits into
mainfrom
feat/eco-canon-2-semantic-identity-red
Closed

ECO-CANON-2: RED witness for semantic callable identity#93
juv4uk wants to merge 29 commits into
mainfrom
feat/eco-canon-2-semantic-identity-red

Conversation

@juv4uk

@juv4uk juv4uk commented Sep 12, 2026

Copy link
Copy Markdown
Owner

Superseded by PR #95, which implements the stronger architecture: language-level Value::SemanticRef(id) identity derived from the semantic registry, with Rust retained only as the current execution projection.

Why #93 is closed instead of merged:

The RED witness and historical evidence from #93 remain useful, but its implementation path is now obsolete.

juv4uk and others added 29 commits September 13, 2026 01:03
Discovered live 2026-09-13 while cross-compiling my-lisp-cli for
aarch64-linux-gnu (Android TV box oracle experiment): --target= works
only for guix build, not guix install/guix shell (unrecognized option
on both). Working path is guix build --target=<triple> pkg, then use
the resulting /gnu/store path directly. Also confirmed aarch64-linux-musl
is not a supported guix cross-compilation target (gnu/glibc only),
relevant for Android deployment since Bionic needs a fully-static
glibc binary rather than the usual musl-static approach.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…n mesboot)

guix build --target=aarch64-linux-gnu gcc-toolchain fails at the
gawk-mesboot-3.1.8 configure phase: config.sub does not recognize
aarch64. Confirmed reproducible before AND after guix pull (0cc8f41 ->
a39872ef5) -- the frozen mesboot bootstrap tarball is not updated by
channel pulls, since it exists specifically to stay frozen for
bootstrap reproducibility. Practical implication: full-source-bootstrap
Guix cross-compilation to aarch64-linux-gnu is not viable on this host
without patching mesboot or sourcing a prebuilt substitute elsewhere.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Live-verified 2026-09-13: my-lisp --tcp oracle deployed and confirmed
working on an Android TV box (X96 Max Plus2), not just the usual
x86_64 droplet. Key finding: the box's Android userspace is 32-bit
only (armeabi-v7a) despite a 64-bit-capable SoC -- an aarch64 build
fails with a misleading not-executable error unrelated to noexec/
SELinux; armv7-linux-androideabi via NDK r30 fixed it immediately.
Full TCP oracle-check round-trip confirmed through adb forward, and a
simple shell watchdog loop confirmed to actually restart the process
after kill -9 within 2s (new server-generation observed). Reboot
persistence intentionally left unsolved (would need Termux:Boot APK
install, not done without separate owner permission).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…onfirmed live

Two real device reboots (not simulated): first reboot after installing
termux-boot via adb and writing ~/.termux/boot/start-oracle.sh produced
zero autostart -- Android keeps never-manually-opened apps in a
stopped state that blocks even correctly-registered BOOT_COMPLETED
receivers. Fixed via one-time  on both com.termux and
com.termux.boot launchable activities. Second real reboot confirmed
the oracle auto-starts under the correct Termux uid with no manual
intervention, and answers a real TCP oracle-check round-trip
immediately after boot.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Align Ukrainian staging profile with canonical .lisp source extension
…tatus doc

Direct TCP oracle-check round-trips confirm three live my-lisp
oracles simultaneously: local WSL host, the droplet (which
docs/swarm-remote-access.md still lists as DOWN from its 2026-08-13
audit -- stale, someone revived it since without updating the address
book), and the new Android TV box deployment. Practical lesson: verify
oracle status via a live TCP request, not by trusting the address
books last-audit date.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@juv4uk juv4uk closed this Sep 13, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant