c695b1ff

By: ValentinHirschi <valentin.hirschi@gmail.com>

verify both symbolica reproducers against official main

The full unchanged GL262 builder input reproduces the same panic on official main ba3737137c2a2ccd7bb39f0441837d38ec867e78. Preserve a standalone loader adapting the new default inlining field with exact certificates, matching GL00 control, and final corrected-license receipts. The unchanged nonsymmetric import test passes upstream. Record the 32-bit product-length hypothesis and portable complete-input bundle.

Validation: standalone fmt/check/build/clippy pass; GL00 operation counts match; fresh-process import write/read pass; GL262 independently reproduces the original panic with H1/CPE5, one core and 500 GB guard, without generated evaluator compilation.

User request:
Do you have a standalone Symbolica reproducer? In a folder, as a rust-script? Because it'd be nice to see if it also fails in the latest version of symbolic, i.e. current `main` branch (and not our fork). Also test the `import_remap.rs` there (this problem should be gone).
This new symbolica version requires a new license key.

The user subsequently supplied a corrected S-format key; license values are omitted from versioned artifacts. Actual upstream environment variable is SYMBOLICA_LICENSE.

Total

35m03s

Time to Start Worker time Duration Time to finish Idle
Config 1s 8s 8s 10s 1s
Eval 7s 1m12s 1m12s 1m19s 0s
Build 1m06s 1h18m00s 26m49s 27m55s 24s
Test 6m15s 18m20s 28m26s 34m42s 0s
Deploy 34m42s 22s 22s 35m05s 4s
Suite 1s 1h38m04s 35m03s 35m05s 30s

Timeline

Config Eval Build Test Deploy Failed Retried Compressed gap
0s1m2m3m4m5m6m7m8m9m10m11m12m13m14m15m16m17m18m19m20m21m22m23m24m25m26m27m28m29m30m31m32m33m34m35m