b04fd885

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

fix LU initialization and preserve empty integrated-UV residues

Skip constant radial energies when choosing LU guesses, validate every active
cut root before constructing cut kinematics, and handle representationally
exhausted brackets through the existing root-consistency checks. Preserve
typed-zero cut-order support for empty direct-3D integrated-UV projections.

Update agent guidance so routine fixture/obsolete-expectation fixes do not
require confirmation; non-trivial or deep expectation changes still do.

Validation: 23 focused tests, workspace cargo check, clippy, and formatting
passed. The unchanged physical GL638 integrated-UV reproducer and focused
zero-projection regression passed on the phase-fix branch, where that UV fix
was already pushed as cbcffb4aed35b19da5b626a0a706564f9ac385df.

Current user request (full):
Ok commit+push what you got here?
Then us there anyway you can think of improving the situation regarding the remaining integrable singularity here? What kind of sampling would be needed, can you experiment with various solutions to see if something can further improve the situation? Especially around max weights found?
You can now do longer integration run to monitor the behaviour, on 20 cores. Make sure you don't waste tokens though when waiting for the outcome.
GitHub
Time to Start Worker time Duration Time to finish Idle
Config 1s 8s 8s 10s 1s
Eval 7s 1m08s 1m08s 1m16s 0s
Build 1m08s 2h50m39s - - 26s
Test 1m31s 30m35s - - 5m03s
Deploy - - - - 1h32m39s
Suite 1s 3h22m31s - - 1h38m10s

Timeline

Config Eval Build Test Failed Retried Compressed gap
0s1m2m3m4m5m6m7m8m9m10m11m12m13m14m15m16m17m18m19m20m21m22m23m24m25m26m27m28m29m30m31m32m33m34m35m36m37m38m39m40m41m44m45m46m47m48m49m50m51m52m