By: ValentinHirschi <valentin.hirschi@gmail.com>
reject nonfinite stability probes at every precision Check raw probes before the single-sample shortcut and at the final precision in both existing stability checks. Preserve the primary value for the existing validity flag and sanitization, but do not assign a finite accuracy or Stable status to a nonfinite comparison. Validation: formatting, locked workspace/all-target check and focused clippy passed; four tests passed under the repository's curated test_gammaloop profile. A broader selection also exposed the unchanged quarantined zero-duration renderer fixture (ns expected, microseconds rendered); no expectations changed. Nineteen full-GL638 replays passed: 17 finite controls retain identical values and flags, while both exact A/P failures now end Arb Unstable with no accuracy. This corrects reporting; it does not cure the coincident-residue evaluation. User requests: Continue as planned, but look into the current implementation, it should already offer such flexibility. But feel free to add some if you need be (but make sure it's steered through runtime integrand toml parameters so the user can easily access it). And it's only weakly tested, so if you find any bug there you can also adjust it. 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.