leonarski_fandClaude Opus 4.8
35607057d9
Build Packages / build:windows:cuda (push) Failing after 4m30s
Build Packages / build:windows:nocuda (push) Successful in 10m0s
Build Packages / build:rpm (ubuntu2404_nocuda) (push) Successful in 13m43s
Build Packages / build:rpm (rocky8_nocuda) (push) Successful in 14m46s
Build Packages / build:rpm (rocky8_sls9) (push) Successful in 14m53s
Build Packages / build:rpm (rocky8) (push) Successful in 14m54s
Build Packages / build:rpm (ubuntu2204_nocuda) (push) Successful in 15m10s
Build Packages / build:rpm (rocky9_nocuda) (push) Successful in 15m34s
Build Packages / build:rpm (rocky9_sls9) (push) Successful in 15m37s
Build Packages / XDS test (durin plugin) (push) Successful in 8m5s
Build Packages / Generate python client (push) Successful in 31s
Build Packages / XDS test (JFJoch plugin) (push) Successful in 8m50s
Build Packages / Create release (push) Skipped
Build Packages / XDS test (neggia plugin) (push) Successful in 8m44s
Build Packages / Build documentation (push) Successful in 1m7s
Build Packages / build:rpm (ubuntu2404) (push) Successful in 10m45s
Build Packages / build:rpm (ubuntu2204) (push) Successful in 12m5s
Build Packages / build:rpm (rocky9) (push) Successful in 13m8s
Build Packages / DIALS test (push) Successful in 12m47s
Build Packages / Unit tests (push) Successful in 1h0m40s
MergeOnTheFly::AddImage picked each image's CC1/2 half from a shared mt19937 drawn in call order (and before Mask), so the split depended on iteration/thread order and on how many images were masked. The class is mutex-guarded for concurrent "on-the-fly" use, so any parallel merge would make CC1/2 non-reproducible - a latent race. Assign the half as a splitmix64 hash of the image's stable index instead, computed after Mask. The split is now reproducible run-to-run, independent of AddImage order, parallel-safe, and decoupled from masking. Callers pass the outcome's vector index as the image id. Verified: lyso_ref two-pass -M -P rot3d gives identical CC1/2 across runs (overall 99.6%, P41212); hash split is balanced ~50/50. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>