Repository navigation
Openmvs 240 - #2074
Open
DodgySpaniard wants to merge 4 commits into
Open
Openmvs 240#2074DodgySpaniard wants to merge 4 commits into
DodgySpaniard wants to merge 4 commits into
Conversation
6 tasks
smathermather
self-requested a review
September 9, 2026 13:45
smathermather
marked this pull request as ready for review
September 10, 2026 16:07
MJohnson459
force-pushed
the
openmvs-240
branch
from
September 15, 2026 13:45
cc898c9 to
63b7b84
Compare
- Pined a commit hash (v2.4.0) - Updated pixi dependencies: cgal 5.6 -> 6, added nanoflann. - Added new arguments to preserver default behaviour from v2.2.0
We have the policy for empty variables: IF(POLICY CMP0104) # Initialize CMAKE_CUDA_ARCHITECTURES when CMAKE_CUDA_COMPILER_ID is NVIDIA. Raise an error if CUDA_ARCHITECTURES is empty. CMAKE_POLICY(SET CMP0104 NEW) ENDIF() and then: UNSET(CMAKE_CUDA_ARCHITECTURES) Setting: -DCMAKE_CUDA_ARCHITECTURES=OFF Prevents the policy violation while preserving the intent.
MJohnson459
force-pushed
the
openmvs-240
branch
2 times, most recently
from
September 16, 2026 10:45
6644b33 to
6d88ba6
Compare
The image builds have been dying on the free GitHub runners at the same point: once the small externals finish, ninja starts openpointclass, draco, opensfm, openmvs, pypopsift and poissonrecon together, each running its own cores+2 inner build. Output stops, and some minutes to an hour later GitHub reports "The hosted runner lost communication with the server". The GPU image failed four of five attempts this way, and the CPU image then failed on both amd64 and arm64 in one run. Pass CMAKE_BUILD_PARALLEL_LEVEL=3 from both Docker workflows as a build argument. cmake --build honours it at the SuperBuild level and inside every ExternalProject that builds through cmake, so the total is bounded at 9 jobs. The value stays in the workflows because it is tuned to the runner; each Dockerfile only declares the ARG, so local builds are unchanged. The native-builds workflow already sets the same variable as a step env. Two SuperBuild changes make the cap reach everything. The top-level nproc block now runs before the External-*.cmake files are included, so they can share it. External-OpenSfM.cmake no longer computes its own ProcessorCount and passes --parallel; that was left over from when OpenSfM built with plain make, and under Ninja it only overrode the cap.
MJohnson459
force-pushed
the
openmvs-240
branch
from
September 16, 2026 11:04
6d88ba6 to
ce76e41
Compare
The macOS native build failed linking DensifyPointCloud and ReconstructMesh with "library not found for -ljxl": the v2.4.0 pkg-config macro fell back to a bare -ljxl because conda-forge ships libjxl only as a shared library. The fork's omvs-v240 branch now carries upstream's fix for that macro on top of v2.4.0, and drops the upstream CI workflow that cannot run without the vcpkg manifest the fork removes.
MJohnson459
force-pushed
the
openmvs-240
branch
from
September 16, 2026 11:54
ce76e41 to
969ae57
Compare
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bumping OpenMVS to v2.4.0 branch
AI Disclaimer: LLMs were used for research and testing