ENH: Ingest ITKCuberille into Modules/Filtering by hjmjohnson · Pull Request #6205 · InsightSoftwareConsortium/ITK

hjmjohnson · 2026-05-04T18:47:35Z

Ingests Cuberille from the standalone remote module InsightSoftwareConsortium/ITKCuberille into ITK at Modules/Filtering/Cuberille/ using the v4 ingestion pipeline (PR #6204). Stacked on #6204 — that PR must merge first.

Local validation: pre-commit run --all-files clean; CuberilleTestDriver builds; 25/25 Cuberille tests pass locally in 5.3s.

Pipeline metrics

Metric	Value
Upstream commits → rewritten	200 → 106
Upstream merges → preserved	59 → 30 (no linearization, per `feedback_ingest_merge_topology.md`)
Blobs scanned in sanitize	148 (cxx=103, py=1, cmake=28, text=5, skip=11)
Blobs reformatted	95 (clang-format / black / gersemi / trailing-ws / EOF)
Commit subjects auto-prefixed	36
Merges visible in `git log -- Modules/Filtering/Cuberille`	6

v4 deny-pass bug surfaced and fixed during this run

The first attempt at this ingest leaked test/Docker/{Dockerfile,build.sh,run.sh,test.sh} into ITK history, because the v4 ingest driver had only a whitelist pass — no --invert-paths --path-glob deny-pass for CI/packaging scaffolding that lives inside whitelisted directories. Fixed in PR #6204 by porting v3's deny-glob list (Docker, Jenkinsfile, .github, azure-pipelines, .clang-format inside the module subtree, etc.). Re-running this ingest after the fix produced a clean tree with no scaffolding leakage.

Auto-prefix decisions (sample)

prefix counts across the 36 changed subjects:
  ENH:   28
  STYLE:  4
  COMP:   2
  BUG:    1
  DOC:    1

Every commit on the rewritten history starts with a valid ITK prefix (BUG:/COMP:/DOC:/ENH:/PERF:/STYLE:/WIP:). The Utilities/Maintenance/RemoteModuleIngest/sanitize-history.py log records the (sha, chosen-prefix, original-subject) for each decision so the operator can audit before merge.

Follow-up commits in this PR

After the merge commit (ENH: Ingest ITKCuberille into Modules/Filtering), the standard cleanup commits land on top:

STYLE: Apply gersemi 0.19.3 to Cuberille wrapping test CMakeLists. The v4 sanitize pass runs gersemi 0.24.0; ITK's pre-commit pins 0.19.3. The post-merge pre-commit run --all-files gate caught one CMake file that 0.24.0 considered clean but 0.19.3 wanted re-formatted. Working as designed.
DOC: Add Modules/Filtering/Cuberille/README.md pointing at the archived upstream.
COMP: Remove Modules/Remote/Cuberille.remote.cmake (now in-tree).
ENH: Enable Module_Cuberille:BOOL=ON in pyproject.toml configure-ci.

Phase B (upstream archival) — deferred

Per the v4 design (PR #6204), the upstream archival step is intentionally separated. After this PR merges into ITK main, run:

~/src/ITK/Utilities/Maintenance/RemoteModuleIngest/archive-remote-module.sh \
    Cuberille --itk-pr-number <THIS-PR-NUMBER>

That step pushes the deletion + MIGRATION_README.md → README.md promotion to upstream and flips archived=true. This PR makes no upstream changes; the upstream ITKCuberille repository remains live until the operator manually runs Phase B.

QuadricDecimationQuadEdgeMeshFilter.

Refactor as an ITKv4 module.

Add Docker test configuration.

ENH: Add wrapping.

COMP: Remove unused typedefs in the test.

Addresses: /home/matt/src/ITK/Modules/Remote/Cuberille/include/itkCuberilleImageToMeshFilter.h: In instantiation of ‘class itk::CuberilleImageToMeshFilter<itk::Image<unsigned char, 3u>, itk::Mesh<unsigned char, 3u>, itk::LinearInterpolateImageFunction<itk::Image<unsigned char, 3u> > >’: /home/matt/src/ITK/Modules/Remote/Cuberille/test/CuberilleTest01.cxx:123:16: required from here /home/matt/src/ITK/Modules/Core/Common/include/itkProcessObject.h:522:16: warning: ‘virtual void itk::ProcessObject::SetInput(const DataObjectIdentifierType&, itk::DataObject*)’ was hidden [-Woverloaded-virtual] virtual void SetInput(const DataObjectIdentifierType & key, DataObject *input); ^ In file included from /home/matt/src/ITK/Modules/Remote/Cuberille/include/itkCuberilleImageToMeshFilter.h:347:0, from /home/matt/src/ITK/Modules/Remote/Cuberille/test/CuberilleTest01.cxx:27: /home/matt/src/ITK/Modules/Remote/Cuberille/include/itkCuberilleImageToMeshFilter.hxx:53:1: warning: by ‘void itk::CuberilleImageToMeshFilter<TInputImage, TOutputMesh, TInterpolator>::SetInput(const InputImageType*) [with TInputImage = itk::Image<unsigned char, 3u>; TOutputMesh = itk::Mesh<unsigned char, 3u>; TInterpolator = itk::LinearInterpolateImageFunction<itk::Image<unsigned char, 3u> >; itk::CuberilleImageToMeshFilter<TInputImage, TOutputMesh, TInterpolator>::InputImageType = itk::Image<unsigned char, 3u>]’ [-Woverloaded-virtual] CuberilleImageToMeshFilter<TInputImage,TOutputMesh,TInterpolator> ^

COMP: Address SetInput override virtual warning.

BUG: Address KWStyle tests.

The defined directory is not correct when built as a Remote module.

BUG: Fix test temporary output dir.

COMP: Wsign-compare warning

…ITKNULLPTR STYLE: Null to ITK_NULLPTR

dzenanz

Mostly looks good.

Brings Cuberille from a configure-time remote fetch into the ITK source tree at Modules/Filtering/Cuberille/ using the v4 ingestion pipeline (whitelist filter-repo + per-commit clang-format + black + commit-prefix sanitization). Upstream repo: https://github.com/InsightSoftwareConsortium/ITKCuberille.git Upstream tip: 4ba447f256a93428e56507178f92124bb91e3f3c Ingest date: 2026-05-04 Whitelist: Cuberille.list Per-commit transforms applied across all 105 commits: - filter-repo --paths-from-file (whitelist) - filter-repo --to-subdirectory-filter Modules/Filtering/Cuberille - clang-format -style=file (ITK main's .clang-format) for *.cxx/.h/.hxx/... - black for *.py - heuristic ITK prefix added to commit subjects without one Merge topology preserved: 59 -> 30 merge(s). Primary author: Matt McCormick <matt.mccormick@kitware.com> Co-authored-by: Davis Marc Vigneault <davis.vigneault@gmail.com> Co-authored-by: Davis Vigneault <davis.vigneault@gmail.com> Co-authored-by: DVigneault <davis.vigneault@gmail.com> Co-authored-by: Dženan Zukić <dzenan.zukic@kitware.com> Co-authored-by: Hans Johnson <hans-johnson@uiowa.edu> Co-authored-by: Hastings Greer <hastings.greer@kitware.com> Co-authored-by: Jon Haitz Legarreta <jhlegarreta@vicomtech.org> Co-authored-by: Jon Haitz Legarreta Gorroño <jon.haitz.legarreta@gmail.com> Co-authored-by: Mathew Seng <mathewseng@gmail.com> Co-authored-by: Matt McCormick <matt@mmmccormick.com> Co-authored-by: root <root@insight-journal.org> Co-authored-by: Tom Birdsong <tom.birdsong@kitware.com> Co-authored-by: Zach Williamson <zachary-williamson@uiowa.edu>

The four implementation-control macros (DEBUG_PRINT, USE_GRADIENT_RECURSIVE_GAUSSIAN, USE_ADVANCED_PROJECTION, USE_LINESEARCH_PROJECTION) were defined unconditionally at the top of itkCuberilleImageToMeshFilter.h. Any translation unit including the public header silently inherited DEBUG_PRINT=0, masking a same-named macro any consumer might want to test with #ifdef. Move the four #define lines to itkCuberilleImageToMeshFilter.hxx so they remain visible to the template implementation that depends on them via #if, but are no longer emitted into consumer translation units that only include the .h. Greptile P1 on PR InsightSoftwareConsortium#6205.

Both operators applied >= / <= on m_Y in their first branch, which short-circuited to true/false whenever the Y values were equal — before the X tiebreaker could be evaluated. E.g. {Y=1,X=0} >= {Y=1,X=5} resolved (1 >= 1) == true and returned true even though the lhs is strictly less than the rhs. Replace the first-branch comparison with strict > / < so equal-Y cases fall through to the X tiebreaker, mirroring the existing operator> and operator< definitions. std::map calls only operator<, so existing callers do not observe the bug, but the broken operators are class-level API and would silently misbehave under direct use. Greptile P1 on PR InsightSoftwareConsortium#6205.

Convert the 20 Cuberille tests registered via plain add_test() to itk_add_test() so they pick up the standard ITK CTest labels and properties (matching the three already-itk_add_test'd entries below them). Tests with bare add_test are invisible to dashboard-specific test selectors that filter on those labels. GTest conversion of these tests is a separate follow-up. Greptile P2 on PR InsightSoftwareConsortium#6205.

DEBUG_PRINT was a debug-tracing toggle defined to 0 (compile-time disabled) since the module's first commit; the eight #if DEBUG_PRINT...#endif blocks they guarded contained std::cout traces from algorithm bring-up that were never re-enabled. Removing the macro and the dead code per the reviewer's preference for cleaner module headers. Behavior is unchanged because all guarded blocks were #if 0.

Replaces three preprocessor toggles USE_GRADIENT_RECURSIVE_GAUSSIAN USE_ADVANCED_PROJECTION USE_LINESEARCH_PROJECTION with class-level `static constexpr bool` members of the same intent. Preprocessor `#if FLAG` becomes `if constexpr (Flag)` and a `#if/#else` selecting between two GradientFilterType type aliases becomes `std::conditional_t<...>`. Eliminates global-namespace preprocessor pollution flagged in the Greptile draft review (P1) and matches dzenanz's reviewer preference for class-level configuration over file-level macros. Default values remain false (matching the prior `#define X 0`) so behavior is unchanged. The dead linesearch branch had a stray top-level `break;` outside any loop (a long-standing bit-rot in the disabled-since-2014 alternative path). Replaced with `return;` (the void function's natural early-out) so the branch parses cleanly under `if constexpr` instantiation rules.

The four implementation-control macros (DEBUG_PRINT, USE_GRADIENT_RECURSIVE_GAUSSIAN, USE_ADVANCED_PROJECTION, USE_LINESEARCH_PROJECTION) were defined unconditionally at the top of itkCuberilleImageToMeshFilter.h. Any translation unit including the public header silently inherited DEBUG_PRINT=0, masking a same-named macro any consumer might want to test with #ifdef. Move the four #define lines to itkCuberilleImageToMeshFilter.hxx so they remain visible to the template implementation that depends on them via #if, but are no longer emitted into consumer translation units that only include the .h. Greptile P1 on PR InsightSoftwareConsortium#6205.

Both operators applied >= / <= on m_Y in their first branch, which short-circuited to true/false whenever the Y values were equal — before the X tiebreaker could be evaluated. E.g. {Y=1,X=0} >= {Y=1,X=5} resolved (1 >= 1) == true and returned true even though the lhs is strictly less than the rhs. Replace the first-branch comparison with strict > / < so equal-Y cases fall through to the X tiebreaker, mirroring the existing operator> and operator< definitions. std::map calls only operator<, so existing callers do not observe the bug, but the broken operators are class-level API and would silently misbehave under direct use. Greptile P1 on PR InsightSoftwareConsortium#6205.

Convert the 20 Cuberille tests registered via plain add_test() to itk_add_test() so they pick up the standard ITK CTest labels and properties (matching the three already-itk_add_test'd entries below them). Tests with bare add_test are invisible to dashboard-specific test selectors that filter on those labels. GTest conversion of these tests is a separate follow-up. Greptile P2 on PR InsightSoftwareConsortium#6205.

…est-cuberille ENH: Ingest ITKCuberille into Modules/Filtering

Fixes the four classes of ghostflow errors that surfaced on the first Cuberille v4 ingest (PR InsightSoftwareConsortium#6205) — the only allowable remaining error should be the unavoidable upstream root commit. - Strip *.orig / *.rej / *.BACKUP.* / *.LOCAL.* / *.REMOTE.* / *.BASE.* merge-conflict-artifact files from history (catches leftover-conflict-marker errors like the .h.orig on Cuberille). - sanitize-history.py: cap every commit subject at 78 chars (ghostflow / kw-commit-msg rule); excess words move into the body. - sanitize-history.py: insert a blank line between subject and body when the original commit message lacked one (two-line subject error). - sanitize-history.py: clear the executable bit (mode 100755 -> 100644) on text-file extensions so root-commit "executable permissions but file does not look executable" stops firing.

Validated on Cuberille (InsightSoftwareConsortium#6205): ghostflow now reports only the unavoidable upstream root-commit error. - truncate_subject_to_body() for subjects >78 chars - ensure_blank_line_after_subject() inserts missing blank line - commit_callback clears exec-bit on text-classified blobs - deny-pass strips *.orig / *.rej / *.BACKUP.* / *.LOCAL.* / *.REMOTE.* / *.BASE.*

The four implementation-control macros (DEBUG_PRINT, USE_GRADIENT_RECURSIVE_GAUSSIAN, USE_ADVANCED_PROJECTION, USE_LINESEARCH_PROJECTION) were defined unconditionally at the top of itkCuberilleImageToMeshFilter.h. Any translation unit including the public header silently inherited DEBUG_PRINT=0, masking a same-named macro any consumer might want to test with #ifdef. Move the four #define lines to itkCuberilleImageToMeshFilter.hxx so they remain visible to the template implementation that depends on them via #if, but are no longer emitted into consumer translation units that only include the .h. Greptile P1 on PR InsightSoftwareConsortium#6205.

Both operators applied >= / <= on m_Y in their first branch, which short-circuited to true/false whenever the Y values were equal — before the X tiebreaker could be evaluated. E.g. {Y=1,X=0} >= {Y=1,X=5} resolved (1 >= 1) == true and returned true even though the lhs is strictly less than the rhs. Replace the first-branch comparison with strict > / < so equal-Y cases fall through to the X tiebreaker, mirroring the existing operator> and operator< definitions. std::map calls only operator<, so existing callers do not observe the bug, but the broken operators are class-level API and would silently misbehave under direct use. Greptile P1 on PR InsightSoftwareConsortium#6205.

Convert the 20 Cuberille tests registered via plain add_test() to itk_add_test() so they pick up the standard ITK CTest labels and properties (matching the three already-itk_add_test'd entries below them). Tests with bare add_test are invisible to dashboard-specific test selectors that filter on those labels. GTest conversion of these tests is a separate follow-up. Greptile P2 on PR InsightSoftwareConsortium#6205.

…est-cuberille ENH: Ingest ITKCuberille into Modules/Filtering

Fixes the four classes of ghostflow errors that surfaced on the first Cuberille v4 ingest (PR InsightSoftwareConsortium#6205) — the only allowable remaining error should be the unavoidable upstream root commit. - Strip *.orig / *.rej / *.BACKUP.* / *.LOCAL.* / *.REMOTE.* / *.BASE.* merge-conflict-artifact files from history (catches leftover-conflict-marker errors like the .h.orig on Cuberille). - sanitize-history.py: cap every commit subject at 78 chars (ghostflow / kw-commit-msg rule); excess words move into the body. - sanitize-history.py: insert a blank line between subject and body when the original commit message lacked one (two-line subject error). - sanitize-history.py: clear the executable bit (mode 100755 -> 100644) on text-file extensions so root-commit "executable permissions but file does not look executable" stops firing.

root and others added 30 commits August 22, 2011 09:47

ENH: revision 3

49774c0

BUG: Add missing LICENSE file.

b935d6d

ENH: Add initial itk-module.cmake file.

9adadd4

ENH: Reorganize as an ITKv4 module.

1729732

ENH: Update test CMake configuration for modularization.

522c87a

COMP: QuadEdgeMeshQuadricDecimation renamed to

996bdfb

QuadricDecimationQuadEdgeMeshFilter.

COMP: TimeProbe::GetMeanTime renamed to TimeProbe::GetMean.

c6d8260

BUG: Update copyright headers.

d024e2c

STYLE: Remove unnecessary include's.

21eb1c8

STYLE: Update header names for ITKv4.

69128aa

BUG: Fix CuberilleInDoxygenGroup test.

3d83af3

ENH: Merge pull request #1 from thewtex/module

667b880

Refactor as an ITKv4 module.

COMP: Fix .txx/.hxx #include.

5cdad19

ENH: Merge pull request #2 from thewtex/docker

8a29c33

Add Docker test configuration.

ENH: Add wrapping.

0947553

ENH: Merge pull request #4 from thewtex/wrapping

f34f3d3

ENH: Add wrapping.

COMP: Remove unused typedefs in the test.

f2c43ef

ENH: Merge pull request #6 from thewtex/unused-typedefs

9c6e139

COMP: Remove unused typedefs in the test.

DOC: Fix description typo.

694b32f

ENH: Merge pull request #8 from thewtex/override-virtual

a17cc39

COMP: Address SetInput override virtual warning.

BUG: Address KWStyle tests.

90645dc

STYLE: Merge pull request #10 from thewtex/kwstyletests

c480ff8

BUG: Address KWStyle tests.

BUG: Fix test temporary output dir.

0056f70

The defined directory is not correct when built as a Remote module.

ENH: Merge pull request #11 from thewtex/TestOutputDir

f3a378e

BUG: Fix test temporary output dir.

COMP: Wsign-compare warning

0d1b3a3

STYLE: Null to ITK_NULLPTR

e37827b

ENH: Merge pull request #13 from DVigneault/WsignCompare

055e52c

COMP: Wsign-compare warning

ENH: Merge pull request InsightSoftwareConsortium#14 from DVigneault/…

a687af5

…ITKNULLPTR STYLE: Null to ITK_NULLPTR

BUG: Fixing appearance of Not-A-Number values when gradient is 0

2d177e8

dzenanz reviewed May 4, 2026

View reviewed changes

Comment thread Modules/Filtering/Cuberille/include/itkCuberilleImageToMeshFilter.hxx Outdated

hjmjohnson and others added 5 commits May 4, 2026 17:29

STYLE: Apply gersemi 0.19.3 to Cuberille wrapping test CMakeLists

8df9883

DOC: Add Cuberille module README pointing at archived upstream

30c3dbe

COMP: Remove Cuberille remote-fetch cmake (now in-tree)

e0fbc8e

ENH: Enable Module_Cuberille in configure-ci pixi task

f6a46d2

hjmjohnson changed the title ~~ENH: Ingest ITKCuberille into Modules/Filtering (stacked on #6204)~~ ENH: Ingest ITKCuberille into Modules/Filtering May 4, 2026

hjmjohnson force-pushed the ingest-cuberille branch from c04bb52 to f6a46d2 Compare May 4, 2026 22:34

hjmjohnson added 5 commits May 4, 2026 17:38

hjmjohnson requested a review from dzenanz May 4, 2026 23:01

dzenanz approved these changes May 4, 2026

View reviewed changes

hjmjohnson merged commit 9658520 into InsightSoftwareConsortium:main May 5, 2026
18 of 19 checks passed

hjmjohnson mentioned this pull request May 5, 2026

ENH: Master tracking — Remote-module ingestion, archival, and consolidation #6160

Open

63 tasks

hjmjohnson added this to the ITK 6.0 Release Candidate 1 milestone May 5, 2026

hjmjohnson mentioned this pull request May 6, 2026

COMP: Fix Mac nightly compile errors (VTI test + ImageAlgorithm CTAD) #6219

Merged

hjmjohnson added a commit to hjmjohnson/ITK that referenced this pull request May 6, 2026

Merge pull request InsightSoftwareConsortium#6205 from hjmjohnson/ing…

110bcb4

…est-cuberille ENH: Ingest ITKCuberille into Modules/Filtering

hjmjohnson added a commit to hjmjohnson/ITK that referenced this pull request May 12, 2026

Merge pull request InsightSoftwareConsortium#6205 from hjmjohnson/ing…

779ebf1

…est-cuberille ENH: Ingest ITKCuberille into Modules/Filtering

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

ENH: Ingest ITKCuberille into Modules/Filtering#6205

ENH: Ingest ITKCuberille into Modules/Filtering#6205
hjmjohnson merged 115 commits into
InsightSoftwareConsortium:mainfrom
hjmjohnson:ingest-cuberille

hjmjohnson commented May 4, 2026

Uh oh!

dzenanz left a comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Uh oh!

Conversation

hjmjohnson commented May 4, 2026

Uh oh!

dzenanz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants