BUG: C21: Wrong or misleading diagnostics in 10 filters

オープン
#1,880 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 6 日以内に返信

まだ誰も着手していません。

評価

難易度
5/5
見積もり時間
1週間以上
初心者へのやさしさ
20/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
cpp
領域
backend

調査の方向性

Start by reading the ten referenced filter implementations and their associated tests; the checklist links identify each finding and location. Reproduce each finding with a failing test, then run the relevant tests for both Direct and Scanline paths where masks are involved. Done means all ten diagnostics or mask behaviors match the stated rules, tests pass, and the AlignSectionsMutualInformation mask behavior decision is resolved.

索引モデルが issue の本文から書いたものです。

説明

bug

Summary

Some filters give diagnostics that are wrong. Two sub-patterns occur:

  • Masked cells validated before the mask test. A per-cell loop validates a value (for example a Phase) or grows a feature before it tests the mask. A cell that the mask excludes can then fail execute or join a feature.
  • Diagnostic message wrong, swapped or misleading. An error, warning or help text formats the wrong variable, swaps arguments or messages, fires on the wrong condition, adds counts across arrays, or gives advice that the user cannot follow.

As of develop 419f9db35 (2026-10-21), there are 10 open findings in 10 filters: 0 CRASH, 0 WRONG_DATA, 10 MINOR. All findings are candidates. They were found and re-checked by reading the code. Nobody ran the code to confirm them.

Mechanism

Most messages are written by hand at each call site. No test asserts on the message text, so a wrong argument or a swapped message is not detected. Example: ReadEnsembleInfo gives the group name where the key name must go, and the key name where the group name must go:

https://github.com/BlueQuartzSoftware/simplnx/blob/419f9db357d0c5e9adbdb0ab4511d36a433aecfb/src/Plugins/OrientationAnalysis/src/OrientationAnalysis/Filters/Algorithms/ReadEnsembleInfo.cpp#L103-L110

    if(inputIni[group].count(k_CrystalStructure) == 0)
    {
      return MakeErrorResult(-13005, fmt::format("Could not find the key {} in phase group {} from the input file '{}'", group, k_CrystalStructure, m_InputValues->InputFile));
    }
    if(inputIni[group].count(k_PhaseType) == 0)
    {
      return MakeErrorResult(-13006, fmt::format("Could not find the key {} in phase group {} from the input file '{}'", group, k_PhaseType, m_InputValues->InputFile));
    }

For the mask sub-pattern, the loop counts a bad value before it reads the mask. In ComputeFZQuaternions the count of out-of-range phases includes cells that the mask excludes. Execute then fails with -49008, although those cells produce no output (ComputeFZQuaternions.cpp#L181-L186).

No filter of this class was fixed on develop or in an open pull request.

Rule

  • In each per-cell loop, test the mask first. Validate, count or grow only the cells that the mask enables. Apply this in every algorithm path (Direct and Scanline).
  • Each diagnostic must name the correct values, in the correct order, and must fire only on the condition that its text describes. Advice in a message must be possible to follow.
  • Each distinct condition gets its own error code and message.
  • Decision needed for AlignSectionsMutualInformationFilter#3. DREAM3D 6.5.171 also lets masked cells join features during neighbor growth, but the filter documentation says that masked cells are excluded. Options:
    1. Add the mask test to neighbor growth (behavior changes from 6.5.171).
    2. Keep the 6.5.171 behavior and correct the documentation (masked cells are only excluded as seeds).

How to fix

  1. Write the failing test first (see "How to test").
  2. For a mask finding, move the mask test before the validation or the counting step. Do this in each algorithm path that has the loop.
  3. For a message finding, correct the fmt::format arguments, the condition, or the text. Keep the existing error code when the condition stays the same.
  4. When you add a new error code, use the range of that filter. Filter error codes are more negative than -10000, and each filter uses one range of 100 codes (first code -47500 gives -47500..-47599). Check that the code is unique with git grep -nE -- "-NNNN([^0-9]|$)". Do not use \b, because git grep -E ignores it. Also check the open pull request branches.
  5. For a help-text-only change (ExtractVertexGeometryFilter#1), a unit test cannot fail on develop. Write this in the pull request.

How to test

  • Write a unit test that fails on develop before you change the code.
  • Message findings: write an error-path test. Assert the error code (or warning code). Also assert that the message contains the correct values, for example the key name and the group name in the correct positions.
  • Mask findings: write an execute test with a masked-out cell that holds a bad value (for example Phase 5 with 2 Crystal Structures tuples). The test must pass, and the output values of the enabled cells must match an independent expected result. Run Direct and Scanline both through the existing algorithm-selection test scope.
  • Warning findings: assert that the warning appears for the condition that it describes and does not appear for the other condition.
  • If you cannot reproduce a finding, write this with the reason in the pull request and on this issue. Do not force a fix.

Conventions

  • Make one pull request for the whole class, with one commit for each filter.
  • Start each commit message with BUG: and add a Signed-off-by line. Do not add AI attribution.
  • Start each fix with a unit test that fails on develop.
  • Value checks need an independent expected result (oracle), not the output of the code under test.
  • Run ctest serially. Never use -j, because the tests depend on each other.
  • Format with clang-format 19.1.1 and the repository .clang-format.
  • Do not add sentences about input requirements to the filter documentation. The generated parameter table documents them.
  • If a fix touches a filter that has a V&V report (src/Plugins/<Plugin>/vv/<Filter>.md), update every file of its V&V package (report, deviations, provenance, test tables). V&V documents reference DREAM3D 6.5.171 only.
  • Rebase on develop. Do not merge develop into the branch.

Checklist

  • MINOR · ReadImageStackFilter · ReadImageStackFilter.cpp:390 — RGB uint8 slices with Set Image Data Type = true and Output Data Type = uint16: error -23506 tells the user to turn on Convert To GrayScale, but then error -23504 rejects the non-uint8 data, so the advice cannot be followed. ReadImageStackFilter#9
  • MINOR · AlignSectionsMutualInformationFilter · AlignSectionsMutualInformation.cpp:125 — With Use Mask Array = true, neighbor growth tests only Feature Ids and Phases, so a masked-out cell joins a feature, or fails with -53703 if its Phase is out of range. Matches 6.5.171; see "Decision needed". AlignSectionsMutualInformationFilter#3
  • MINOR · ComputeBoundaryStrengthsFilter · ComputeBoundaryStrengths.cpp:88 — The Laue class warning fires when the phase index is not 1, not when the crystal structure is not Cubic m-3m. It fires for a cubic phase 2 and is missing for a hexagonal phase 1. Output values are correct. ComputeBoundaryStrengthsFilter#3 — moved to #1861 C01b.
  • MINOR · ComputeFZQuaternionsFilter · ComputeFZQuaternions.cpp:183 — With Apply to Good Elements Only = true, a masked-out cell with Phase 5 and 2 Crystal Structures tuples is counted as out of range, so execute fails with -49008. Fix every algorithm path. ComputeFZQuaternionsFilter#1
  • MINOR · CreateEnsembleInfoFilter · CreateEnsembleInfo.cpp:82 — Errors -10006 and -10007 print the looked-up number (for example 999) instead of the crystal structure or phase type text that the user typed, and do not give the row number. CreateEnsembleInfoFilter#2
  • MINOR · ReadEnsembleInfoFilter · ReadEnsembleInfo.cpp:105 — An INI group without a CrystalStructure or PhaseType key gives error -13005 or -13006 with the key name and group name swapped. ReadEnsembleInfoFilter#3
  • MINOR · ComputeArrayHistogramByFeatureFilter · ComputeArrayHistogramByFeature.cpp:1393 — The overflow counter is declared once outside the per-array loop and never reset, so the message for array B also counts the out-of-range values of array A. ComputeArrayHistogramByFeatureFilter#10
  • MINOR · ComputeFeatureRectFilter · ComputeFeatureRect.cpp:93 — A Feature Id of -1 gives error -31000, which says that the Feature Attribute Matrix has too few tuples. Negative Feature Ids need their own code and message. ComputeFeatureRectFilter#2
  • MINOR · CreateColorMapFilter · ColorTableUtilities.cpp:251 — Errors -36782 and -36783 are swapped: an unknown preset says "Found ... but no 'RGBPoints'", and a preset without RGBPoints (for example 'KAAMS') says "Unable to find". CreateColorMapFilter#3
  • MINOR · ExtractVertexGeometryFilter · ExtractVertexGeometryFilter.cpp:71 — The Array Handling help text says "[0] Move or [1] Copy", but the choices and ArrayHandlingType use 0 = Copy and 1 = Move. A Python or pipeline user who sets 0 gets Copy. ExtractVertexGeometryFilter#1

Out of scope

  • Findings of this class in filters that have a V&V report are not in this issue. There is 1 such finding, in ComputeIPFColorsFilter. It is handled on a separate V&V branch.
主要言語
C++
スター
18
フォーク
15
平均マージ
4日 23時間
マージ済み PR(30日)
42

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

BlueQuartzSoftware/simplnx のほかの issue

BlueQuartzSoftware/simplnx の issue をすべて見る

似ている issue

C++ の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。