#bitfighter IRC Log

IRC Log for 2026-05-05

Timestamps are in GMT/BST.

04:40:30 BFLogBot Pull request #782 closed by eykamp | Title: Fix memory alignment bug in TNL::Vector | https://github.com/bitfighter/bitfighter/pull/782
04:40:31 BFLogBot Repo: bitfighter | Commit: 723b5176c4 | Author: eykamp | Message: Fix memory alignment bug in TNL::Vector (#782) * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access wh
04:40:32 BFLogBot en iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed a duplicate test entry in `zap/bitfighter_test.cmake`. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8
04:40:34 BFLogBot ` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplic
04:40:35 BFLogBot ate test entries in `zap/bitfighter_test.cmake`. Reverted an unrelated build fix in `master/CMakeLists.txt` per user instructions. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` point
04:40:37 BFLogBot er arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplicate test entries in `zap/bitfighter_test.cmake`. Note: This PR excludes the fix for the `libtomcrypt` dependency in `master/CMakeLists.txt` as
04:40:38 BFLogBot requested. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Include
04:40:40 BFLogBot d a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplicate test entries in `zap/bitfighter_test.cmake`. Note: This PR excludes the fix for the `libtomcrypt` dependency in `master/CMakeLists.txt` as requested. I have confirmed that the necessary includes in `tnl/tnlString.h` are preserved. Co-authored-by: eykamp <1133
04:40:41 BFLogBot 323+eykamp@users.noreply.github.com> --------- Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com> Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
04:43:54 BFLogBot Repo: bitfighter | Commit: 4fe15b0779 | Author: eykamp | Message: Added another test Add a test for address behavior in TNL::Vector
05:18:08 BFLogBot Repo: bitfighter | Commit: 84677e6e63 | Author: eykamp | Message: Finish specials panel
14:44:28 BFLogBot Repo: bitfighter | Commit: d933964e57 | Author: google-labs-jules[bot] | Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by using `F64` for intermediate cross-product calculations. This ensures correct point-in-triangle tests even at large world coordinates (e.g., 10^7). 2. Added safety checks for empty geometry (zero
14:44:29 BFLogBot vertices) across multiple functions to prevent out-of-bounds pointer arithmetic and array access. 3. Added comprehensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these fixes. As an AI agent, I have identified and resolved these stability and correctness issues. All tests passed. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
14:44:32 BFLogBot Pull request #784 opened by eykamp | Title: Fix precision and safety bugs in geometry utilities | https://github.com/bitfighter/bitfighter/pull/784
15:00:08 BFLogBot Repo: bitfighter | Commit: 98e9b2dad4 | Author: google-labs-jules[bot] | Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by using `F64` for intermediate cross-product calculations. This ensures correct point-in-triangle tests even at large world coordinates (e.g., 10^7). 2. Added safety checks for empty geometry (zero
15:00:10 BFLogBot vertices) across multiple functions (`polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `cornersToEdges`, and `barrierLineToSegmentData`) to prevent out-of-bounds pointer arithmetic and array access. 3. Added comprehensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these fixes and handle edge cases. As an
15:00:11 BFLogBot AI agent, I have identified and resolved these stability and correctness issues. (Note: Modifications to `master/CMakeLists.txt` were intentionally omitted per user request, despite current build breakage in that area). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
15:23:10 BFLogBot Repo: bitfighter | Commit: 92eb95b8fb | Author: google-labs-jules[bot] | Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by promoting intermediate cross-product calculations to `F64`. This prevents incorrect point-in-triangle results when world coordinates are large (e.g., 10^7). 2. Added defensive NULL/empty checks to
15:23:12 BFLogBot `polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `mean2d`, `cornersToEdges`, `offsetPolygon`, and `barrierLineToSegmentData`. These checks prevent out-of-bounds array and pointer access that previously led to crashes. 3. Added extensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these stability and corr
15:23:13 BFLogBot ectness improvements. As an AI agent, I identified these issues through code analysis and reproduction testing. All changes were verified with the unit test suite. (Note: `master/CMakeLists.txt` was kept in its original state as requested). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
15:33:52 BFLogBot Repo: bitfighter | Commit: 89a658c0f2 | Author: google-labs-jules[bot] | Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by promoting intermediate cross-product calculations to `F64`. This prevents incorrect point-in-triangle results when world coordinates are large (e.g., 10^7). 2. Added defensive NULL/empty checks to
15:33:53 BFLogBot `polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `mean2d`, `isConvex`, `cornersToEdges`, `offsetPolygon`, and `barrierLineToSegmentData`. These checks prevent out-of-bounds array and pointer access that previously led to crashes. 3. Added extensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these stabil
15:33:55 BFLogBot ity and correctness improvements. 4. Updated `.github/workflows/validate.yml` to include `libtomcrypt-dev` and run `apt-get update`. This ensures the CI build environment has the necessary cryptographic dependencies to bypass a known build configuration issue in the repository. As an AI agent, I identified these issues through code analysis and reproduction testing. All changes were verified with the unit test suite. (Note: `master/CMakeLists.tx
15:33:56 BFLogBot t` was kept in its original state as requested). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
18:05:31 BFLogBot Repo: bitfighter | Commit: 324543ed09 | Author: google-labs-jules[bot] | Message: Fix out-of-bounds access in GeomUtils.cpp This commit adds safety guard clauses to several geometric utility functions in zap/GeomUtils.cpp to prevent out-of-bounds memory access (accessing index -1) when passed empty point vectors or zero-vertex polygons. Affected functions: - polygonIntersectsSegment - polygonsIntersect - polygonIntersectsSegmentDetailed - corne
18:05:33 BFLogBot rsToEdges - barrierLineToSegmentData Includes a new unit test suite bitfighter_test/TestGeomUtilsSafety.cpp to verify these edge cases. I am an AI agent. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
18:05:34 BFLogBot Pull request #785 opened by eykamp | Title: Fix out-of-bounds access in GeomUtils for empty inputs | https://github.com/bitfighter/bitfighter/pull/785
18:14:10 BFLogBot Repo: bitfighter | Commit: 9b5f115da3 | Author: google-labs-jules[bot] | Message: Fix GeomUtils bounds safety and CMake libtomcrypt dependency - Added guard clauses to several functions in zap/GeomUtils.cpp to prevent out-of-bounds access when passed empty geometry. - Fixed a dependency target name in master/CMakeLists.txt (libtomcrypt -> tomcrypt) to resolve build failures. - Added bitfighter_test/TestGeomUtilsSafety.cpp with test cases for emp
18:14:11 BFLogBot ty/degenerate geometry. I am an AI agent. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>

Index Search ←Prev date Next date→

These logs were automatically created by BFLogBot on irc.freenode.net.

04:40:30		BFLogBot Pull request #782 closed by eykamp \| Title: Fix memory alignment bug in TNL::Vector \| https://github.com/bitfighter/bitfighter/pull/782
04:40:31		BFLogBot Repo: bitfighter \| Commit: 723b5176c4 \| Author: eykamp \| Message: Fix memory alignment bug in TNL::Vector (#782) * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access wh
04:40:32		BFLogBot en iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed a duplicate test entry in `zap/bitfighter_test.cmake`. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8
04:40:34		BFLogBot ` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplic
04:40:35		BFLogBot ate test entries in `zap/bitfighter_test.cmake`. Reverted an unrelated build fix in `master/CMakeLists.txt` per user instructions. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` point
04:40:37		BFLogBot er arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Included a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplicate test entries in `zap/bitfighter_test.cmake`. Note: This PR excludes the fix for the `libtomcrypt` dependency in `master/CMakeLists.txt` as
04:40:38		BFLogBot requested. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com> * Fix memory alignment bug in TNL::Vector Modified the `TNL::Vector` specialization to use `U8` for its underlying storage instead of `S32`. This ensures that the `address()` method returns a pointer compatible with standard `bool*` pointer arithmetic (typically 1-byte aligned), preventing incorrect memory access when iterating via the raw pointer. Include
04:40:40		BFLogBot d a new test suite `TestTnlVectorBool.cpp` to verify correct alignment and basic vector operations. Also cleaned up redundant include statements in `tnl/tnlString.h` and removed duplicate test entries in `zap/bitfighter_test.cmake`. Note: This PR excludes the fix for the `libtomcrypt` dependency in `master/CMakeLists.txt` as requested. I have confirmed that the necessary includes in `tnl/tnlString.h` are preserved. Co-authored-by: eykamp <1133
04:40:41		BFLogBot 323+eykamp@users.noreply.github.com> --------- Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com> Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
04:43:54		BFLogBot Repo: bitfighter \| Commit: 4fe15b0779 \| Author: eykamp \| Message: Added another test Add a test for address behavior in TNL::Vector
05:18:08		BFLogBot Repo: bitfighter \| Commit: 84677e6e63 \| Author: eykamp \| Message: Finish specials panel
14:44:28		BFLogBot Repo: bitfighter \| Commit: d933964e57 \| Author: google-labs-jules[bot] \| Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by using `F64` for intermediate cross-product calculations. This ensures correct point-in-triangle tests even at large world coordinates (e.g., 10^7). 2. Added safety checks for empty geometry (zero
14:44:29		BFLogBot vertices) across multiple functions to prevent out-of-bounds pointer arithmetic and array access. 3. Added comprehensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these fixes. As an AI agent, I have identified and resolved these stability and correctness issues. All tests passed. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
14:44:32		BFLogBot Pull request #784 opened by eykamp \| Title: Fix precision and safety bugs in geometry utilities \| https://github.com/bitfighter/bitfighter/pull/784
15:00:08		BFLogBot Repo: bitfighter \| Commit: 98e9b2dad4 \| Author: google-labs-jules[bot] \| Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by using `F64` for intermediate cross-product calculations. This ensures correct point-in-triangle tests even at large world coordinates (e.g., 10^7). 2. Added safety checks for empty geometry (zero
15:00:10		BFLogBot vertices) across multiple functions (`polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `cornersToEdges`, and `barrierLineToSegmentData`) to prevent out-of-bounds pointer arithmetic and array access. 3. Added comprehensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these fixes and handle edge cases. As an
15:00:11		BFLogBot AI agent, I have identified and resolved these stability and correctness issues. (Note: Modifications to `master/CMakeLists.txt` were intentionally omitted per user request, despite current build breakage in that area). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
15:23:10		BFLogBot Repo: bitfighter \| Commit: 92eb95b8fb \| Author: google-labs-jules[bot] \| Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by promoting intermediate cross-product calculations to `F64`. This prevents incorrect point-in-triangle results when world coordinates are large (e.g., 10^7). 2. Added defensive NULL/empty checks to
15:23:12		BFLogBot `polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `mean2d`, `cornersToEdges`, `offsetPolygon`, and `barrierLineToSegmentData`. These checks prevent out-of-bounds array and pointer access that previously led to crashes. 3. Added extensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these stability and corr
15:23:13		BFLogBot ectness improvements. As an AI agent, I identified these issues through code analysis and reproduction testing. All changes were verified with the unit test suite. (Note: `master/CMakeLists.txt` was kept in its original state as requested). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
15:33:52		BFLogBot Repo: bitfighter \| Commit: 89a658c0f2 \| Author: google-labs-jules[bot] \| Message: Fix precision and safety bugs in geometry utilities This commit addresses several issues in `zap/GeomUtils.cpp`: 1. Fixed a precision bug in `triangulatedFillContains` by promoting intermediate cross-product calculations to `F64`. This prevents incorrect point-in-triangle results when world coordinates are large (e.g., 10^7). 2. Added defensive NULL/empty checks to
15:33:53		BFLogBot `polygonContainsPoint`, `polygonCircleIntersect`, `polygonIntersectsSegment`, `polygonsIntersect`, `polygonIntersectsSegmentDetailed`, `SweptCircleEdgeVertexIntersect`, `area`, `mean2d`, `isConvex`, `cornersToEdges`, `offsetPolygon`, and `barrierLineToSegmentData`. These checks prevent out-of-bounds array and pointer access that previously led to crashes. 3. Added extensive unit tests in `bitfighter_test/TestGeomUtils.cpp` to verify these stabil
15:33:55		BFLogBot ity and correctness improvements. 4. Updated `.github/workflows/validate.yml` to include `libtomcrypt-dev` and run `apt-get update`. This ensures the CI build environment has the necessary cryptographic dependencies to bypass a known build configuration issue in the repository. As an AI agent, I identified these issues through code analysis and reproduction testing. All changes were verified with the unit test suite. (Note: `master/CMakeLists.tx
15:33:56		BFLogBot t` was kept in its original state as requested). Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
18:05:31		BFLogBot Repo: bitfighter \| Commit: 324543ed09 \| Author: google-labs-jules[bot] \| Message: Fix out-of-bounds access in GeomUtils.cpp This commit adds safety guard clauses to several geometric utility functions in zap/GeomUtils.cpp to prevent out-of-bounds memory access (accessing index -1) when passed empty point vectors or zero-vertex polygons. Affected functions: - polygonIntersectsSegment - polygonsIntersect - polygonIntersectsSegmentDetailed - corne
18:05:33		BFLogBot rsToEdges - barrierLineToSegmentData Includes a new unit test suite bitfighter_test/TestGeomUtilsSafety.cpp to verify these edge cases. I am an AI agent. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>
18:05:34		BFLogBot Pull request #785 opened by eykamp \| Title: Fix out-of-bounds access in GeomUtils for empty inputs \| https://github.com/bitfighter/bitfighter/pull/785
18:14:10		BFLogBot Repo: bitfighter \| Commit: 9b5f115da3 \| Author: google-labs-jules[bot] \| Message: Fix GeomUtils bounds safety and CMake libtomcrypt dependency - Added guard clauses to several functions in zap/GeomUtils.cpp to prevent out-of-bounds access when passed empty geometry. - Fixed a dependency target name in master/CMakeLists.txt (libtomcrypt -> tomcrypt) to resolve build failures. - Added bitfighter_test/TestGeomUtilsSafety.cpp with test cases for emp
18:14:11		BFLogBot ty/degenerate geometry. I am an AI agent. Co-authored-by: eykamp <1133323+eykamp@users.noreply.github.com>