Skip to content

Misc Fixes for 2026/09/12 - #1824

Open
Areloch wants to merge 1 commit into
TorqueGameEngines:developmentfrom
Areloch:MiscFixes_20260912
Open

Areloch wants to merge 1 commit into
TorqueGameEngines:developmentfrom
Areloch:MiscFixes_20260912

Conversation

@Areloch

@Areloch Areloch commented Sep 13, 2026

Copy link
Copy Markdown
Contributor
  • Adjusted the handling for testing loading cached dts' and enumerating shape files in assetImport to more correctly and reliably populate shape contents for import
  • Tweak to FrameTemp constructor so if FrameAllocator::alloc were to ever return NULL, have the size be appropriately 0 for better checks handling, rather than the assumed numElements size that failed
  • Error check and handling in Stream::_read() so if the alloc fails, we can handle it and error log properly
  • Tweaked logic in the tokenizer so we gate the quote token while loop as part of the loop argument, rather than internal to it to make it a bit cleaner and forces the assert to happen as an after-catch
  • Have it so both GuiInspectorTypeRangedS32 and GuiInspectorTypeRangedF32 handle both altCommand and validate properly. As in, hitting enter validates the value as normal, but clicking off of the control and it losing focus now also validates the new value, so clicking away doesn't "lose" the change
  • Added a sanity check to the renderProbeMgr's submitProbe
  • Added a memcpy to the ambientChannels processing to minimize the chance of the memory being dropped once out of immediate scope
  • When doing TSSkinMesh::createSkinBatchData(), tweaked the loop range to more correctly cap to BatchData::maxBonePerVert, rather than a potential overrun if batchOperations.last().transformCount ends up being bigger than it

…ng shape files in assetImport to more correctly and reliably populate shape contents for import

- Tweak to FrameTemp constructor so if FrameAllocator::alloc were to ever return NULL, have the size be appropriately 0 for better checks handling, rather than the assumed numElements size that failed
- Error check and handling in Stream::_read() so if the alloc fails, we can handle it and error log properly
- Tweaked logic in the tokenizer so we gate the quote token while loop as part of the loop argument, rather than internal to it to make it a bit cleaner and forces the assert to happen as an after-catch
- Have it so both GuiInspectorTypeRangedS32 and GuiInspectorTypeRangedF32 handle both altCommand and validate properly. As in, hitting enter validates the value as normal, but clicking off of the control and it losing focus now also validates the new value, so clicking away doesn't "lose" the change
- Added a sanity check to the renderProbeMgr's submitProbe
- Added a memcpy to the ambientChannels processing to minimize the chance of the memory being dropped once out of immediate scope
- When doing TSSkinMesh::createSkinBatchData(), tweaked the loop range to more correctly cap to BatchData::maxBonePerVert, rather than a potential overrun if batchOperations.last().transformCount ends up being bigger than it
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant