Implement word boundary checks in regular expressions #14129
Conversation
In the current implementation of regular expressions intended to match the GLSL keywords "in", "out", and "varying", it has been found that the patterns may match parts of variable names or within comments that contain these keywords, leading to inaccurate results. To improve accuracy, word boundaries (\b) have been applied around each keyword in the regular expressions. This ensures that the engine would only match complete words, not substrings within longer words or comments. The regular expressions have been updated as follows: const outRegex = /(flat\s)?\s*\bout\b /; const inRegex = /(flat\s)?\s*\bin\b /; const varyingRegex = /(flat\s)?\s*\bvarying\b/ ; While this improves the keyword matching accuracy, please note that particularly advanced cases like keyword usage inside comments might still not be handled perfectly. For deeper parsing needs, consider using a dedicated GLSL parser or compiler plugin.
|
Please make sure to label your PR with "bug", "new feature" or "breaking change" label(s). |
|
Snapshot stored with reference name: Test environment: To test a playground add it to the URL, for example: https://babylonsnapshots.z22.web.core.windows.net/refs/pull/14129/merge/index.html#WGZLGJ#4600 Links to test babylon tools with this snapshot: https://playground.babylonjs.com/?snapshot=refs/pull/14129/merge To test the snapshot in the playground with a playground ID add it after the snapshot query string: https://playground.babylonjs.com/?snapshot=refs/pull/14129/merge#BCU1XR#0 |
packages/dev/core/src/Engines/WebGPU/webgpuShaderProcessorsGLSL.ts
Outdated
Show resolved
Hide resolved
remove the trailing spaces
|
Thanks! |
See WebGPU use glsl bug ： non-opaque uniforms outside a block - Bugs - Babylon.js
In the current implementation of regular expressions intended to match the GLSL keywords "in", "out", and "varying", it has been found that the patterns may match parts of variable names or within comments that contain these keywords, leading to inaccurate results.
Example
To improve accuracy, word boundaries (\b) have been applied around each keyword in the regular expressions. This ensures that the engine would only match complete words, not substrings within longer words or comments.
The regular expressions have been updated as follows: