* pre filter detectors that include the keywords in the chunk.
* Optimize the engine to prevent iterating overing all detectors.
* use sync.Map for concurrent access.
* lint.
* use correct verify.
* allow versioned detectors.
* Break apart Start.
* cleanup.
* Update benchmark.
* add comment.
* remove Engine prefix.
* update comments.
* use regular map.
* delete the pool.
* remove old code.
* refactor ahocorasickcore into own file.
* update comments
* move structs to ahocorasickcore
* update comments
* fix
* address comments
* exported some methods and constructor since it will need to be be used by the enterprise pipeline as well
* remove extra log
* Use heuristic to choose the most likely UTF-16 decoded string
* Assume ASCII and include valid BE and LE bytes
* Remove unused code
* Assume ASCII and return nil when not utf16
---------
Co-authored-by: bill-rich <bill.rich@gmail.com>
* Add Base64URLSafe decoder
* Add decoder that can decode base64 strings with '_' and '-' instead of
of '+' and '/'.
* Combine url-safe b64 decoder into b64 decoder
* Small amount of code clean up.
* Rename sem to concurrency for better readability and to remove an extra comment.
* fix stashing issue.
Co-authored-by: Ahrav Dutta <ahrav.dutta@trufflesec.com>
