* Add delayed verdict support for response body in nano attachment
Summary:
* Implement TRAFFIC_VERDICT_DELAYED handling in SendResponseBody function * Add delayed verdict thread spawning when response body returns delayed verdict This change mirrors the existing delayed verdict handling in SendRequestBody to ensure consistent behavior for both request and response body processing.
* Redirect kong nano attachment output from stdout to stderr
* fix large response body
* fix large response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix response body
* fix specific path
* return correct path
* fresh start
* fix send_bodu in lua_attachment_wrapper.c
* change branch
* add timeout
* add timeout
* remove the header filter
* try chunk hashing
* proper fini session
* add more check
* try setting last chunk
* remove finilizing session
* return to basic version
* proper fini of session data
* add some check for internal traffic
* envoy functionality...
* proper fini session
* proper fini session
* fini session on accept also
* for testing change the log level
* remove fini session from header
* remove fini session from header
* remove fini session from header
* remove inpect check
* fix wrong close statement
* correct inspection_complete handling
* fix oom
* fix oom
* fix oom
* fix oom
* fix oom
* fix init attachment
* nano_ffi free memoty
* GC implementation
* optimize
* clean
* increase timeout
* increase timeout
* using is_session_fini, the commit before this one is working ...
* returning to last working version ...
* just for testing
* fix block page
* fix block page
* fix block page
* fix block page
* fix block page
* fix fini session
* add collect(stop)
* return to default
* fix oom caused by using freed values
* fix oom caused by using freed values
* add large response body handling
* implement timeout
* implement timeout
* remove GC
* testing
* add end_inspection
* remove comments
* ngx.arg[1] = nil
* Skip inspection for health checks and internal requests
* add new flag for bypass inspections
* move ngx.arg[1] to the beginning
* try without coolectgarbage
* remove internal traffic check
* remove internal traffic check
* I don't know
* I don't know
* try add chunk assignment
* check why data session is nil in header but availablein body
* remove end inspection
* remove logs
* ctx.timeout_passthrough
* remove ctx.timeout_passthrough
* add GC
* remove the logs
* return the logs
* last modification....
* add logs
* revert to working version
* remove stupid cat
* fini_session only in log phase
* last try
* remove processed_requests
* remove unused variable
* remove nano_ffi changes
* add debuging message for testing
* get worker workers inside the init_attachmetn
* test now
* move check
* add accept
* add accept
* remove endinspection
* fix typo
* get req body in chunks
* test body chunks:
* add timeout in req body
* fixing
* add logs for test
* fix accept verdic in body_filter
* add more logs
* fix ngx time
* add more logs
* add getter functions for timeout
* add more logs
* ready for review
* use right content-length
* add missing content-length update
* try content-length nil
* add nil to header filter
* revert shcema changes
* try exit in header filter
* try content-length nil
* remove the test
* refuse connection when prevnet occures in response body
---------
Co-authored-by: wiaamm <wiaamm@checkpoint.com>
* fix large file inspection and free memory
* remove duplicated function
* fix cleanup calls
* return the old functionality
* send the all the body
* fix cleanup
* remove debug messages
---------
Co-authored-by: wiaamm <wiaamm@checkpoint.com>
* add kong plugin to open-appsec
* fix url in rockspec file
* add the attachment prefix to the paths
* check
* fix branch
---------
Co-authored-by: wiaamm <wiaamm@checkpoint.com>
