SLeeLa Debugger Architecture
Version: 0.6.0
System model
Interface
|-- Terminal
|-- Voice
|-- IDE / DAP
|-- Test Suite
|-- Plugins
|
v
Typed Action Controller
|
v
Debugger Session
|
+----+----+
| |
Event Bus Evidence Chain
| |
+----+----+
|
v
Native Backend
|-- Linux ptrace
|-- macOS LLDB
|-- Windows Debug API
|-- Portable
|
v
Target Process
Evidence model
Every significant observation should be correlatable through:
Observation -> Event -> Source -> Thread -> Stack -> Registers/Memory -> Action -> Backend Result -> Diagnostic.
Observed facts and derived interpretations must remain distinguishable.
Stop model
A stop record should identify:
- reason
- process/thread
- source location
- function
- breakpoint/watchpoint
- exception/signal
- triggering action
- previous relevant events
- available evidence
Identity model
A debug session should track:
- executable hash
- source identity
- symbol identity
- compiler/toolchain
- linker
- build ID
- architecture
- OS
- debugger version
Domain integration
The architecture permits SLeeLa-specific observers for HTTP/server traffic, tests, sanitizers, logs, and other diagnostic domains without changing native backend semantics.
Security boundary
Debugger commands are typed actions. They are never interpreted as arbitrary shell input or target-program input.