use synchronous reads for call stack computation, but extend artifact cache with the ability to say that a non-retryable failure occurred, so that we can fall back to stale call stacks

project.raddbg

@@ -9,13 +9,6 @@ target:
   debug_subprocesses: 0
 }
 target:
-{
-  executable: "build/raddbg.exe"
-  working_directory: "../raddebugger"
-  arguments: "--user:test123"
-  debug_subprocesses: 0
-}
-target:
 {
   executable: "build/mule_main.exe"
   working_directory: build
@@ -25,7 +18,6 @@ target:
 {
   executable: "build/torture.exe"
   working_directory: build
-  arguments: "raddbg/*"
 }
 target:
 {
@@ -40,20 +32,13 @@ target:
 }
 target:
 {
-  executable: "build/release_version/raddbg.exe"
-  working_directory: "build/release_version"
-}
-target:
-{
-  executable: "build/radbin.exe"
+  executable: "build/ryan_scratch.exe"
   working_directory: build
-  arguments: "--breakpad fat.so --capture"
-}
-target:
-{
-  executable: "build/radbin.exe"
-  working_directory: build
-  arguments: "--rdi raddbg"
   enabled: 1
 }
-watch_pin: expression: tag_hash_slots
+target:
+{
+  executable: "build/raddbg.exe"
+  working_directory: "build/"
+  arguments: "--bin --dump test123.rdi"
+}