notes

project.raddbg

@@ -7,12 +7,19 @@ target:
   working_directory: "../raddebugger"
   arguments: "--user:C:/devel/raddebugger/build/raddbg_test.user"
   debug_subprocesses: 0
+}
+target:
+{
+  executable: "build/mule_main.exe"
+  working_directory: build
+  output_label: output_a
   enabled: 1
 }
 target:
 {
   executable: "build/mule_main.exe"
   working_directory: build
+  output_label: output_b
 }
 target:
 {

src/dbg_engine/dbg_engine_ctrl.c

@@ -3579,7 +3579,6 @@ d_ctrl_thread__eval_scope_begin(Arena *arena, D_BreakpointList *user_bps, D_Enti
     E_BaseCtx *ctx = &scope->base_ctx;
     
     //- rjf: fill instruction pointer info
-    ctx->thread_ip_vaddr     = thread_rip_vaddr;
     ctx->thread_ip_voff      = thread_rip_voff;
     ctx->thread_arch         = thread->arch;
     ctx->thread_reg_space = e_space_make(D_EvalSpaceKind_Entity);

src/eval/eval_core.h

@@ -786,7 +786,6 @@ typedef struct E_BaseCtx E_BaseCtx;
 struct E_BaseCtx
 {
   // rjf: instruction pointer info
-  U64 thread_ip_vaddr;
   U64 thread_ip_voff;
   E_Space thread_reg_space;
   E_Space thread_process_space;

src/raddbg/raddbg_core.c

@@ -11955,7 +11955,6 @@ rd_frame(void)
       E_BaseCtx *ctx = eval_base_ctx;
       
       //- rjf: fill instruction pointer info
-      ctx->thread_ip_vaddr     = rip_vaddr;
       ctx->thread_ip_voff      = rip_voff;
       ctx->thread_reg_space    = rd_eval_space_from_ctrl_entity(thread, D_EvalSpaceKind_Entity);
       ctx->thread_process_space= rd_eval_space_from_ctrl_entity(process, D_EvalSpaceKind_Entity);

src/raddbg/raddbg_main.c

@@ -33,7 +33,6 @@
 // [ ] need to eliminate accelerators from evaluation context, and build them on the fly instead -
 //     this is necessary because, for instance, the correct locals_map varies by expression, if we
 //     want to (we do) support features like "look up call stack to find local"
-// [ ] unit / module name qualification
 //
 //- memory view pass
 // [ ] toggleable ascii column
@@ -83,8 +82,6 @@
 //     a way we can defer to the underlying shell in a non-horrible way...?
 //
 //- stepping or breakpoint oddness/fixes
-// [ ] halting during a spoof-ridden step leaves the spoofs in place!!!
-//     (repro via LOTS of code on one line & halting)
 // [ ] stepping-onto a line with a conditional breakpoint, which fails, causes a
 // single step over the first instruction of that line, even if the thread
 // would've stopped at the first instruction due to the step, were that bp not
@@ -185,7 +182,6 @@
 // [ ] investigate /DEBUG:FASTLINK - can we somehow alert that we do not
 //     support it?
 // [ ] just-in-time debugging
-// [ ] step-out-of-loop
 //
 //- late-conversion performance improvements
 // [ ] live++ investigations - ctrl+alt+f11 in UE?
@@ -281,6 +277,10 @@
 // [x] unwinding in jit page - fail
 // [x] test concepts -> base
 // [x] debugger, d2r, p2r exemplar / determinism testing
+// [x] halting during a spoof-ridden step leaves the spoofs in place!!!
+//     (repro via LOTS of code on one line & halting)
+// [x] unit / module name qualification
+// [x] step-out-of-loop
 
 ////////////////////////////////
 //~ rjf: Build Options