[Lldb-commits] [lldb] [lldb-dap] assembly breakpoints (PR #139969)

[Ely Ronnen via lldb-commits]

================
@@ -35,29 +36,34 @@ SourceBreakpoint::SourceBreakpoint(DAP &dap,
       m_line(breakpoint.line),
       m_column(breakpoint.column.value_or(LLDB_INVALID_COLUMN_NUMBER)) {}
 
-void SourceBreakpoint::SetBreakpoint(const protocol::Source &source) {
+llvm::Error SourceBreakpoint::SetBreakpoint(const protocol::Source &source) {
   lldb::SBMutex lock = m_dap.GetAPIMutex();
   std::lock_guard<lldb::SBMutex> guard(lock);
 
   if (m_line == 0)
-    return;
+    return llvm::createStringError(llvm::inconvertibleErrorCode(),
+                                   "Invalid line number.");
 
   if (source.sourceReference) {
     // breakpoint set by assembly source.
     lldb::SBAddress source_address(*source.sourceReference, m_dap.target);
     if (!source_address.IsValid())
-      return;
+      return llvm::createStringError(llvm::inconvertibleErrorCode(),
----------------
eronnen wrote:

I think for this specific case it shouldn't make it to the handler because in the same `setBreakpoints` requests there could be other successful breakpoints set. According to the protocol seems that the `message` field is the way to return errors for specific breakpoints:

```javascript
 /**
   * A message about the state of the breakpoint.
   * This is shown to the user and can be used to explain why a breakpoint could
   * not be verified.
   */
  message?: string;
```

https://github.com/llvm/llvm-project/pull/139969