[lldb][NFC] Fix a build failure with MSVC

[igorkudrin]

This fixes a build error with msvc for code introduced in #106442:

...\lldb\source\Expression\DiagnosticManager.cpp(37): error C2668: 'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo': ambiguous call to overloaded function
...\llvm\include\llvm/Support/Error.h(357): note: could be 'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo(std::error_code)', which inherits 'lldb_private::CloneableECError::CloneableECError(std::error_code)' via base class 'lldb_private::ExpressionErrorBase'
...\llvm\include\llvm/Support/Error.h(357): note: or       'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo(std::error_code,std::string)', which inherits 'lldb_private::ExpressionErrorBase::ExpressionErrorBase(std::error_code,std::string)' via base class 'lldb_private::ExpressionErrorBase'
...\lldb\source\Expression\DiagnosticManager.cpp(37): note: while trying to match the argument list '(std::error_code)'

[llvmbot]

@llvm/pr-subscribers-lldb

Author: Igor Kudrin (igorkudrin)

Changes

This fixes a build error with msvc for code introduced in #106442:

...\lldb\source\Expression\DiagnosticManager.cpp(37): error C2668: 'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo': ambiguous call to overloaded function
...\llvm\include\llvm/Support/Error.h(357): note: could be 'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo(std::error_code)', which inherits 'lldb_private::CloneableECError::CloneableECError(std::error_code)' via base class 'lldb_private::ExpressionErrorBase'
...\llvm\include\llvm/Support/Error.h(357): note: or       'llvm::ErrorInfo<lldb_private::ExpressionError,lldb_private::ExpressionErrorBase>::ErrorInfo(std::error_code,std::string)', which inherits 'lldb_private::ExpressionErrorBase::ExpressionErrorBase(std::error_code,std::string)' via base class 'lldb_private::ExpressionErrorBase'
...\lldb\source\Expression\DiagnosticManager.cpp(37): note: while trying to match the argument list '(std::error_code)'

---

Full diff: https://github.com/llvm/llvm-project/pull/111231.diff

1 Files Affected:

• (modified) lldb/include/lldb/Utility/Status.h (+1-2)

diff --git a/lldb/include/lldb/Utility/Status.h b/lldb/include/lldb/Utility/Status.h
index 3910c26d115a09..2911ffb804c6fb 100644
--- a/lldb/include/lldb/Utility/Status.h
+++ b/lldb/include/lldb/Utility/Status.h
@@ -83,8 +83,7 @@ class ExpressionErrorBase
     : public llvm::ErrorInfo<ExpressionErrorBase, CloneableECError> {
 public:
   using llvm::ErrorInfo<ExpressionErrorBase, CloneableECError>::ErrorInfo;
-  ExpressionErrorBase(std::error_code ec, std::string msg = {})
-      : ErrorInfo(ec) {}
+  ExpressionErrorBase(std::error_code ec) : ErrorInfo(ec) {}
   lldb::ErrorType GetErrorType() const override;
   static char ID;
 };

[labath]

Looks good, I guess. Is that constructor even necessary, given that it just forwards the argument to the base class, and we already have the forwarding using declaration?

[igorkudrin]

Looks good, I guess. Is that constructor even necessary, given that it just forwards the argument to the base class, and we already have the forwarding using declaration?

The constructor looks redundant at the moment, perhaps it is just a legacy from the early stages of patch #106442. Maybe some other constructors can also be reduced, for example, Win32Error::Win32Error(). However, I may not fully understand the intentions of the author of the original patch, so I'd prefer @adrian-prantl to do the cleanup if necessary. My patch is only intended to fix the immediate problem.

[adrian-prantl]

Looks good, I guess. Is that constructor even necessary, given that it just forwards the argument to the base class, and we already have the forwarding using declaration?

The constructor looks redundant at the moment, perhaps it is just a legacy from the early stages of patch #106442. Maybe some other constructors can also be reduced, for example, Win32Error::Win32Error(). However, I may not fully understand the intentions of the author of the original patch, so I'd prefer @adrian-prantl to do the cleanup if necessary. My patch is only intended to fix the immediate problem.

This constructor is more a leftover after a bunch of refactoring, your change LGTM!