[lldb] Add interface to check if UserExpression::Parse() is cacheable (…

…#66826)

When setting conditional breakpoints, we currently assume that a call to
UserExpression::Parse() can be cached and resued multiple times. This
may not be true for every user expression. Add a new method so
subclasses of UserExpression can customize if they are parseable or not.

lldb/include/lldb/Expression/UserExpression.h

@@ -192,6 +192,14 @@ class UserExpression : public Expression {
   /// expression.  Text() should contain the definition of this function.
   const char *FunctionName() override { return "$__lldb_expr"; }
 
+  /// Returns whether the call to Parse on this user expression is cacheable.
+  /// This function exists to provide an escape hatch for supporting languages
+  /// where parsing an expression in the exact same context is unsafe. For
+  /// example, languages where generic functions aren't monomorphized, but
+  /// implement some other mechanism to represent generic values, may be unsafe
+  /// to cache, as the concrete type substitution may be different in every
+  /// expression evaluation.
+  virtual bool IsParseCacheable() { return true; }
   /// Return the language that should be used when parsing.  To use the
   /// default, return eLanguageTypeUnknown.
   lldb::LanguageType Language() const override { return m_language; }

lldb/source/Breakpoint/BreakpointLocation.cpp

@@ -250,6 +250,7 @@ bool BreakpointLocation::ConditionSaysStop(ExecutionContext &exe_ctx,
   DiagnosticManager diagnostics;
 
   if (condition_hash != m_condition_hash || !m_user_expression_sp ||
+      !m_user_expression_sp->IsParseCacheable() ||
       !m_user_expression_sp->MatchesContext(exe_ctx)) {
     LanguageType language = eLanguageTypeUnknown;
     // See if we can figure out the language from the frame, otherwise use the