hanchenye-llvm-project/lldb/source/Core/ValueObjectDynamicValue.cpp

433 lines
14 KiB
C++
Raw Normal View History

//===-- ValueObjectDynamicValue.cpp ---------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "lldb/Core/ValueObjectDynamicValue.h"
// C Includes
// C++ Includes
// Other libraries and framework includes
// Project includes
#include "lldb/Core/Log.h"
#include "lldb/Core/Module.h"
#include "lldb/Core/ValueObjectList.h"
#include "lldb/Core/Value.h"
#include "lldb/Core/ValueObject.h"
#include "lldb/Symbol/ClangASTType.h"
#include "lldb/Symbol/ObjectFile.h"
#include "lldb/Symbol/SymbolContext.h"
#include "lldb/Symbol/Type.h"
#include "lldb/Symbol/Variable.h"
#include "lldb/Target/ExecutionContext.h"
#include "lldb/Target/LanguageRuntime.h"
#include "lldb/Target/Process.h"
#include "lldb/Target/RegisterContext.h"
#include "lldb/Target/Target.h"
#include "lldb/Target/Thread.h"
using namespace lldb_private;
ValueObjectDynamicValue::ValueObjectDynamicValue (ValueObject &parent, lldb::DynamicValueType use_dynamic) :
ValueObject(parent),
m_address (),
m_dynamic_type_info(),
m_use_dynamic (use_dynamic)
{
SetName (parent.GetName());
}
ValueObjectDynamicValue::~ValueObjectDynamicValue()
{
m_owning_valobj_sp.reset();
}
ClangASTType
ValueObjectDynamicValue::GetClangTypeImpl ()
{
const bool success = UpdateValueIfNeeded(false);
if (success)
{
if (m_dynamic_type_info.HasType())
return m_value.GetClangType();
else
return m_parent->GetClangType();
}
return m_parent->GetClangType();
}
ConstString
ValueObjectDynamicValue::GetTypeName()
{
const bool success = UpdateValueIfNeeded(false);
if (success)
{
if (m_dynamic_type_info.HasName())
return m_dynamic_type_info.GetName();
}
return m_parent->GetTypeName();
}
TypeImpl
ValueObjectDynamicValue::GetTypeImpl ()
{
const bool success = UpdateValueIfNeeded(false);
if (success && m_type_impl.IsValid())
{
return m_type_impl;
}
return m_parent->GetTypeImpl();
}
ConstString
ValueObjectDynamicValue::GetQualifiedTypeName()
{
const bool success = UpdateValueIfNeeded(false);
if (success)
{
if (m_dynamic_type_info.HasName())
return m_dynamic_type_info.GetName();
Introduce the concept of a "display name" for types Rationale: Pretty simply, the idea is that sometimes type names are way too long and contain way too many details for the average developer to care about. For instance, a plain ol' vector of int might be shown as std::__1::vector<int, std::__1::allocator<.... rather than the much simpler std::vector<int> form, which is what most developers would actually type in their code Proposed solution: Introduce a notion of "display name" and a corresponding API GetDisplayTypeName() to return such a crafted for visual representation type name Obviously, the display name and the fully qualified (or "true") name are not necessarily the same - that's the whole point LLDB could choose to pick the "display name" as its one true notion of a type name, and if somebody really needs the fully qualified version of it, let them deal with the problem Or, LLDB could rename what it currently calls the "type name" to be the "display name", and add new APIs for the fully qualified name, making the display name the default choice The choice that I am making here is that the type name will keep meaning the same, and people who want a type name suited for display will explicitly ask for one It is the less risky/disruptive choice - and it should eventually make it fairly obvious when someone is asking for the wrong type Caveats: - for now, GetDisplayTypeName() == GetTypeName(), there is no logic to produce customized display type names yet. - while the fully-qualified type name is still the main key to the kingdom of data formatters, if we start showing custom names to people, those should match formatters llvm-svn: 209072
2014-05-18 03:14:17 +08:00
}
return m_parent->GetQualifiedTypeName();
}
ConstString
ValueObjectDynamicValue::GetDisplayTypeName()
{
const bool success = UpdateValueIfNeeded(false);
if (success)
{
if (m_dynamic_type_info.HasType())
Introduce the concept of a "display name" for types Rationale: Pretty simply, the idea is that sometimes type names are way too long and contain way too many details for the average developer to care about. For instance, a plain ol' vector of int might be shown as std::__1::vector<int, std::__1::allocator<.... rather than the much simpler std::vector<int> form, which is what most developers would actually type in their code Proposed solution: Introduce a notion of "display name" and a corresponding API GetDisplayTypeName() to return such a crafted for visual representation type name Obviously, the display name and the fully qualified (or "true") name are not necessarily the same - that's the whole point LLDB could choose to pick the "display name" as its one true notion of a type name, and if somebody really needs the fully qualified version of it, let them deal with the problem Or, LLDB could rename what it currently calls the "type name" to be the "display name", and add new APIs for the fully qualified name, making the display name the default choice The choice that I am making here is that the type name will keep meaning the same, and people who want a type name suited for display will explicitly ask for one It is the less risky/disruptive choice - and it should eventually make it fairly obvious when someone is asking for the wrong type Caveats: - for now, GetDisplayTypeName() == GetTypeName(), there is no logic to produce customized display type names yet. - while the fully-qualified type name is still the main key to the kingdom of data formatters, if we start showing custom names to people, those should match formatters llvm-svn: 209072
2014-05-18 03:14:17 +08:00
return GetClangType().GetDisplayTypeName();
if (m_dynamic_type_info.HasName())
return m_dynamic_type_info.GetName();
}
Introduce the concept of a "display name" for types Rationale: Pretty simply, the idea is that sometimes type names are way too long and contain way too many details for the average developer to care about. For instance, a plain ol' vector of int might be shown as std::__1::vector<int, std::__1::allocator<.... rather than the much simpler std::vector<int> form, which is what most developers would actually type in their code Proposed solution: Introduce a notion of "display name" and a corresponding API GetDisplayTypeName() to return such a crafted for visual representation type name Obviously, the display name and the fully qualified (or "true") name are not necessarily the same - that's the whole point LLDB could choose to pick the "display name" as its one true notion of a type name, and if somebody really needs the fully qualified version of it, let them deal with the problem Or, LLDB could rename what it currently calls the "type name" to be the "display name", and add new APIs for the fully qualified name, making the display name the default choice The choice that I am making here is that the type name will keep meaning the same, and people who want a type name suited for display will explicitly ask for one It is the less risky/disruptive choice - and it should eventually make it fairly obvious when someone is asking for the wrong type Caveats: - for now, GetDisplayTypeName() == GetTypeName(), there is no logic to produce customized display type names yet. - while the fully-qualified type name is still the main key to the kingdom of data formatters, if we start showing custom names to people, those should match formatters llvm-svn: 209072
2014-05-18 03:14:17 +08:00
return m_parent->GetDisplayTypeName();
}
size_t
ValueObjectDynamicValue::CalculateNumChildren()
{
const bool success = UpdateValueIfNeeded(false);
if (success && m_dynamic_type_info.HasType())
return GetClangType().GetNumChildren (true);
else
return m_parent->GetNumChildren();
}
uint64_t
ValueObjectDynamicValue::GetByteSize()
{
const bool success = UpdateValueIfNeeded(false);
if (success && m_dynamic_type_info.HasType())
return m_value.GetValueByteSize(NULL);
else
return m_parent->GetByteSize();
}
lldb::ValueType
ValueObjectDynamicValue::GetValueType() const
{
return m_parent->GetValueType();
}
static TypeAndOrName
FixupTypeAndOrName (const TypeAndOrName& type_andor_name,
ValueObject& parent)
{
TypeAndOrName ret(type_andor_name);
if (type_andor_name.HasType())
{
// The type will always be the type of the dynamic object. If our parent's type was a pointer,
// then our type should be a pointer to the type of the dynamic object. If a reference, then the original type
// should be okay...
ClangASTType orig_type = type_andor_name.GetClangASTType();
ClangASTType corrected_type = orig_type;
if (parent.IsPointerType())
corrected_type = orig_type.GetPointerType ();
else if (parent.IsPointerOrReferenceType())
corrected_type = orig_type.GetLValueReferenceType ();
ret.SetClangASTType(corrected_type);
}
else /*if (m_dynamic_type_info.HasName())*/
{
// If we are here we need to adjust our dynamic type name to include the correct & or * symbol
std::string corrected_name (type_andor_name.GetName().GetCString());
if (parent.IsPointerType())
corrected_name.append(" *");
else if (parent.IsPointerOrReferenceType())
corrected_name.append(" &");
// the parent type should be a correctly pointer'ed or referenc'ed type
ret.SetClangASTType(parent.GetClangType());
ret.SetName(corrected_name.c_str());
}
return ret;
}
bool
ValueObjectDynamicValue::UpdateValue ()
{
SetValueIsValid (false);
m_error.Clear();
if (!m_parent->UpdateValueIfNeeded(false))
{
2011-05-30 08:49:24 +08:00
// The dynamic value failed to get an error, pass the error along
if (m_error.Success() && m_parent->GetError().Fail())
m_error = m_parent->GetError();
return false;
}
// Setting our type_sp to NULL will route everything back through our
// parent which is equivalent to not using dynamic values.
if (m_use_dynamic == lldb::eNoDynamicValues)
{
m_dynamic_type_info.Clear();
return true;
}
ExecutionContext exe_ctx (GetExecutionContextRef());
Target *target = exe_ctx.GetTargetPtr();
if (target)
{
m_data.SetByteOrder(target->GetArchitecture().GetByteOrder());
m_data.SetAddressByteSize(target->GetArchitecture().GetAddressByteSize());
}
// First make sure our Type and/or Address haven't changed:
Process *process = exe_ctx.GetProcessPtr();
if (!process)
return false;
TypeAndOrName class_type_or_name;
Address dynamic_address;
bool found_dynamic_type = false;
lldb::LanguageType known_type = m_parent->GetObjectRuntimeLanguage();
if (known_type != lldb::eLanguageTypeUnknown && known_type != lldb::eLanguageTypeC)
{
LanguageRuntime *runtime = process->GetLanguageRuntime (known_type);
if (runtime)
found_dynamic_type = runtime->GetDynamicTypeAndAddress (*m_parent, m_use_dynamic, class_type_or_name, dynamic_address);
}
else
{
LanguageRuntime *cpp_runtime = process->GetLanguageRuntime (lldb::eLanguageTypeC_plus_plus);
if (cpp_runtime)
found_dynamic_type = cpp_runtime->GetDynamicTypeAndAddress (*m_parent, m_use_dynamic, class_type_or_name, dynamic_address);
if (!found_dynamic_type)
{
LanguageRuntime *objc_runtime = process->GetLanguageRuntime (lldb::eLanguageTypeObjC);
if (objc_runtime)
found_dynamic_type = objc_runtime->GetDynamicTypeAndAddress (*m_parent, m_use_dynamic, class_type_or_name, dynamic_address);
}
}
// Getting the dynamic value may have run the program a bit, and so marked us as needing updating, but we really
// don't...
m_update_point.SetUpdated();
if (found_dynamic_type)
{
if (class_type_or_name.HasType())
{
// TypeSP are always generated from debug info
if (!class_type_or_name.HasTypeSP() && class_type_or_name.GetClangASTType().IsRuntimeGeneratedType())
{
m_type_impl = TypeImpl(m_parent->GetClangType(),FixupTypeAndOrName(class_type_or_name, *m_parent).GetClangASTType());
class_type_or_name.SetClangASTType(ClangASTType());
}
else
{
m_type_impl = TypeImpl(FixupTypeAndOrName(class_type_or_name, *m_parent).GetClangASTType());
}
}
else
{
m_type_impl.Clear();
}
}
else
{
m_type_impl.Clear();
}
// If we don't have a dynamic type, then make ourselves just a echo of our parent.
// Or we could return false, and make ourselves an echo of our parent?
if (!found_dynamic_type)
{
if (m_dynamic_type_info)
SetValueDidChange(true);
ClearDynamicTypeInformation();
m_dynamic_type_info.Clear();
m_value = m_parent->GetValue();
m_error = m_value.GetValueAsData (&exe_ctx, m_data, 0, GetModule().get());
return m_error.Success();
}
Value old_value(m_value);
Log *log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_TYPES));
bool has_changed_type = false;
if (!m_dynamic_type_info)
{
m_dynamic_type_info = class_type_or_name;
has_changed_type = true;
}
else if (class_type_or_name != m_dynamic_type_info)
{
// We are another type, we need to tear down our children...
m_dynamic_type_info = class_type_or_name;
SetValueDidChange (true);
has_changed_type = true;
}
if (has_changed_type)
ClearDynamicTypeInformation ();
if (!m_address.IsValid() || m_address != dynamic_address)
{
if (m_address.IsValid())
SetValueDidChange (true);
// We've moved, so we should be fine...
m_address = dynamic_address;
lldb::TargetSP target_sp (GetTargetSP());
lldb::addr_t load_address = m_address.GetLoadAddress(target_sp.get());
m_value.GetScalar() = load_address;
}
m_dynamic_type_info = FixupTypeAndOrName(m_dynamic_type_info, *m_parent);
//m_value.SetContext (Value::eContextTypeClangType, corrected_type);
m_value.SetClangType (m_dynamic_type_info.GetClangASTType());
// Our address is the location of the dynamic type stored in memory. It isn't a load address,
// because we aren't pointing to the LOCATION that stores the pointer to us, we're pointing to us...
m_value.SetValueType(Value::eValueTypeScalar);
if (has_changed_type && log)
log->Printf("[%s %p] has a new dynamic type %s", GetName().GetCString(),
static_cast<void*>(this), GetTypeName().GetCString());
if (m_address.IsValid() && m_dynamic_type_info)
{
// The variable value is in the Scalar value inside the m_value.
// We can point our m_data right to it.
m_error = m_value.GetValueAsData (&exe_ctx, m_data, 0, GetModule().get());
if (m_error.Success())
{
if (GetClangType().IsAggregateType ())
{
// this value object represents an aggregate type whose
// children have values, but this object does not. So we
// say we are changed if our location has changed.
SetValueDidChange (m_value.GetValueType() != old_value.GetValueType() || m_value.GetScalar() != old_value.GetScalar());
}
SetValueIsValid (true);
return true;
}
}
// We get here if we've failed above...
SetValueIsValid (false);
return false;
}
bool
ValueObjectDynamicValue::IsInScope ()
{
return m_parent->IsInScope();
}
bool
ValueObjectDynamicValue::SetValueFromCString (const char *value_str, Error& error)
{
if (!UpdateValueIfNeeded(false))
{
error.SetErrorString("unable to read value");
return false;
}
uint64_t my_value = GetValueAsUnsigned(UINT64_MAX);
uint64_t parent_value = m_parent->GetValueAsUnsigned(UINT64_MAX);
if (my_value == UINT64_MAX || parent_value == UINT64_MAX)
{
error.SetErrorString("unable to read value");
return false;
}
// if we are at an offset from our parent, in order to set ourselves correctly we would need
// to change the new value so that it refers to the correct dynamic type. we choose not to deal
// with that - if anything more than a value overwrite is required, you should be using the
// expression parser instead of the value editing facility
if (my_value != parent_value)
{
// but NULL'ing out a value should always be allowed
if (strcmp(value_str,"0"))
{
error.SetErrorString("unable to modify dynamic value, use 'expression' command");
return false;
}
}
bool ret_val = m_parent->SetValueFromCString(value_str,error);
SetNeedsUpdate();
return ret_val;
}
bool
ValueObjectDynamicValue::SetData (DataExtractor &data, Error &error)
{
if (!UpdateValueIfNeeded(false))
{
error.SetErrorString("unable to read value");
return false;
}
uint64_t my_value = GetValueAsUnsigned(UINT64_MAX);
uint64_t parent_value = m_parent->GetValueAsUnsigned(UINT64_MAX);
if (my_value == UINT64_MAX || parent_value == UINT64_MAX)
{
error.SetErrorString("unable to read value");
return false;
}
// if we are at an offset from our parent, in order to set ourselves correctly we would need
// to change the new value so that it refers to the correct dynamic type. we choose not to deal
// with that - if anything more than a value overwrite is required, you should be using the
// expression parser instead of the value editing facility
if (my_value != parent_value)
{
// but NULL'ing out a value should always be allowed
lldb::offset_t offset = 0;
if (data.GetPointer(&offset) != 0)
{
error.SetErrorString("unable to modify dynamic value, use 'expression' command");
return false;
}
}
bool ret_val = m_parent->SetData(data, error);
SetNeedsUpdate();
return ret_val;
}