[SharedCache] Make the shared cache file lookup more straightforward

We already had separated this part out into the `CacheProcessor` but its more like a builder than anything, so this refactors that out into a `SharedCacheBuilder`.

- Separate out the `CacheProcessor` and simplify its implementation
- Move more implementation specific stuff to the shared cache view
- Prepare for more validation of the shared cache to detect irregular or incomplete shared cache information
- Deduplicate a lot of file vs. project file lookup stuff
- Stop copying all images when getting the number of images to log
- Allow user to select another directory to scan for shared cache files, might make this a warning instead to prevent users from relying on this behavior

We will likely follow this commit up soon with better open behavior, maybe open with options can specify a list of cache entry files? Shouldn't be too much effort aside from making the UI modal for that.

Author: emesare

view/sharedcache/core/SharedCache.cpp

@@ -537,197 +537,3 @@ std::optional<CacheSymbol> SharedCache::GetSymbolWithName(const std::string& nam
 		return std::nullopt;
 	return GetSymbolAt(it->second);
 }
-
-CacheProcessor::CacheProcessor(Ref<BinaryView> view)
-{
-	m_view = std::move(view);
-	m_logger = new Logger("SharedCache.Processor", m_view->GetFile()->GetSessionId());
-}
-
-bool CacheProcessor::ProcessCache(SharedCache& cache)
-{
-	// If we are in a project, use the project cache processor.
-	if (m_view->GetFile()->GetProjectFile())
-		return ProcessProjectCache(cache);
-	return ProcessFileCache(cache);
-}
-
-bool CacheProcessor::ProcessFileCache(SharedCache& cache)
-{
-	// We assume that the binary view location has all the files we need.
-	// If we ever want to allow users to override the shared cache file location
-	// we should really make a cache processor constructor with entry file paths.
-	std::string baseFilePath = m_view->GetFile()->GetOriginalFilename();
-	std::string baseFileName = BaseFileName(baseFilePath);
-
-	// If we don't have an original file path, prompt the user to select one.
-	if (baseFilePath.empty())
-	{
-		if (!GetOpenFileNameInput(baseFilePath, "Please select the base shared cache file"))
-		{
-			// User did not give a file, tell them that we will try and scan the file directory.
-			m_logger->LogWarn("Failed to get original file path, will try and scan the file directory.");
-		}
-		// Update so next load we don't need to prompt the user.
-		// TODO: What happens for a remote project? This will point at what exactly?
-		m_view->GetFile()->SetOriginalFilename(baseFilePath);
-	}
-
-	// Add this file to the entries
-	try
-	{
-		auto baseEntry = CacheEntry::FromFile(baseFilePath, baseFileName, CacheEntryType::Primary);
-		if (!baseEntry.has_value())
-			return false;
-
-		// Before we do anything else, add this to the cache so it's available to other entries.
-		cache.AddEntry(std::move(*baseEntry));
-	}
-	catch (const std::exception& e)
-	{
-		// Just return false so the view init can continue.
-		m_logger->LogErrorF("Failed to load base entry {}... {}", baseFileName, e.what());
-		return false;
-	}
-
-	// Locate all possible related entry files and add them to the cache.
-	std::filesystem::path basePath = std::filesystem::path(baseFilePath).parent_path();
-	for (const auto& entry : std::filesystem::directory_iterator(basePath))
-	{
-		if (!entry.is_regular_file())
-			continue;
-		auto currentFilePath = entry.path().string();
-		auto currentFileName = BaseFileName(currentFilePath);
-		// Skip our base file, obviously.
-		if (currentFilePath == baseFilePath)
-			continue;
-		// Filter files that don't contain the base file name i.e. "dyld_shared_cache_arm64e"
-		if (currentFilePath.find(baseFileName) == std::string::npos)
-			continue;
-		// Skip map files, they contain some nice information... we don't use.
-		if (entry.path().extension() == ".map")
-			continue;
-		// Skip bndb files!
-		if (entry.path().extension() == ".bndb")
-			continue;
-		try
-		{
-			auto additionalEntry = CacheEntry::FromFile(currentFilePath, currentFileName, CacheEntryType::Secondary);
-			if (!additionalEntry.has_value())
-			{
-				m_logger->LogErrorF("Failed to load entry {}...", currentFileName);
-				continue;
-			}
-
-			// Add this file as an entry to the cache
-			cache.AddEntry(std::move(*additionalEntry));
-		}
-		catch (const std::exception& e)
-		{
-			m_logger->LogErrorF("Failed to load entry {}... {}", currentFileName, e.what());
-		}
-	}
-
-	return true;
-}
-
-bool CacheProcessor::ProcessProjectCache(SharedCache& cache)
-{
-	auto baseProjectFile = m_view->GetFile()->GetProjectFile();
-	auto baseProjectFileFolder = baseProjectFile->GetFolder();
-	std::string baseFilePath = baseProjectFile->GetPathOnDisk();
-	std::string baseFileName = baseProjectFile->GetName();
-	// This will point to the path on disk for original non-bndb file.
-	std::string originalFilePath = m_view->GetFile()->GetOriginalFilename();
-
-	// If we don't have an original file path, prompt the user to select one.
-	if (originalFilePath.empty())
-	{
-		if (!GetOpenFileNameInput(originalFilePath, "Please select the base shared cache file"))
-		{
-			// User did not give a file, tell them that we will try and scan the project directory.
-			m_logger->LogWarn("Failed to get original file path, will try and scan the project directory.");
-		}
-		// Update so next load we don't need to prompt the user.
-		m_view->GetFile()->SetOriginalFilename(originalFilePath);
-	}
-
-	// Remove the .bndb extension if present ("dyld_shared_cache_arm64e.bndb" => "dyld_shared_cache_arm64e)
-	// TODO: This is a little annoying, we need to do this because the file accessor we have is separate
-	// TODO: from the view file accessor. If we either made it so that we can parse from the BNDB file accessor,
-	// TODO: or... something better than this.
-	if (baseFileName.find(".bndb") != std::string::npos)
-	{
-		baseFileName = baseFileName.substr(0, baseFileName.size() - 5);
-		// Search for the backing file.
-		for (const auto& projectFile : baseProjectFile->GetProject()->GetFiles())
-		{
-			// Skip files not in the same folder.
-			auto projectFileFolder = projectFile->GetFolder();
-			if (baseProjectFileFolder && projectFileFolder)
-				if (baseProjectFileFolder->GetId() != projectFileFolder->GetId())
-					continue;
-
-			auto projectFilePath = projectFile->GetPathOnDisk();
-			auto projectFileName = projectFile->GetName();
-			if (projectFileName == baseFileName || projectFilePath == originalFilePath)
-			{
-				// Use the real file instead.
-				baseFilePath = projectFilePath;
-				baseFileName = projectFileName;
-				baseProjectFile = projectFile;
-				break;
-			}
-		}
-	}
-
-	// Add this file to the entries
-	auto baseEntry = CacheEntry::FromFile(baseFilePath, baseFileName, CacheEntryType::Primary);
-	if (!baseEntry.has_value())
-		return false;
-
-	// Before we do anything else, add this to the cache so it's available to other entries.
-	cache.AddEntry(std::move(*baseEntry));
-
-	// Enumerate the project files folder to gather the necessary sub caches.
-	const auto project = baseProjectFile->GetProject();
-	const auto folder = baseProjectFile->GetFolder();
-	for (const auto& projectFile : project->GetFiles())
-	{
-		auto projectFilePath = projectFile->GetPathOnDisk();
-		auto projectFileName = projectFile->GetName();
-		auto currentFolder = projectFile->GetFolder();
-		// Skip our base project file, obviously.
-		if (projectFile->GetId() == baseProjectFile->GetId())
-			continue;
-		// Filter files that don't contain the base file name i.e. "dyld_shared_cache_arm64e"
-		if (projectFileName.find(baseFileName) == std::string::npos)
-			continue;
-		// Filter out .map files, they contain some nice info for rebasing... that we don't do.
-		if (projectFileName.find(".map") != std::string::npos)
-			continue;
-		// Filter out .bndb files!
-		if (projectFileName.find(".bndb") != std::string::npos)
-			continue;
-		// If both top level, or we are in the same folder as the base project file add it.
-		if ((!folder && !currentFolder) || (folder && currentFolder))
-		{
-			try
-			{
-				auto additionalEntry = CacheEntry::FromFile(projectFilePath, projectFileName, CacheEntryType::Secondary);
-				if (!additionalEntry.has_value())
-				{
-					m_logger->LogErrorF("Failed to load entry {}...", projectFileName);
-					continue;
-				}
-
-				// Add this file as an entry to the cache
-				cache.AddEntry(std::move(*additionalEntry));
-			}
-			catch (const std::exception& e)
-			{}
-		}
-	}
-
-	return true;
-}

view/sharedcache/core/SharedCache.h

@@ -7,28 +7,22 @@
 #include "MachO.h"
 #include "VirtualMemory.h"
 
-class SharedCache;
-
 struct CacheSymbol
 {
 	BNSymbolType type;
 	uint64_t address;
 	std::string name;
 
 	CacheSymbol() = default;
-
 	CacheSymbol(BNSymbolType type, uint64_t address, std::string name) :
 		type(type), address(address), name(std::move(name))
 	{}
-
 	~CacheSymbol() = default;
 
 	CacheSymbol(const CacheSymbol& other) = default;
-
 	CacheSymbol& operator=(const CacheSymbol& other) = default;
 
 	CacheSymbol(CacheSymbol&& other) noexcept = default;
-
 	CacheSymbol& operator=(CacheSymbol&& other) noexcept = default;
 
 	std::pair<std::string, BinaryNinja::Ref<BinaryNinja::Type>> DemangledName(BinaryNinja::BinaryView& view) const;
@@ -56,15 +50,12 @@ struct CacheRegion
 	BNSegmentFlag flags;
 
 	CacheRegion() = default;
-
 	~CacheRegion() = default;
 
 	CacheRegion(const CacheRegion& other) = default;
-
 	CacheRegion& operator=(const CacheRegion& other) = default;
 
 	CacheRegion(CacheRegion&& other) noexcept = default;
-
 	CacheRegion& operator=(CacheRegion&& other) noexcept = default;
 
 	AddressRange AsAddressRange() const { return {start, start + size}; }
@@ -95,15 +86,12 @@ struct CacheImage
 	std::shared_ptr<SharedCacheMachOHeader> header;
 
 	CacheImage() = default;
-
 	~CacheImage() = default;
 
 	CacheImage(const CacheImage& other) = default;
-
 	CacheImage& operator=(const CacheImage& other) = default;
 
 	CacheImage(CacheImage&& other) noexcept = default;
-
 	CacheImage& operator=(CacheImage&& other) noexcept = default;
 
 	// Get the file name from the path.
@@ -145,11 +133,9 @@ class CacheEntry
 		std::vector<dyld_cache_mapping_info> mappings, std::vector<std::pair<std::string, dyld_cache_image_info>> images);
 
 	CacheEntry() = default;
-
 	CacheEntry(const CacheEntry&) = default;
 
 	CacheEntry(CacheEntry&&) = default;
-
 	CacheEntry& operator=(CacheEntry&&) = default;
 
 	// Construct a cache entry from the file on disk.
@@ -213,6 +199,7 @@ class SharedCache
 	bool AddNonOverlappingRegion(CacheRegion region);
 
 public:
+	SharedCache() = default;
 	explicit SharedCache(uint64_t addressSize);
 
 	SharedCache(const SharedCache &) = delete;
@@ -269,23 +256,3 @@ class SharedCache
 
 	std::optional<CacheSymbol> GetSymbolWithName(const std::string& name);
 };
-
-// This constructs a Cache, give it a file path, and it will add all relevant cache entries.
-class CacheProcessor
-{
-	BinaryNinja::Ref<BinaryNinja::BinaryView> m_view;
-	BinaryNinja::Ref<BinaryNinja::Logger> m_logger;
-
-public:
-	explicit CacheProcessor(BinaryNinja::Ref<BinaryNinja::BinaryView> view);
-
-	// Construct a cache from the root file, this will parse the cache header and locate all
-	// applicable cache entries and add them as well.
-	bool ProcessCache(SharedCache& cache);
-
-	// Process a cache on the file system, this is for when not using a project.
-	bool ProcessFileCache(SharedCache& cache);
-
-	// Process a cache using Binary Ninja's project system.
-	bool ProcessProjectCache(SharedCache& cache);
-};