//===- VirtualFileSystem.h - Virtual File System Layer ----------*- C++ -*-===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception // //===----------------------------------------------------------------------===// // /// \file /// Defines the virtual file system interface vfs::FileSystem. // //===----------------------------------------------------------------------===// #ifndef LLVM_SUPPORT_VIRTUALFILESYSTEM_H #define LLVM_SUPPORT_VIRTUALFILESYSTEM_H #include "llvm/ADT/IntrusiveRefCntPtr.h" #include "llvm/ADT/None.h" #include "llvm/ADT/Optional.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringRef.h" #include "llvm/Support/Chrono.h" #include "llvm/Support/ErrorOr.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/Path.h" #include "llvm/Support/SourceMgr.h" #include #include #include #include #include #include #include #include #include namespace llvm { class MemoryBuffer; class MemoryBufferRef; class Twine; namespace vfs { /// The result of a \p status operation. class Status { std::string Name; llvm::sys::fs::UniqueID UID; llvm::sys::TimePoint<> MTime; uint32_t User; uint32_t Group; uint64_t Size; llvm::sys::fs::file_type Type = llvm::sys::fs::file_type::status_error; llvm::sys::fs::perms Perms; public: // FIXME: remove when files support multiple names bool IsVFSMapped = false; Status() = default; Status(const llvm::sys::fs::file_status &Status); Status(const Twine &Name, llvm::sys::fs::UniqueID UID, llvm::sys::TimePoint<> MTime, uint32_t User, uint32_t Group, uint64_t Size, llvm::sys::fs::file_type Type, llvm::sys::fs::perms Perms); /// Get a copy of a Status with a different name. static Status copyWithNewName(const Status &In, const Twine &NewName); static Status copyWithNewName(const llvm::sys::fs::file_status &In, const Twine &NewName); /// Returns the name that should be used for this file or directory. StringRef getName() const { return Name; } /// @name Status interface from llvm::sys::fs /// @{ llvm::sys::fs::file_type getType() const { return Type; } llvm::sys::fs::perms getPermissions() const { return Perms; } llvm::sys::TimePoint<> getLastModificationTime() const { return MTime; } llvm::sys::fs::UniqueID getUniqueID() const { return UID; } uint32_t getUser() const { return User; } uint32_t getGroup() const { return Group; } uint64_t getSize() const { return Size; } /// @} /// @name Status queries /// These are static queries in llvm::sys::fs. /// @{ bool equivalent(const Status &Other) const; bool isDirectory() const; bool isRegularFile() const; bool isOther() const; bool isSymlink() const; bool isStatusKnown() const; bool exists() const; /// @} }; /// Represents an open file. class File { public: /// Destroy the file after closing it (if open). /// Sub-classes should generally call close() inside their destructors. We /// cannot do that from the base class, since close is virtual. virtual ~File(); /// Get the status of the file. virtual llvm::ErrorOr status() = 0; /// Get the name of the file virtual llvm::ErrorOr getName() { if (auto Status = status()) return Status->getName().str(); else return Status.getError(); } /// Get the contents of the file as a \p MemoryBuffer. virtual llvm::ErrorOr> getBuffer(const Twine &Name, int64_t FileSize = -1, bool RequiresNullTerminator = true, bool IsVolatile = false) = 0; /// Closes the file. virtual std::error_code close() = 0; }; /// A member of a directory, yielded by a directory_iterator. /// Only information available on most platforms is included. class directory_entry { std::string Path; llvm::sys::fs::file_type Type = llvm::sys::fs::file_type::type_unknown; public: directory_entry() = default; directory_entry(std::string Path, llvm::sys::fs::file_type Type) : Path(std::move(Path)), Type(Type) {} llvm::StringRef path() const { return Path; } llvm::sys::fs::file_type type() const { return Type; } }; namespace detail { /// An interface for virtual file systems to provide an iterator over the /// (non-recursive) contents of a directory. struct DirIterImpl { virtual ~DirIterImpl(); /// Sets \c CurrentEntry to the next entry in the directory on success, /// to directory_entry() at end, or returns a system-defined \c error_code. virtual std::error_code increment() = 0; directory_entry CurrentEntry; }; } // namespace detail /// An input iterator over the entries in a virtual path, similar to /// llvm::sys::fs::directory_iterator. class directory_iterator { std::shared_ptr Impl; // Input iterator semantics on copy public: directory_iterator(std::shared_ptr I) : Impl(std::move(I)) { assert(Impl.get() != nullptr && "requires non-null implementation"); if (Impl->CurrentEntry.path().empty()) Impl.reset(); // Normalize the end iterator to Impl == nullptr. } /// Construct an 'end' iterator. directory_iterator() = default; /// Equivalent to operator++, with an error code. directory_iterator &increment(std::error_code &EC) { assert(Impl && "attempting to increment past end"); EC = Impl->increment(); if (Impl->CurrentEntry.path().empty()) Impl.reset(); // Normalize the end iterator to Impl == nullptr. return *this; } const directory_entry &operator*() const { return Impl->CurrentEntry; } const directory_entry *operator->() const { return &Impl->CurrentEntry; } bool operator==(const directory_iterator &RHS) const { if (Impl && RHS.Impl) return Impl->CurrentEntry.path() == RHS.Impl->CurrentEntry.path(); return !Impl && !RHS.Impl; } bool operator!=(const directory_iterator &RHS) const { return !(*this == RHS); } }; class FileSystem; namespace detail { /// Keeps state for the recursive_directory_iterator. struct RecDirIterState { std::stack> Stack; bool HasNoPushRequest = false; }; } // end namespace detail /// An input iterator over the recursive contents of a virtual path, /// similar to llvm::sys::fs::recursive_directory_iterator. class recursive_directory_iterator { FileSystem *FS; std::shared_ptr State; // Input iterator semantics on copy. public: recursive_directory_iterator(FileSystem &FS, const Twine &Path, std::error_code &EC); /// Construct an 'end' iterator. recursive_directory_iterator() = default; /// Equivalent to operator++, with an error code. recursive_directory_iterator &increment(std::error_code &EC); const directory_entry &operator*() const { return *State->Stack.top(); } const directory_entry *operator->() const { return &*State->Stack.top(); } bool operator==(const recursive_directory_iterator &Other) const { return State == Other.State; // identity } bool operator!=(const recursive_directory_iterator &RHS) const { return !(*this == RHS); } /// Gets the current level. Starting path is at level 0. int level() const { assert(!State->Stack.empty() && "Cannot get level without any iteration state"); return State->Stack.size() - 1; } void no_push() { State->HasNoPushRequest = true; } }; /// The virtual file system interface. class FileSystem : public llvm::ThreadSafeRefCountedBase { public: virtual ~FileSystem(); /// Get the status of the entry at \p Path, if one exists. virtual llvm::ErrorOr status(const Twine &Path) = 0; /// Get a \p File object for the file at \p Path, if one exists. virtual llvm::ErrorOr> openFileForRead(const Twine &Path) = 0; /// This is a convenience method that opens a file, gets its content and then /// closes the file. llvm::ErrorOr> getBufferForFile(const Twine &Name, int64_t FileSize = -1, bool RequiresNullTerminator = true, bool IsVolatile = false); /// Get a directory_iterator for \p Dir. /// \note The 'end' iterator is directory_iterator(). virtual directory_iterator dir_begin(const Twine &Dir, std::error_code &EC) = 0; /// Set the working directory. This will affect all following operations on /// this file system and may propagate down for nested file systems. virtual std::error_code setCurrentWorkingDirectory(const Twine &Path) = 0; /// Get the working directory of this file system. virtual llvm::ErrorOr getCurrentWorkingDirectory() const = 0; /// Gets real path of \p Path e.g. collapse all . and .. patterns, resolve /// symlinks. For real file system, this uses `llvm::sys::fs::real_path`. /// This returns errc::operation_not_permitted if not implemented by subclass. virtual std::error_code getRealPath(const Twine &Path, SmallVectorImpl &Output) const; /// Check whether a file exists. Provided for convenience. bool exists(const Twine &Path); /// Is the file mounted on a local filesystem? virtual std::error_code isLocal(const Twine &Path, bool &Result); /// Make \a Path an absolute path. /// /// Makes \a Path absolute using the current directory if it is not already. /// An empty \a Path will result in the current directory. /// /// /absolute/path => /absolute/path /// relative/../path => /relative/../path /// /// \param Path A path that is modified to be an absolute path. /// \returns success if \a path has been made absolute, otherwise a /// platform-specific error_code. virtual std::error_code makeAbsolute(SmallVectorImpl &Path) const; }; /// Gets an \p vfs::FileSystem for the 'real' file system, as seen by /// the operating system. /// The working directory is linked to the process's working directory. /// (This is usually thread-hostile). IntrusiveRefCntPtr getRealFileSystem(); /// Create an \p vfs::FileSystem for the 'real' file system, as seen by /// the operating system. /// It has its own working directory, independent of (but initially equal to) /// that of the process. std::unique_ptr createPhysicalFileSystem(); /// A file system that allows overlaying one \p AbstractFileSystem on top /// of another. /// /// Consists of a stack of >=1 \p FileSystem objects, which are treated as being /// one merged file system. When there is a directory that exists in more than /// one file system, the \p OverlayFileSystem contains a directory containing /// the union of their contents. The attributes (permissions, etc.) of the /// top-most (most recently added) directory are used. When there is a file /// that exists in more than one file system, the file in the top-most file /// system overrides the other(s). class OverlayFileSystem : public FileSystem { using FileSystemList = SmallVector, 1>; /// The stack of file systems, implemented as a list in order of /// their addition. FileSystemList FSList; public: OverlayFileSystem(IntrusiveRefCntPtr Base); /// Pushes a file system on top of the stack. void pushOverlay(IntrusiveRefCntPtr FS); llvm::ErrorOr status(const Twine &Path) override; llvm::ErrorOr> openFileForRead(const Twine &Path) override; directory_iterator dir_begin(const Twine &Dir, std::error_code &EC) override; llvm::ErrorOr getCurrentWorkingDirectory() const override; std::error_code setCurrentWorkingDirectory(const Twine &Path) override; std::error_code isLocal(const Twine &Path, bool &Result) override; std::error_code getRealPath(const Twine &Path, SmallVectorImpl &Output) const override; using iterator = FileSystemList::reverse_iterator; using const_iterator = FileSystemList::const_reverse_iterator; using reverse_iterator = FileSystemList::iterator; using const_reverse_iterator = FileSystemList::const_iterator; /// Get an iterator pointing to the most recently added file system. iterator overlays_begin() { return FSList.rbegin(); } const_iterator overlays_begin() const { return FSList.rbegin(); } /// Get an iterator pointing one-past the least recently added file system. iterator overlays_end() { return FSList.rend(); } const_iterator overlays_end() const { return FSList.rend(); } /// Get an iterator pointing to the least recently added file system. reverse_iterator overlays_rbegin() { return FSList.begin(); } const_reverse_iterator overlays_rbegin() const { return FSList.begin(); } /// Get an iterator pointing one-past the most recently added file system. reverse_iterator overlays_rend() { return FSList.end(); } const_reverse_iterator overlays_rend() const { return FSList.end(); } }; /// By default, this delegates all calls to the underlying file system. This /// is useful when derived file systems want to override some calls and still /// proxy other calls. class ProxyFileSystem : public FileSystem { public: explicit ProxyFileSystem(IntrusiveRefCntPtr FS) : FS(std::move(FS)) {} llvm::ErrorOr status(const Twine &Path) override { return FS->status(Path); } llvm::ErrorOr> openFileForRead(const Twine &Path) override { return FS->openFileForRead(Path); } directory_iterator dir_begin(const Twine &Dir, std::error_code &EC) override { return FS->dir_begin(Dir, EC); } llvm::ErrorOr getCurrentWorkingDirectory() const override { return FS->getCurrentWorkingDirectory(); } std::error_code setCurrentWorkingDirectory(const Twine &Path) override { return FS->setCurrentWorkingDirectory(Path); } std::error_code getRealPath(const Twine &Path, SmallVectorImpl &Output) const override { return FS->getRealPath(Path, Output); } std::error_code isLocal(const Twine &Path, bool &Result) override { return FS->isLocal(Path, Result); } protected: FileSystem &getUnderlyingFS() { return *FS; } private: IntrusiveRefCntPtr FS; virtual void anchor(); }; namespace detail { class InMemoryDirectory; class InMemoryFile; } // namespace detail /// An in-memory file system. class InMemoryFileSystem : public FileSystem { std::unique_ptr Root; std::string WorkingDirectory; bool UseNormalizedPaths = true; /// If HardLinkTarget is non-null, a hardlink is created to the To path which /// must be a file. If it is null then it adds the file as the public addFile. bool addFile(const Twine &Path, time_t ModificationTime, std::unique_ptr Buffer, Optional User, Optional Group, Optional Type, Optional Perms, const detail::InMemoryFile *HardLinkTarget); public: explicit InMemoryFileSystem(bool UseNormalizedPaths = true); ~InMemoryFileSystem() override; /// Add a file containing a buffer or a directory to the VFS with a /// path. The VFS owns the buffer. If present, User, Group, Type /// and Perms apply to the newly-created file or directory. /// \return true if the file or directory was successfully added, /// false if the file or directory already exists in the file system with /// different contents. bool addFile(const Twine &Path, time_t ModificationTime, std::unique_ptr Buffer, Optional User = None, Optional Group = None, Optional Type = None, Optional Perms = None); /// Add a hard link to a file. /// Here hard links are not intended to be fully equivalent to the classical /// filesystem. Both the hard link and the file share the same buffer and /// status (and thus have the same UniqueID). Because of this there is no way /// to distinguish between the link and the file after the link has been /// added. /// /// The To path must be an existing file or a hardlink. The From file must not /// have been added before. The To Path must not be a directory. The From Node /// is added as a hard link which points to the resolved file of To Node. /// \return true if the above condition is satisfied and hardlink was /// successfully created, false otherwise. bool addHardLink(const Twine &From, const Twine &To); /// Add a buffer to the VFS with a path. The VFS does not own the buffer. /// If present, User, Group, Type and Perms apply to the newly-created file /// or directory. /// \return true if the file or directory was successfully added, /// false if the file or directory already exists in the file system with /// different contents. bool addFileNoOwn(const Twine &Path, time_t ModificationTime, const llvm::MemoryBufferRef &Buffer, Optional User = None, Optional Group = None, Optional Type = None, Optional Perms = None); std::string toString() const; /// Return true if this file system normalizes . and .. in paths. bool useNormalizedPaths() const { return UseNormalizedPaths; } llvm::ErrorOr status(const Twine &Path) override; llvm::ErrorOr> openFileForRead(const Twine &Path) override; directory_iterator dir_begin(const Twine &Dir, std::error_code &EC) override; llvm::ErrorOr getCurrentWorkingDirectory() const override { return WorkingDirectory; } /// Canonicalizes \p Path by combining with the current working /// directory and normalizing the path (e.g. remove dots). If the current /// working directory is not set, this returns errc::operation_not_permitted. /// /// This doesn't resolve symlinks as they are not supported in in-memory file /// system. std::error_code getRealPath(const Twine &Path, SmallVectorImpl &Output) const override; std::error_code isLocal(const Twine &Path, bool &Result) override; std::error_code setCurrentWorkingDirectory(const Twine &Path) override; }; /// Get a globally unique ID for a virtual file or directory. llvm::sys::fs::UniqueID getNextVirtualUniqueID(); /// Gets a \p FileSystem for a virtual file system described in YAML /// format. std::unique_ptr getVFSFromYAML(std::unique_ptr Buffer, llvm::SourceMgr::DiagHandlerTy DiagHandler, StringRef YAMLFilePath, void *DiagContext = nullptr, IntrusiveRefCntPtr ExternalFS = getRealFileSystem()); struct YAMLVFSEntry { template YAMLVFSEntry(T1 &&VPath, T2 &&RPath, bool IsDirectory = false) : VPath(std::forward(VPath)), RPath(std::forward(RPath)), IsDirectory(IsDirectory) {} std::string VPath; std::string RPath; bool IsDirectory = false; }; class VFSFromYamlDirIterImpl; class RedirectingFileSystemParser; /// A virtual file system parsed from a YAML file. /// /// Currently, this class allows creating virtual directories and mapping /// virtual file paths to existing external files, available in \c ExternalFS. /// /// The basic structure of the parsed file is: /// \verbatim /// { /// 'version': , /// /// 'roots': [ /// /// ] /// } /// \endverbatim /// /// All configuration options are optional. /// 'case-sensitive': /// 'use-external-names': /// 'overlay-relative': /// 'fallthrough': /// /// Virtual directories are represented as /// \verbatim /// { /// 'type': 'directory', /// 'name': , /// 'contents': [ ] /// } /// \endverbatim /// /// The default attributes for virtual directories are: /// \verbatim /// MTime = now() when created /// Perms = 0777 /// User = Group = 0 /// Size = 0 /// UniqueID = unspecified unique value /// \endverbatim /// /// Re-mapped files are represented as /// \verbatim /// { /// 'type': 'file', /// 'name': , /// 'use-external-name': # Optional /// 'external-contents': /// } /// \endverbatim /// /// and inherit their attributes from the external contents. /// /// In both cases, the 'name' field may contain multiple path components (e.g. /// /path/to/file). However, any directory that contains more than one child /// must be uniquely represented by a directory entry. class RedirectingFileSystem : public vfs::FileSystem { public: enum EntryKind { EK_Directory, EK_File }; /// A single file or directory in the VFS. class Entry { EntryKind Kind; std::string Name; public: Entry(EntryKind K, StringRef Name) : Kind(K), Name(Name) {} virtual ~Entry() = default; StringRef getName() const { return Name; } EntryKind getKind() const { return Kind; } }; class RedirectingDirectoryEntry : public Entry { std::vector> Contents; Status S; public: RedirectingDirectoryEntry(StringRef Name, std::vector> Contents, Status S) : Entry(EK_Directory, Name), Contents(std::move(Contents)), S(std::move(S)) {} RedirectingDirectoryEntry(StringRef Name, Status S) : Entry(EK_Directory, Name), S(std::move(S)) {} Status getStatus() { return S; } void addContent(std::unique_ptr Content) { Contents.push_back(std::move(Content)); } Entry *getLastContent() const { return Contents.back().get(); } using iterator = decltype(Contents)::iterator; iterator contents_begin() { return Contents.begin(); } iterator contents_end() { return Contents.end(); } static bool classof(const Entry *E) { return E->getKind() == EK_Directory; } }; class RedirectingFileEntry : public Entry { public: enum NameKind { NK_NotSet, NK_External, NK_Virtual }; private: std::string ExternalContentsPath; NameKind UseName; public: RedirectingFileEntry(StringRef Name, StringRef ExternalContentsPath, NameKind UseName) : Entry(EK_File, Name), ExternalContentsPath(ExternalContentsPath), UseName(UseName) {} StringRef getExternalContentsPath() const { return ExternalContentsPath; } /// whether to use the external path as the name for this file. bool useExternalName(bool GlobalUseExternalName) const { return UseName == NK_NotSet ? GlobalUseExternalName : (UseName == NK_External); } NameKind getUseName() const { return UseName; } static bool classof(const Entry *E) { return E->getKind() == EK_File; } }; private: friend class VFSFromYamlDirIterImpl; friend class RedirectingFileSystemParser; bool shouldUseExternalFS() const { return IsFallthrough; } /// Canonicalize path by removing ".", "..", "./", components. This is /// a VFS request, do not bother about symlinks in the path components /// but canonicalize in order to perform the correct entry search. std::error_code makeCanonical(SmallVectorImpl &Path) const; // In a RedirectingFileSystem, keys can be specified in Posix or Windows // style (or even a mixture of both), so this comparison helper allows // slashes (representing a root) to match backslashes (and vice versa). Note // that, other than the root, path components should not contain slashes or // backslashes. bool pathComponentMatches(llvm::StringRef lhs, llvm::StringRef rhs) const { if ((CaseSensitive ? lhs.equals(rhs) : lhs.equals_lower(rhs))) return true; return (lhs == "/" && rhs == "\\") || (lhs == "\\" && rhs == "/"); } /// The root(s) of the virtual file system. std::vector> Roots; /// The current working directory of the file system. std::string WorkingDirectory; /// The file system to use for external references. IntrusiveRefCntPtr ExternalFS; /// If IsRelativeOverlay is set, this represents the directory /// path that should be prefixed to each 'external-contents' entry /// when reading from YAML files. std::string ExternalContentsPrefixDir; /// @name Configuration /// @{ /// Whether to perform case-sensitive comparisons. /// /// Currently, case-insensitive matching only works correctly with ASCII. bool CaseSensitive = #ifdef _WIN32 false; #else true; #endif /// IsRelativeOverlay marks whether a ExternalContentsPrefixDir path must /// be prefixed in every 'external-contents' when reading from YAML files. bool IsRelativeOverlay = false; /// Whether to use to use the value of 'external-contents' for the /// names of files. This global value is overridable on a per-file basis. bool UseExternalNames = true; /// Whether to attempt a file lookup in external file system after it wasn't /// found in VFS. bool IsFallthrough = true; /// @} RedirectingFileSystem(IntrusiveRefCntPtr ExternalFS); /// Looks up the path [Start, End) in \p From, possibly /// recursing into the contents of \p From if it is a directory. ErrorOr lookupPath(llvm::sys::path::const_iterator Start, llvm::sys::path::const_iterator End, Entry *From) const; /// Get the status of a given an \c Entry. ErrorOr status(const Twine &Path, Entry *E); public: /// Looks up \p Path in \c Roots. ErrorOr lookupPath(StringRef Path) const; /// Parses \p Buffer, which is expected to be in YAML format and /// returns a virtual file system representing its contents. static std::unique_ptr create(std::unique_ptr Buffer, SourceMgr::DiagHandlerTy DiagHandler, StringRef YAMLFilePath, void *DiagContext, IntrusiveRefCntPtr ExternalFS); /// Redirect each of the remapped files from first to second. static std::unique_ptr create(ArrayRef> RemappedFiles, bool UseExternalNames, FileSystem &ExternalFS); ErrorOr status(const Twine &Path) override; ErrorOr> openFileForRead(const Twine &Path) override; std::error_code getRealPath(const Twine &Path, SmallVectorImpl &Output) const override; llvm::ErrorOr getCurrentWorkingDirectory() const override; std::error_code setCurrentWorkingDirectory(const Twine &Path) override; std::error_code isLocal(const Twine &Path, bool &Result) override; std::error_code makeAbsolute(SmallVectorImpl &Path) const override; directory_iterator dir_begin(const Twine &Dir, std::error_code &EC) override; void setExternalContentsPrefixDir(StringRef PrefixDir); StringRef getExternalContentsPrefixDir() const; void setFallthrough(bool Fallthrough); std::vector getRoots() const; void dump(raw_ostream &OS) const; void dumpEntry(raw_ostream &OS, Entry *E, int NumSpaces = 0) const; #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) LLVM_DUMP_METHOD void dump() const; #endif }; /// Collect all pairs of entries from the /// \p YAMLFilePath. This is used by the module dependency collector to forward /// the entries into the reproducer output VFS YAML file. void collectVFSFromYAML( std::unique_ptr Buffer, llvm::SourceMgr::DiagHandlerTy DiagHandler, StringRef YAMLFilePath, SmallVectorImpl &CollectedEntries, void *DiagContext = nullptr, IntrusiveRefCntPtr ExternalFS = getRealFileSystem()); class YAMLVFSWriter { std::vector Mappings; Optional IsCaseSensitive; Optional IsOverlayRelative; Optional UseExternalNames; std::string OverlayDir; void addEntry(StringRef VirtualPath, StringRef RealPath, bool IsDirectory); public: YAMLVFSWriter() = default; void addFileMapping(StringRef VirtualPath, StringRef RealPath); void addDirectoryMapping(StringRef VirtualPath, StringRef RealPath); void setCaseSensitivity(bool CaseSensitive) { IsCaseSensitive = CaseSensitive; } void setUseExternalNames(bool UseExtNames) { UseExternalNames = UseExtNames; } void setOverlayDir(StringRef OverlayDirectory) { IsOverlayRelative = true; OverlayDir.assign(OverlayDirectory.str()); } const std::vector &getMappings() const { return Mappings; } void write(llvm::raw_ostream &OS); }; } // namespace vfs } // namespace llvm #endif // LLVM_SUPPORT_VIRTUALFILESYSTEM_H