Index: llvm/trunk/include/llvm/Support/TarWriter.h =================================================================== --- llvm/trunk/include/llvm/Support/TarWriter.h +++ llvm/trunk/include/llvm/Support/TarWriter.h @@ -0,0 +1,32 @@ +//===-- llvm/Support/TarWriter.h - Tar archive file creator -----*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_SUPPORT_TAR_WRITER_H +#define LLVM_SUPPORT_TAR_WRITER_H + +#include "llvm/ADT/StringRef.h" +#include "llvm/Support/Error.h" +#include "llvm/Support/raw_ostream.h" + +namespace llvm { +class TarWriter { +public: + static Expected> create(StringRef OutputPath, + StringRef BaseDir); + + void append(StringRef Path, StringRef Data); + +private: + TarWriter(int FD, StringRef BaseDir); + raw_fd_ostream OS; + std::string BaseDir; +}; +} + +#endif Index: llvm/trunk/lib/Support/CMakeLists.txt =================================================================== --- llvm/trunk/lib/Support/CMakeLists.txt +++ llvm/trunk/lib/Support/CMakeLists.txt @@ -90,6 +90,7 @@ StringSaver.cpp StringRef.cpp SystemUtils.cpp + TarWriter.cpp TargetParser.cpp ThreadPool.cpp Timer.cpp Index: llvm/trunk/lib/Support/TarWriter.cpp =================================================================== --- llvm/trunk/lib/Support/TarWriter.cpp +++ llvm/trunk/lib/Support/TarWriter.cpp @@ -0,0 +1,169 @@ +//===-- TarWriter.cpp - Tar archive file creator --------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// TarWriter class provides a feature to create a tar archive file. +// +// I put emphasis on simplicity over comprehensiveness when +// implementing this class because we don't need a full-fledged +// archive file generator in LLVM at the moment. +// +// The filename field in the Unix V7 tar header is 100 bytes, which is +// apparently too small. Various extensions were proposed and +// implemented to fix the issue. The writer implemented in this file +// uses PAX extension headers. +// +// Note that we emit PAX headers even if filenames fit in the V7 +// header for the sake of simplicity. So, generated files are N +// kilobyte larger than the ideal where N is the number of files in +// archives. In practice, I think you don't need to worry about that. +// +// The PAX header is standardized in IEEE Std 1003.1-2001. +// +// The struct definition of UstarHeader is copied from +// https://www.freebsd.org/cgi/man.cgi?query=tar&sektion=5 +// +//===----------------------------------------------------------------------===// + +#include "llvm/Support/TarWriter.h" +#include "llvm/ADT/StringRef.h" +#include "llvm/Support/FileSystem.h" +#include "llvm/Support/MathExtras.h" + +using namespace llvm; + +// Each file in an archive must be aligned to this block size. +static const int BlockSize = 512; + +struct UstarHeader { + char Name[100]; + char Mode[8]; + char Uid[8]; + char Gid[8]; + char Size[12]; + char Mtime[12]; + char Checksum[8]; + char TypeFlag; + char Linkname[100]; + char Magic[6]; + char Version[2]; + char Uname[32]; + char Gname[32]; + char DevMajor[8]; + char DevMinor[8]; + char Prefix[155]; + char Pad[12]; +}; +static_assert(sizeof(UstarHeader) == BlockSize, "invalid Ustar header"); + +// A PAX attribute is in the form of " =\n" +// where is the length of the entire string including +// the length field itself. An example string is this. +// +// 25 ctime=1084839148.1212\n +// +// This function create such string. +static std::string formatPax(StringRef Key, const Twine &Val) { + int Len = Key.size() + Val.str().size() + 3; // +3 for " ", "=" and "\n" + + // We need to compute total size twice because appending + // a length field could change total size by one. + int Total = Len + Twine(Len).str().size(); + Total = Len + Twine(Total).str().size(); + return (Twine(Total) + " " + Key + "=" + Val + "\n").str(); +} + +// Headers in tar files must be aligned to 512 byte boundaries. +// This function writes null bytes so that the file is a multiple +// of 512 bytes. +static void pad(raw_fd_ostream &OS) { + uint64_t Pos = OS.tell(); + OS.seek(alignTo(Pos, BlockSize)); +} + +// Computes a checksum for a tar header. +static void computeChecksum(UstarHeader &Hdr) { + // Before computing a checksum, checksum field must be + // filled with space characters. + memset(Hdr.Checksum, ' ', sizeof(Hdr.Checksum)); + + // Compute a checksum and set it to the checksum field. + unsigned Chksum = 0; + for (size_t I = 0; I < sizeof(Hdr); ++I) + Chksum += reinterpret_cast(&Hdr)[I]; + sprintf(Hdr.Checksum, "%06o", Chksum); +} + +// Create a tar header and write it to a given output stream. +static void writePaxHeader(raw_fd_ostream &OS, const Twine &Path) { + // A PAX header consists of a 512-byte header followed + // by key-value strings. First, create key-value strings. + std::string PaxAttr = formatPax("path", Path); + + // Create a 512-byte header. + UstarHeader Hdr = {}; + sprintf(Hdr.Size, "%011lo", PaxAttr.size()); + Hdr.TypeFlag = 'x'; // PAX magic + memcpy(Hdr.Magic, "ustar", 6); // Ustar magic + computeChecksum(Hdr); + + // Write them down. + OS << StringRef(reinterpret_cast(&Hdr), sizeof(Hdr)); + OS << PaxAttr; + pad(OS); +} + +// The PAX header is an extended format, so a PAX header needs +// to be followed by a "real" header. +static void writeUstarHeader(raw_fd_ostream &OS, size_t Size) { + UstarHeader Hdr = {}; + strcpy(Hdr.Mode, "0000664"); + sprintf(Hdr.Size, "%011lo", Size); + memcpy(Hdr.Magic, "ustar", 6); + + computeChecksum(Hdr); + OS << StringRef(reinterpret_cast(&Hdr), sizeof(Hdr)); +} + +// We want to use '/' as a path separator even on Windows. +// This function canonicalizes a given path. +static std::string canonicalize(std::string S) { +#ifdef LLVM_ON_WIN32 + std::replace(S.begin(), S.end(), '\\', '/'); +#endif + return S; +} + +// Creates a TarWriter instance and returns it. +Expected> TarWriter::create(StringRef OutputPath, + StringRef BaseDir) { + int FD; + if (std::error_code EC = openFileForWrite(OutputPath, FD, sys::fs::F_None)) + return make_error("cannot open " + OutputPath, EC); + return std::unique_ptr(new TarWriter(FD, BaseDir)); +} + +TarWriter::TarWriter(int FD, StringRef BaseDir) + : OS(FD, /*shouldClose=*/true, /*unbuffered=*/false), BaseDir(BaseDir) {} + +// Append a given file to an archive. +void TarWriter::append(StringRef Path, StringRef Data) { + // Write Path and Data. + writePaxHeader(OS, BaseDir + "/" + canonicalize(Path)); + writeUstarHeader(OS, Data.size()); + OS << Data; + pad(OS); + + // POSIX requires tar archives end with two null blocks. + // Here, we write the terminator and then seek back, so that + // the file being output is terminated correctly at any moment. + uint64_t Pos = OS.tell(); + OS << std::string(BlockSize * 2, '\0'); + OS.seek(Pos); + OS.flush(); +}