Andrew's git - gitweb.git/blob

   1/*
   2 * Copyright (c) 2005, 2006 Rene Scharfe
   3 */
   4#include "cache.h"
   5#include "config.h"
   6#include "tar.h"
   7#include "archive.h"
   8#include "streaming.h"
   9#include "run-command.h"
  10
  11#define RECORDSIZE      (512)
  12#define BLOCKSIZE       (RECORDSIZE * 20)
  13
  14static char block[BLOCKSIZE];
  15static unsigned long offset;
  16
  17static int tar_umask = 002;
  18
  19static int write_tar_filter_archive(const struct archiver *ar,
  20                                    struct archiver_args *args);
  21
  22/*
  23 * This is the max value that a ustar size header can specify, as it is fixed
  24 * at 11 octal digits. POSIX specifies that we switch to extended headers at
  25 * this size.
  26 *
  27 * Likewise for the mtime (which happens to use a buffer of the same size).
  28 */
  29#if ULONG_MAX == 0xFFFFFFFF
  30#define USTAR_MAX_SIZE ULONG_MAX
  31#define USTAR_MAX_MTIME ULONG_MAX
  32#else
  33#define USTAR_MAX_SIZE 077777777777UL
  34#define USTAR_MAX_MTIME 077777777777UL
  35#endif
  36
  37/* writes out the whole block, but only if it is full */
  38static void write_if_needed(void)
  39{
  40        if (offset == BLOCKSIZE) {
  41                write_or_die(1, block, BLOCKSIZE);
  42                offset = 0;
  43        }
  44}
  45
  46/*
  47 * queues up writes, so that all our write(2) calls write exactly one
  48 * full block; pads writes to RECORDSIZE
  49 */
  50static void do_write_blocked(const void *data, unsigned long size)
  51{
  52        const char *buf = data;
  53
  54        if (offset) {
  55                unsigned long chunk = BLOCKSIZE - offset;
  56                if (size < chunk)
  57                        chunk = size;
  58                memcpy(block + offset, buf, chunk);
  59                size -= chunk;
  60                offset += chunk;
  61                buf += chunk;
  62                write_if_needed();
  63        }
  64        while (size >= BLOCKSIZE) {
  65                write_or_die(1, buf, BLOCKSIZE);
  66                size -= BLOCKSIZE;
  67                buf += BLOCKSIZE;
  68        }
  69        if (size) {
  70                memcpy(block + offset, buf, size);
  71                offset += size;
  72        }
  73}
  74
  75static void finish_record(void)
  76{
  77        unsigned long tail;
  78        tail = offset % RECORDSIZE;
  79        if (tail)  {
  80                memset(block + offset, 0, RECORDSIZE - tail);
  81                offset += RECORDSIZE - tail;
  82        }
  83        write_if_needed();
  84}
  85
  86static void write_blocked(const void *data, unsigned long size)
  87{
  88        do_write_blocked(data, size);
  89        finish_record();
  90}
  91
  92/*
  93 * The end of tar archives is marked by 2*512 nul bytes and after that
  94 * follows the rest of the block (if any).
  95 */
  96static void write_trailer(void)
  97{
  98        int tail = BLOCKSIZE - offset;
  99        memset(block + offset, 0, tail);
 100        write_or_die(1, block, BLOCKSIZE);
 101        if (tail < 2 * RECORDSIZE) {
 102                memset(block, 0, offset);
 103                write_or_die(1, block, BLOCKSIZE);
 104        }
 105}
 106
 107/*
 108 * queues up writes, so that all our write(2) calls write exactly one
 109 * full block; pads writes to RECORDSIZE
 110 */
 111static int stream_blocked(const unsigned char *sha1)
 112{
 113        struct git_istream *st;
 114        enum object_type type;
 115        unsigned long sz;
 116        char buf[BLOCKSIZE];
 117        ssize_t readlen;
 118
 119        st = open_istream(sha1, &type, &sz, NULL);
 120        if (!st)
 121                return error("cannot stream blob %s", sha1_to_hex(sha1));
 122        for (;;) {
 123                readlen = read_istream(st, buf, sizeof(buf));
 124                if (readlen <= 0)
 125                        break;
 126                do_write_blocked(buf, readlen);
 127        }
 128        close_istream(st);
 129        if (!readlen)
 130                finish_record();
 131        return readlen;
 132}
 133
 134/*
 135 * pax extended header records have the format "%u %s=%s\n".  %u contains
 136 * the size of the whole string (including the %u), the first %s is the
 137 * keyword, the second one is the value.  This function constructs such a
 138 * string and appends it to a struct strbuf.
 139 */
 140static void strbuf_append_ext_header(struct strbuf *sb, const char *keyword,
 141                                     const char *value, unsigned int valuelen)
 142{
 143        int len, tmp;
 144
 145        /* "%u %s=%s\n" */
 146        len = 1 + 1 + strlen(keyword) + 1 + valuelen + 1;
 147        for (tmp = len; tmp > 9; tmp /= 10)
 148                len++;
 149
 150        strbuf_grow(sb, len);
 151        strbuf_addf(sb, "%u %s=", len, keyword);
 152        strbuf_add(sb, value, valuelen);
 153        strbuf_addch(sb, '\n');
 154}
 155
 156/*
 157 * Like strbuf_append_ext_header, but for numeric values.
 158 */
 159static void strbuf_append_ext_header_uint(struct strbuf *sb,
 160                                          const char *keyword,
 161                                          uintmax_t value)
 162{
 163        char buf[40]; /* big enough for 2^128 in decimal, plus NUL */
 164        int len;
 165
 166        len = xsnprintf(buf, sizeof(buf), "%"PRIuMAX, value);
 167        strbuf_append_ext_header(sb, keyword, buf, len);
 168}
 169
 170static unsigned int ustar_header_chksum(const struct ustar_header *header)
 171{
 172        const unsigned char *p = (const unsigned char *)header;
 173        unsigned int chksum = 0;
 174        while (p < (const unsigned char *)header->chksum)
 175                chksum += *p++;
 176        chksum += sizeof(header->chksum) * ' ';
 177        p += sizeof(header->chksum);
 178        while (p < (const unsigned char *)header + sizeof(struct ustar_header))
 179                chksum += *p++;
 180        return chksum;
 181}
 182
 183static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 184{
 185        size_t i = pathlen;
 186        if (i > 1 && path[i - 1] == '/')
 187                i--;
 188        if (i > maxlen)
 189                i = maxlen;
 190        do {
 191                i--;
 192        } while (i > 0 && path[i] != '/');
 193        return i;
 194}
 195
 196static void prepare_header(struct archiver_args *args,
 197                           struct ustar_header *header,
 198                           unsigned int mode, unsigned long size)
 199{
 200        xsnprintf(header->mode, sizeof(header->mode), "%07o", mode & 07777);
 201        xsnprintf(header->size, sizeof(header->size), "%011lo", S_ISREG(mode) ? size : 0);
 202        xsnprintf(header->mtime, sizeof(header->mtime), "%011lo", (unsigned long) args->time);
 203
 204        xsnprintf(header->uid, sizeof(header->uid), "%07o", 0);
 205        xsnprintf(header->gid, sizeof(header->gid), "%07o", 0);
 206        strlcpy(header->uname, "root", sizeof(header->uname));
 207        strlcpy(header->gname, "root", sizeof(header->gname));
 208        xsnprintf(header->devmajor, sizeof(header->devmajor), "%07o", 0);
 209        xsnprintf(header->devminor, sizeof(header->devminor), "%07o", 0);
 210
 211        memcpy(header->magic, "ustar", 6);
 212        memcpy(header->version, "00", 2);
 213
 214        xsnprintf(header->chksum, sizeof(header->chksum), "%07o", ustar_header_chksum(header));
 215}
 216
 217static void write_extended_header(struct archiver_args *args,
 218                                  const unsigned char *sha1,
 219                                  const void *buffer, unsigned long size)
 220{
 221        struct ustar_header header;
 222        unsigned int mode;
 223        memset(&header, 0, sizeof(header));
 224        *header.typeflag = TYPEFLAG_EXT_HEADER;
 225        mode = 0100666;
 226        xsnprintf(header.name, sizeof(header.name), "%s.paxheader", sha1_to_hex(sha1));
 227        prepare_header(args, &header, mode, size);
 228        write_blocked(&header, sizeof(header));
 229        write_blocked(buffer, size);
 230}
 231
 232static int write_tar_entry(struct archiver_args *args,
 233                           const unsigned char *sha1,
 234                           const char *path, size_t pathlen,
 235                           unsigned int mode)
 236{
 237        struct ustar_header header;
 238        struct strbuf ext_header = STRBUF_INIT;
 239        unsigned int old_mode = mode;
 240        unsigned long size, size_in_header;
 241        void *buffer;
 242        int err = 0;
 243
 244        memset(&header, 0, sizeof(header));
 245
 246        if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
 247                *header.typeflag = TYPEFLAG_DIR;
 248                mode = (mode | 0777) & ~tar_umask;
 249        } else if (S_ISLNK(mode)) {
 250                *header.typeflag = TYPEFLAG_LNK;
 251                mode |= 0777;
 252        } else if (S_ISREG(mode)) {
 253                *header.typeflag = TYPEFLAG_REG;
 254                mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 255        } else {
 256                return error("unsupported file mode: 0%o (SHA1: %s)",
 257                             mode, sha1_to_hex(sha1));
 258        }
 259        if (pathlen > sizeof(header.name)) {
 260                size_t plen = get_path_prefix(path, pathlen,
 261                                              sizeof(header.prefix));
 262                size_t rest = pathlen - plen - 1;
 263                if (plen > 0 && rest <= sizeof(header.name)) {
 264                        memcpy(header.prefix, path, plen);
 265                        memcpy(header.name, path + plen + 1, rest);
 266                } else {
 267                        xsnprintf(header.name, sizeof(header.name), "%s.data",
 268                                  sha1_to_hex(sha1));
 269                        strbuf_append_ext_header(&ext_header, "path",
 270                                                 path, pathlen);
 271                }
 272        } else
 273                memcpy(header.name, path, pathlen);
 274
 275        if (S_ISREG(mode) && !args->convert &&
 276            sha1_object_info(sha1, &size) == OBJ_BLOB &&
 277            size > big_file_threshold)
 278                buffer = NULL;
 279        else if (S_ISLNK(mode) || S_ISREG(mode)) {
 280                enum object_type type;
 281                buffer = sha1_file_to_archive(args, path, sha1, old_mode, &type, &size);
 282                if (!buffer)
 283                        return error("cannot read %s", sha1_to_hex(sha1));
 284        } else {
 285                buffer = NULL;
 286                size = 0;
 287        }
 288
 289        if (S_ISLNK(mode)) {
 290                if (size > sizeof(header.linkname)) {
 291                        xsnprintf(header.linkname, sizeof(header.linkname),
 292                                  "see %s.paxheader", sha1_to_hex(sha1));
 293                        strbuf_append_ext_header(&ext_header, "linkpath",
 294                                                 buffer, size);
 295                } else
 296                        memcpy(header.linkname, buffer, size);
 297        }
 298
 299        size_in_header = size;
 300        if (S_ISREG(mode) && size > USTAR_MAX_SIZE) {
 301                size_in_header = 0;
 302                strbuf_append_ext_header_uint(&ext_header, "size", size);
 303        }
 304
 305        prepare_header(args, &header, mode, size_in_header);
 306
 307        if (ext_header.len > 0) {
 308                write_extended_header(args, sha1, ext_header.buf,
 309                                      ext_header.len);
 310        }
 311        strbuf_release(&ext_header);
 312        write_blocked(&header, sizeof(header));
 313        if (S_ISREG(mode) && size > 0) {
 314                if (buffer)
 315                        write_blocked(buffer, size);
 316                else
 317                        err = stream_blocked(sha1);
 318        }
 319        free(buffer);
 320        return err;
 321}
 322
 323static void write_global_extended_header(struct archiver_args *args)
 324{
 325        const unsigned char *sha1 = args->commit_sha1;
 326        struct strbuf ext_header = STRBUF_INIT;
 327        struct ustar_header header;
 328        unsigned int mode;
 329
 330        if (sha1)
 331                strbuf_append_ext_header(&ext_header, "comment",
 332                                         sha1_to_hex(sha1), 40);
 333        if (args->time > USTAR_MAX_MTIME) {
 334                strbuf_append_ext_header_uint(&ext_header, "mtime",
 335                                              args->time);
 336                args->time = USTAR_MAX_MTIME;
 337        }
 338
 339        if (!ext_header.len)
 340                return;
 341
 342        memset(&header, 0, sizeof(header));
 343        *header.typeflag = TYPEFLAG_GLOBAL_HEADER;
 344        mode = 0100666;
 345        xsnprintf(header.name, sizeof(header.name), "pax_global_header");
 346        prepare_header(args, &header, mode, ext_header.len);
 347        write_blocked(&header, sizeof(header));
 348        write_blocked(ext_header.buf, ext_header.len);
 349        strbuf_release(&ext_header);
 350}
 351
 352static struct archiver **tar_filters;
 353static int nr_tar_filters;
 354static int alloc_tar_filters;
 355
 356static struct archiver *find_tar_filter(const char *name, int len)
 357{
 358        int i;
 359        for (i = 0; i < nr_tar_filters; i++) {
 360                struct archiver *ar = tar_filters[i];
 361                if (!strncmp(ar->name, name, len) && !ar->name[len])
 362                        return ar;
 363        }
 364        return NULL;
 365}
 366
 367static int tar_filter_config(const char *var, const char *value, void *data)
 368{
 369        struct archiver *ar;
 370        const char *name;
 371        const char *type;
 372        int namelen;
 373
 374        if (parse_config_key(var, "tar", &name, &namelen, &type) < 0 || !name)
 375                return 0;
 376
 377        ar = find_tar_filter(name, namelen);
 378        if (!ar) {
 379                ar = xcalloc(1, sizeof(*ar));
 380                ar->name = xmemdupz(name, namelen);
 381                ar->write_archive = write_tar_filter_archive;
 382                ar->flags = ARCHIVER_WANT_COMPRESSION_LEVELS;
 383                ALLOC_GROW(tar_filters, nr_tar_filters + 1, alloc_tar_filters);
 384                tar_filters[nr_tar_filters++] = ar;
 385        }
 386
 387        if (!strcmp(type, "command")) {
 388                if (!value)
 389                        return config_error_nonbool(var);
 390                free(ar->data);
 391                ar->data = xstrdup(value);
 392                return 0;
 393        }
 394        if (!strcmp(type, "remote")) {
 395                if (git_config_bool(var, value))
 396                        ar->flags |= ARCHIVER_REMOTE;
 397                else
 398                        ar->flags &= ~ARCHIVER_REMOTE;
 399                return 0;
 400        }
 401
 402        return 0;
 403}
 404
 405static int git_tar_config(const char *var, const char *value, void *cb)
 406{
 407        if (!strcmp(var, "tar.umask")) {
 408                if (value && !strcmp(value, "user")) {
 409                        tar_umask = umask(0);
 410                        umask(tar_umask);
 411                } else {
 412                        tar_umask = git_config_int(var, value);
 413                }
 414                return 0;
 415        }
 416
 417        return tar_filter_config(var, value, cb);
 418}
 419
 420static int write_tar_archive(const struct archiver *ar,
 421                             struct archiver_args *args)
 422{
 423        int err = 0;
 424
 425        write_global_extended_header(args);
 426        err = write_archive_entries(args, write_tar_entry);
 427        if (!err)
 428                write_trailer();
 429        return err;
 430}
 431
 432static int write_tar_filter_archive(const struct archiver *ar,
 433                                    struct archiver_args *args)
 434{
 435        struct strbuf cmd = STRBUF_INIT;
 436        struct child_process filter = CHILD_PROCESS_INIT;
 437        const char *argv[2];
 438        int r;
 439
 440        if (!ar->data)
 441                die("BUG: tar-filter archiver called with no filter defined");
 442
 443        strbuf_addstr(&cmd, ar->data);
 444        if (args->compression_level >= 0)
 445                strbuf_addf(&cmd, " -%d", args->compression_level);
 446
 447        argv[0] = cmd.buf;
 448        argv[1] = NULL;
 449        filter.argv = argv;
 450        filter.use_shell = 1;
 451        filter.in = -1;
 452
 453        if (start_command(&filter) < 0)
 454                die_errno("unable to start '%s' filter", argv[0]);
 455        close(1);
 456        if (dup2(filter.in, 1) < 0)
 457                die_errno("unable to redirect descriptor");
 458        close(filter.in);
 459
 460        r = write_tar_archive(ar, args);
 461
 462        close(1);
 463        if (finish_command(&filter) != 0)
 464                die("'%s' filter reported error", argv[0]);
 465
 466        strbuf_release(&cmd);
 467        return r;
 468}
 469
 470static struct archiver tar_archiver = {
 471        "tar",
 472        write_tar_archive,
 473        ARCHIVER_REMOTE
 474};
 475
 476void init_tar_archiver(void)
 477{
 478        int i;
 479        register_archiver(&tar_archiver);
 480
 481        tar_filter_config("tar.tgz.command", "gzip -cn", NULL);
 482        tar_filter_config("tar.tgz.remote", "true", NULL);
 483        tar_filter_config("tar.tar.gz.command", "gzip -cn", NULL);
 484        tar_filter_config("tar.tar.gz.remote", "true", NULL);
 485        git_config(git_tar_config, NULL);
 486        for (i = 0; i < nr_tar_filters; i++) {
 487                /* omit any filters that never had a command configured */
 488                if (tar_filters[i]->data)
 489                        register_archiver(tar_filters[i]);
 490        }
 491}