const char *
archive_version_details(void)
{
static struct archive_string str;
static int init = 0;
if (!init) {
archive_string_init(&str);
archive_strcat(&str, ARCHIVE_VERSION_STRING);
#ifdef HAVE_ZLIB_H
archive_strcat(&str, " zlib/");
archive_strcat(&str, ZLIB_VERSION);
#endif
#ifdef HAVE_LZMA_H
archive_strcat(&str, " liblzma/");
archive_strcat(&str, LZMA_VERSION_STRING);
#endif
#ifdef HAVE_BZLIB_H
{
const char *p = BZ2_bzlibVersion();
const char *sep = strchr(p, ',');
if (sep == NULL)
sep = p + strlen(p);
archive_strcat(&str, " bz2lib/");
archive_strncat(&str, p, sep - p);
}
#endif
#if defined(HAVE_LZ4_H) && defined(HAVE_LIBLZ4)
archive_string_sprintf(&str, " liblz4/%d.%d.%d",
LZ4_VERSION_MAJOR, LZ4_VERSION_MINOR, LZ4_VERSION_RELEASE);
#endif
}
return str.s;
}
static void
test_archive_strncat(void)
{
struct archive_string s;
archive_string_init(&s);
assertExactString(0, 0, NULL, s);
/* perfect length */
assert(&s == archive_strncat(&s, "snafu", 5));
assertExactString(5, EXTENT, "snafu", s);
/* short read */
assert(&s == archive_strncat(&s, "barbazqux", 3));
assertExactString(8, EXTENT, "snafubar", s);
/* long read is ok too! */
assert(&s == archive_strncat(&s, "snafu", 8));
assertExactString(13, EXTENT, "snafubarsnafu", s);
}
static void
mtree_quote(struct archive_string *s, const char *str)
{
const char *start;
char buf[4];
unsigned char c;
for (start = str; *str != '\0'; ++str) {
if (mtree_safe_char(*str))
continue;
if (start != str)
archive_strncat(s, start, str - start);
c = (unsigned char)*str;
buf[0] = '\\';
buf[1] = (c / 64) + '0';
buf[2] = (c / 8 % 8) + '0';
buf[3] = (c % 8) + '0';
archive_strncat(s, buf, 4);
start = str + 1;
}
if (start != str)
archive_strncat(s, start, str - start);
}
static void
mtree_indent(struct mtree_writer *mtree)
{
int i, fn;
const char *r, *s, *x;
fn = 1;
s = r = mtree->ebuf.s;
x = NULL;
while (*r == ' ')
r++;
while ((r = strchr(r, ' ')) != NULL) {
if (fn) {
fn = 0;
archive_strncat(&mtree->buf, s, r - s);
if (r -s > INDENTNAMELEN) {
archive_strncat(&mtree->buf, " \\\n", 3);
for (i = 0; i < (INDENTNAMELEN + 1); i++)
archive_strappend_char(&mtree->buf, ' ');
} else {
for (i = r -s; i < (INDENTNAMELEN + 1); i++)
archive_strappend_char(&mtree->buf, ' ');
}
s = ++r;
x = NULL;
continue;
}
if (r - s <= MAXLINELEN - 3 - INDENTNAMELEN)
x = r++;
else {
if (x == NULL)
x = r;
archive_strncat(&mtree->buf, s, x - s);
archive_strncat(&mtree->buf, " \\\n", 3);
for (i = 0; i < (INDENTNAMELEN + 1); i++)
archive_strappend_char(&mtree->buf, ' ');
s = r = ++x;
x = NULL;
}
}
if (x != NULL && strlen(s) > MAXLINELEN - 3 - INDENTNAMELEN) {
/* Last keyword is longer. */
archive_strncat(&mtree->buf, s, x - s);
archive_strncat(&mtree->buf, " \\\n", 3);
for (i = 0; i < (INDENTNAMELEN + 1); i++)
archive_strappend_char(&mtree->buf, ' ');
s = ++x;
}
archive_strcat(&mtree->buf, s);
archive_string_empty(&mtree->ebuf);
}
/*
* Copy the given string to the buffer, quoting all shell meta characters
* found.
*/
static void
shar_quote(struct archive_string *buf, const char *str, int in_shell)
{
static const char meta[] = "\n \t'`\";&<>()|*?{}[]\\$!#^~";
size_t len;
while (*str != '\0') {
if ((len = strcspn(str, meta)) != 0) {
archive_strncat(buf, str, len);
str += len;
} else if (*str == '\n') {
if (in_shell)
archive_strcat(buf, "\"\n\"");
else
archive_strcat(buf, "\\n");
++str;
} else {
archive_strappend_char(buf, '\\');
archive_strappend_char(buf, *str);
++str;
}
}
}
const char *
archive_version_details(void)
{
static struct archive_string str;
static int init = 0;
const char *zlib = archive_zlib_version();
const char *liblzma = archive_liblzma_version();
const char *bzlib = archive_bzlib_version();
const char *liblz4 = archive_liblz4_version();
if (!init) {
archive_string_init(&str);
archive_strcat(&str, ARCHIVE_VERSION_STRING);
if (zlib != NULL) {
archive_strcat(&str, " zlib/");
archive_strcat(&str, zlib);
}
if (liblzma) {
archive_strcat(&str, " liblzma/");
archive_strcat(&str, liblzma);
}
if (bzlib) {
const char *p = bzlib;
const char *sep = strchr(p, ',');
if (sep == NULL)
sep = p + strlen(p);
archive_strcat(&str, " bz2lib/");
archive_strncat(&str, p, sep - p);
}
if (liblz4) {
archive_strcat(&str, " liblz4/");
archive_strcat(&str, liblz4);
}
}
return str.s;
}
/* private routines */
static void
xstrftime(struct archive_string *as, const char *fmt, time_t t)
{
/** like strftime(3) but for time_t objects */
struct tm *rt;
#if defined(HAVE_GMTIME_R) || defined(HAVE__GMTIME64_S)
struct tm timeHere;
#endif
char strtime[100];
size_t len;
#ifdef HAVE_GMTIME_R
if ((rt = gmtime_r(&t, &timeHere)) == NULL)
return;
#elif defined(HAVE__GMTIME64_S)
_gmtime64_s(&timeHere, &t);
#else
if ((rt = gmtime(&t)) == NULL)
return;
#endif
/* leave the hard yacker to our role model strftime() */
len = strftime(strtime, sizeof(strtime)-1, fmt, rt);
archive_strncat(as, strtime, len);
}
static int
add_pattern_from_file(struct archive_match *a, struct match_list *mlist,
int mbs, const void *pathname, int nullSeparator)
{
struct archive *ar;
struct archive_entry *ae;
struct archive_string as;
const void *buff;
size_t size;
int64_t offset;
int r;
ar = archive_read_new();
if (ar == NULL) {
archive_set_error(&(a->archive), ENOMEM, "No memory");
return (ARCHIVE_FATAL);
}
r = archive_read_support_format_raw(ar);
r = archive_read_support_format_empty(ar);
if (r != ARCHIVE_OK) {
archive_copy_error(&(a->archive), ar);
archive_read_free(ar);
return (r);
}
if (mbs)
r = archive_read_open_filename(ar, pathname, 512*20);
else
r = archive_read_open_filename_w(ar, pathname, 512*20);
if (r != ARCHIVE_OK) {
archive_copy_error(&(a->archive), ar);
archive_read_free(ar);
return (r);
}
r = archive_read_next_header(ar, &ae);
if (r != ARCHIVE_OK) {
archive_read_free(ar);
if (r == ARCHIVE_EOF) {
return (ARCHIVE_OK);
} else {
archive_copy_error(&(a->archive), ar);
return (r);
}
}
archive_string_init(&as);
while ((r = archive_read_data_block(ar, &buff, &size, &offset))
== ARCHIVE_OK) {
const char *b = (const char *)buff;
while (size) {
const char *s = (const char *)b;
size_t length = 0;
int found_separator = 0;
while (length < size) {
if (nullSeparator) {
if (*b == '\0') {
found_separator = 1;
break;
}
} else {
if (*b == 0x0d || *b == 0x0a) {
found_separator = 1;
break;
}
}
b++;
length++;
}
if (!found_separator) {
archive_strncat(&as, s, length);
/* Read next data block. */
break;
}
b++;
size -= length + 1;
archive_strncat(&as, s, length);
/* If the line is not empty, add the pattern. */
if (archive_strlen(&as) > 0) {
/* Add pattern. */
r = add_pattern_mbs(a, mlist, as.s);
if (r != ARCHIVE_OK) {
archive_read_free(ar);
archive_string_free(&as);
return (r);
}
archive_string_empty(&as);
}
}
}
/* If an error occurred, report it immediately. */
if (r < ARCHIVE_OK) {
archive_copy_error(&(a->archive), ar);
archive_read_free(ar);
archive_string_free(&as);
return (r);
}
/* If the line is not empty, add the pattern. */
if (r == ARCHIVE_EOF && archive_strlen(&as) > 0) {
/* Add pattern. */
r = add_pattern_mbs(a, mlist, as.s);
if (r != ARCHIVE_OK) {
archive_read_free(ar);
archive_string_free(&as);
return (r);
}
}
archive_read_free(ar);
archive_string_free(&as);
return (ARCHIVE_OK);
}
static int
_warc_header(struct archive_write *a, struct archive_entry *entry)
{
struct warc_s *w = a->format_data;
struct archive_string hdr;
#define MAX_HDR_SIZE 512
/* check whether warcinfo record needs outputting */
if (!w->omit_warcinfo) {
ssize_t r;
warc_essential_hdr_t wi = {
WT_INFO,
/*uri*/NULL,
/*urn*/NULL,
/*rtm*/0,
/*mtm*/0,
/*cty*/"application/warc-fields",
/*len*/sizeof(warcinfo) - 1U,
};
wi.rtime = w->now;
wi.mtime = w->now;
archive_string_init(&hdr);
r = _popul_ehdr(&hdr, MAX_HDR_SIZE, wi);
if (r >= 0) {
/* jackpot! */
/* now also use HDR buffer for the actual warcinfo */
archive_strncat(&hdr, warcinfo, sizeof(warcinfo) -1);
/* append end-of-record indicator */
archive_strncat(&hdr, "\r\n\r\n", 4);
/* write to output stream */
__archive_write_output(a, hdr.s, archive_strlen(&hdr));
}
/* indicate we're done with file header writing */
w->omit_warcinfo = 1U;
archive_string_free(&hdr);
}
if (archive_entry_pathname(entry) == NULL) {
archive_set_error(&a->archive, EINVAL,
"Invalid filename");
return (ARCHIVE_WARN);
}
w->typ = archive_entry_filetype(entry);
w->populz = 0U;
if (w->typ == AE_IFREG) {
warc_essential_hdr_t rh = {
WT_RSRC,
/*uri*/NULL,
/*urn*/NULL,
/*rtm*/0,
/*mtm*/0,
/*cty*/NULL,
/*len*/0,
};
ssize_t r;
rh.tgturi = archive_entry_pathname(entry);
rh.rtime = w->now;
rh.mtime = archive_entry_mtime(entry);
rh.cntlen = (size_t)archive_entry_size(entry);
archive_string_init(&hdr);
r = _popul_ehdr(&hdr, MAX_HDR_SIZE, rh);
if (r < 0) {
/* don't bother */
archive_set_error(
&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"cannot archive file");
return (ARCHIVE_WARN);
}
/* otherwise append to output stream */
__archive_write_output(a, hdr.s, r);
/* and let subsequent calls to _data() know about the size */
w->populz = rh.cntlen;
archive_string_free(&hdr);
return (ARCHIVE_OK);
}
/* just resort to erroring as per Tim's advice */
archive_set_error(
&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"WARC can only process regular files");
return (ARCHIVE_FAILED);
}
static void
parse_rockridge(struct iso9660 *iso9660, struct file_info *file,
const unsigned char *p, const unsigned char *end)
{
(void)iso9660; /* UNUSED */
while (p + 4 < end /* Enough space for another entry. */
&& p[0] >= 'A' && p[0] <= 'Z' /* Sanity-check 1st char of name. */
&& p[1] >= 'A' && p[1] <= 'Z' /* Sanity-check 2nd char of name. */
&& p + p[2] <= end) { /* Sanity-check length. */
const unsigned char *data = p + 4;
int data_length = p[2] - 4;
int version = p[3];
/*
* Yes, each 'if' here does test p[0] again.
* Otherwise, the fall-through handling to catch
* unsupported extensions doesn't work.
*/
switch(p[0]) {
case 'C':
if (p[0] == 'C' && p[1] == 'E' && version == 1) {
/*
* CE extension comprises:
* 8 byte sector containing extension
* 8 byte offset w/in above sector
* 8 byte length of continuation
*/
file->ce_offset = toi(data, 4)
* iso9660->logical_block_size
+ toi(data + 8, 4);
file->ce_size = toi(data + 16, 4);
break;
}
/* FALLTHROUGH */
case 'N':
if (p[0] == 'N' && p[1] == 'M' && version == 1
&& *data == 0) {
/* NM extension with flag byte == 0 */
/*
* NM extension comprises:
* one byte flag
* rest is long name
*/
/* TODO: Obey flags. */
char *old_name = file->name;
data++; /* Skip flag byte. */
data_length--;
file->name = (char *)malloc(data_length + 1);
if (file->name != NULL) {
free(old_name);
memcpy(file->name, data, data_length);
file->name[data_length] = '\0';
} else
file->name = old_name;
break;
}
/* FALLTHROUGH */
case 'P':
if (p[0] == 'P' && p[1] == 'D' && version == 1) {
/*
* PD extension is padding;
* contents are always ignored.
*/
break;
}
if (p[0] == 'P' && p[1] == 'X' && version == 1) {
/*
* PX extension comprises:
* 8 bytes for mode,
* 8 bytes for nlinks,
* 8 bytes for uid,
* 8 bytes for gid,
* 8 bytes for inode.
*/
if (data_length == 32) {
file->mode = toi(data, 4);
file->nlinks = toi(data + 8, 4);
file->uid = toi(data + 16, 4);
file->gid = toi(data + 24, 4);
file->inode = toi(data + 32, 4);
}
break;
}
/* FALLTHROUGH */
case 'R':
if (p[0] == 'R' && p[1] == 'R' && version == 1) {
iso9660->seenRockridge = 1;
/*
* RR extension comprises:
* one byte flag value
*/
/* TODO: Handle RR extension. */
break;
}
/* FALLTHROUGH */
case 'S':
if (p[0] == 'S' && p[1] == 'L' && version == 1
&& *data == 0) {
int cont = 1;
/* SL extension with flags == 0 */
/* TODO: handle non-zero flag values. */
data++; /* Skip flag byte. */
data_length--;
while (data_length > 0) {
unsigned char flag = *data++;
unsigned char nlen = *data++;
data_length -= 2;
if (cont == 0)
archive_strcat(&file->symlink, "/");
cont = 0;
switch(flag) {
case 0x01: /* Continue */
archive_strncat(&file->symlink,
(const char *)data, nlen);
cont = 1;
break;
case 0x02: /* Current */
archive_strcat(&file->symlink, ".");
break;
case 0x04: /* Parent */
archive_strcat(&file->symlink, "..");
break;
case 0x08: /* Root */
case 0x10: /* Volume root */
archive_string_empty(&file->symlink);
break;
case 0x20: /* Hostname */
archive_strcat(&file->symlink, "hostname");
break;
case 0:
archive_strncat(&file->symlink,
(const char *)data, nlen);
break;
default:
/* TODO: issue a warning ? */
break;
}
data += nlen;
data_length -= nlen;
}
break;
}
if (p[0] == 'S' && p[1] == 'P'
&& version == 1 && data_length == 7
&& data[0] == (unsigned char)'\xbe'
&& data[1] == (unsigned char)'\xef') {
/*
* SP extension stores the suspOffset
* (Number of bytes to skip between
* filename and SUSP records.)
* It is mandatory by the SUSP standard
* (IEEE 1281).
*
* It allows SUSP to coexist with
* non-SUSP uses of the System
* Use Area by placing non-SUSP data
* before SUSP data.
*
* TODO: Add a check for 'SP' in
* first directory entry, disable all SUSP
* processing if not found.
*/
iso9660->suspOffset = data[2];
break;
}
if (p[0] == 'S' && p[1] == 'T'
&& data_length == 0 && version == 1) {
/*
* ST extension marks end of this
* block of SUSP entries.
*
* It allows SUSP to coexist with
* non-SUSP uses of the System
* Use Area by placing non-SUSP data
* after SUSP data.
*/
return;
}
case 'T':
if (p[0] == 'T' && p[1] == 'F' && version == 1) {
char flag = data[0];
/*
* TF extension comprises:
* one byte flag
* create time (optional)
* modify time (optional)
* access time (optional)
* attribute time (optional)
* Time format and presence of fields
* is controlled by flag bits.
*/
data++;
if (flag & 0x80) {
/* Use 17-byte time format. */
if (flag & 1) /* Create time. */
data += 17;
if (flag & 2) { /* Modify time. */
file->mtime = isodate17(data);
data += 17;
}
if (flag & 4) { /* Access time. */
file->atime = isodate17(data);
data += 17;
}
if (flag & 8) { /* Attribute time. */
file->ctime = isodate17(data);
data += 17;
}
} else {
/* Use 7-byte time format. */
if (flag & 1) /* Create time. */
data += 7;
if (flag & 2) { /* Modify time. */
file->mtime = isodate7(data);
data += 7;
}
if (flag & 4) { /* Access time. */
file->atime = isodate7(data);
data += 7;
}
if (flag & 8) { /* Attribute time. */
file->ctime = isodate7(data);
data += 7;
}
}
break;
}
/* FALLTHROUGH */
default:
/* The FALLTHROUGHs above leave us here for
* any unsupported extension. */
{
const unsigned char *t;
fprintf(stderr, "\nUnsupported RRIP extension for %s\n", file->name);
fprintf(stderr, " %c%c(%d):", p[0], p[1], data_length);
for (t = data; t < data + data_length && t < data + 16; t++)
fprintf(stderr, " %02x", *t);
fprintf(stderr, "\n");
}
}
p += p[2];
}
}
static int
slurp_central_directory(struct archive_read *a, struct zip *zip)
{
unsigned i;
int64_t correction;
static const struct archive_rb_tree_ops rb_ops = {
&cmp_node, &cmp_key
};
static const struct archive_rb_tree_ops rb_rsrc_ops = {
&rsrc_cmp_node, &rsrc_cmp_key
};
/*
* Consider the archive file we are reading may be SFX.
* So we have to calculate a SFX header size to revise
* ZIP header offsets.
*/
correction = zip->end_of_central_directory_offset -
(zip->central_directory_offset + zip->central_directory_size);
/* The central directory offset is relative value, and so
* we revise this offset for SFX. */
zip->central_directory_offset += correction;
__archive_read_seek(a, zip->central_directory_offset, SEEK_SET);
zip->offset = zip->central_directory_offset;
__archive_rb_tree_init(&zip->tree, &rb_ops);
__archive_rb_tree_init(&zip->tree_rsrc, &rb_rsrc_ops);
zip->zip_entries = calloc(zip->central_directory_entries,
sizeof(struct zip_entry));
for (i = 0; i < zip->central_directory_entries; ++i) {
struct zip_entry *zip_entry = &zip->zip_entries[i];
size_t filename_length, extra_length, comment_length;
uint32_t external_attributes;
const char *name, *p, *r;
if ((p = __archive_read_ahead(a, 46, NULL)) == NULL)
return ARCHIVE_FATAL;
if (memcmp(p, "PK\001\002", 4) != 0) {
archive_set_error(&a->archive,
-1, "Invalid central directory signature");
return ARCHIVE_FATAL;
}
zip->have_central_directory = 1;
/* version = p[4]; */
zip_entry->system = p[5];
/* version_required = archive_le16dec(p + 6); */
zip_entry->flags = archive_le16dec(p + 8);
zip_entry->compression = (char)archive_le16dec(p + 10);
zip_entry->mtime = zip_time(p + 12);
zip_entry->crc32 = archive_le32dec(p + 16);
zip_entry->compressed_size = archive_le32dec(p + 20);
zip_entry->uncompressed_size = archive_le32dec(p + 24);
filename_length = archive_le16dec(p + 28);
extra_length = archive_le16dec(p + 30);
comment_length = archive_le16dec(p + 32);
/* disk_start = archive_le16dec(p + 34); */ /* Better be zero. */
/* internal_attributes = archive_le16dec(p + 36); */ /* text bit */
external_attributes = archive_le32dec(p + 38);
zip_entry->local_header_offset =
archive_le32dec(p + 42) + correction;
/* If we can't guess the mode, leave it zero here;
when we read the local file header we might get
more information. */
zip_entry->mode = 0;
if (zip_entry->system == 3) {
zip_entry->mode = external_attributes >> 16;
}
/*
* Mac resource fork files are stored under the
* "__MACOSX/" directory, so we should check if
* it is.
*/
/* Make sure we have the file name. */
if ((p = __archive_read_ahead(a, 46 + filename_length, NULL))
== NULL)
return ARCHIVE_FATAL;
name = p + 46;
r = rsrc_basename(name, filename_length);
if (filename_length >= 9 &&
strncmp("__MACOSX/", name, 9) == 0) {
/* If this file is not a resource fork nor
* a directory. We should treat it as a non
* resource fork file to expose it. */
if (name[filename_length-1] != '/' &&
(r - name < 3 || r[0] != '.' || r[1] != '_')) {
__archive_rb_tree_insert_node(&zip->tree,
&zip_entry->node);
/* Expose its parent directories. */
expose_parent_dirs(zip, name, filename_length);
} else {
/* This file is a resource fork file or
* a directory. */
archive_strncpy(&(zip_entry->rsrcname), name,
filename_length);
__archive_rb_tree_insert_node(&zip->tree_rsrc,
&zip_entry->node);
}
} else {
/* Generate resource fork name to find its resource
* file at zip->tree_rsrc. */
archive_strcpy(&(zip_entry->rsrcname), "__MACOSX/");
archive_strncat(&(zip_entry->rsrcname), name, r - name);
archive_strcat(&(zip_entry->rsrcname), "._");
archive_strncat(&(zip_entry->rsrcname),
name + (r - name), filename_length - (r - name));
/* Register an entry to RB tree to sort it by
* file offset. */
__archive_rb_tree_insert_node(&zip->tree,
&zip_entry->node);
}
/* We don't read the filename until we get to the
local file header. Reading it here would speed up
table-of-contents operations (removing the need to
find and read local file header to get the
filename) at the cost of requiring a lot of extra
space. */
/* We don't read the extra block here. We assume it
will be duplicated at the local file header. */
__archive_read_consume(a,
46 + filename_length + extra_length + comment_length);
}