Import libarchive-2.5.1b.

This commit is contained in:
joerg 2008-04-25 15:07:17 +00:00
parent 5548060f79
commit 65a789b8e7
33 changed files with 1515 additions and 610 deletions

View file

@ -1,4 +1,6 @@
Mar 30, 2008: libarchive 2.5.1b released
Mar 15, 2008: libarchive 2.5.0b released
Mar 15, 2008: bsdcpio now seems to correctly write hardlinks into newc,
ustar, and old cpio archives. Just a little more testing before

View file

@ -1,6 +1,6 @@
#! /bin/sh
# Guess values for system-dependent variables and create Makefiles.
# Generated by GNU Autoconf 2.61 for libarchive 2.5.0b.
# Generated by GNU Autoconf 2.61 for libarchive 2.5.1b.
#
# Report bugs to <kientzle@freebsd.org>.
#
@ -728,8 +728,8 @@ SHELL=${CONFIG_SHELL-/bin/sh}
# Identity of this package.
PACKAGE_NAME='libarchive'
PACKAGE_TARNAME='libarchive'
PACKAGE_VERSION='2.5.0b'
PACKAGE_STRING='libarchive 2.5.0b'
PACKAGE_VERSION='2.5.1b'
PACKAGE_STRING='libarchive 2.5.1b'
PACKAGE_BUGREPORT='kientzle@freebsd.org'
ac_unique_file="libarchive"
@ -1408,7 +1408,7 @@ if test "$ac_init_help" = "long"; then
# Omit some internal or obsolete options to make the list less imposing.
# This message is too long to be a string in the A/UX 3.1 sh.
cat <<_ACEOF
\`configure' configures libarchive 2.5.0b to adapt to many kinds of systems.
\`configure' configures libarchive 2.5.1b to adapt to many kinds of systems.
Usage: $0 [OPTION]... [VAR=VALUE]...
@ -1478,7 +1478,7 @@ fi
if test -n "$ac_init_help"; then
case $ac_init_help in
short | recursive ) echo "Configuration of libarchive 2.5.0b:";;
short | recursive ) echo "Configuration of libarchive 2.5.1b:";;
esac
cat <<\_ACEOF
@ -1596,7 +1596,7 @@ fi
test -n "$ac_init_help" && exit $ac_status
if $ac_init_version; then
cat <<\_ACEOF
libarchive configure 2.5.0b
libarchive configure 2.5.1b
generated by GNU Autoconf 2.61
Copyright (C) 1992, 1993, 1994, 1995, 1996, 1998, 1999, 2000, 2001,
@ -1610,7 +1610,7 @@ cat >config.log <<_ACEOF
This file contains any messages produced by compilers while
running configure, to aid debugging if configure makes a mistake.
It was created by libarchive $as_me 2.5.0b, which was
It was created by libarchive $as_me 2.5.1b, which was
generated by GNU Autoconf 2.61. Invocation command line was
$ $0 $@
@ -2304,7 +2304,7 @@ fi
# Define the identity of the package.
PACKAGE='libarchive'
VERSION='2.5.0b'
VERSION='2.5.1b'
cat >>confdefs.h <<_ACEOF
@ -2455,21 +2455,21 @@ am__tar='${AMTAR} chof - "$$tardir"'; am__untar='${AMTAR} xf -'
# platforms. At least on FreeBSD, libtool uses an overly complex
# convention that attempts to solve problems that most people just
# don't have and which just causes confusion for most end users.
ARCHIVE_MAJOR=$(( 2005000 / 1000000 ))
ARCHIVE_MINOR=$(( (2005000 / 1000) % 1000 ))
ARCHIVE_REVISION=$(( 2005000 % 1000 ))
ARCHIVE_MAJOR=$(( 2005001 / 1000000 ))
ARCHIVE_MINOR=$(( (2005001 / 1000) % 1000 ))
ARCHIVE_REVISION=$(( 2005001 % 1000 ))
ARCHIVE_LIBTOOL_MAJOR=`echo $((${ARCHIVE_MAJOR} + ${ARCHIVE_MINOR}))`
ARCHIVE_LIBTOOL_VERSION=$ARCHIVE_LIBTOOL_MAJOR:$ARCHIVE_REVISION:$ARCHIVE_MINOR
# Stick the version numbers into config.h
cat >>confdefs.h <<\_ACEOF
#define LIBARCHIVE_VERSION_STRING "2.5.0b"
#define LIBARCHIVE_VERSION_STRING "2.5.1b"
_ACEOF
cat >>confdefs.h <<_ACEOF
#define LIBARCHIVE_VERSION_NUMBER "2005000"
#define LIBARCHIVE_VERSION_NUMBER "2005001"
_ACEOF
@ -2479,7 +2479,7 @@ _ACEOF
cat >>confdefs.h <<\_ACEOF
#define BSDTAR_VERSION_STRING "2.5.0b"
#define BSDTAR_VERSION_STRING "2.5.1b"
_ACEOF
@ -2487,9 +2487,9 @@ _ACEOF
# below, but the shell variable names apparently cannot be the same as
# the m4 macro names above. Why? Ask autoconf.
BSDCPIO_VERSION_STRING=0.9.9a
BSDTAR_VERSION_STRING=2.5.0b
LIBARCHIVE_VERSION_STRING=2.5.0b
LIBARCHIVE_VERSION_NUMBER=2005000
BSDTAR_VERSION_STRING=2.5.1b
LIBARCHIVE_VERSION_STRING=2.5.1b
LIBARCHIVE_VERSION_NUMBER=2005001
# Substitute the above version numbers into the various files below.
# Yes, I believe this is the fourth time we define what are essentially
@ -27783,7 +27783,7 @@ exec 6>&1
# report actual input values of CONFIG_FILES etc. instead of their
# values after options handling.
ac_log="
This file was extended by libarchive $as_me 2.5.0b, which was
This file was extended by libarchive $as_me 2.5.1b, which was
generated by GNU Autoconf 2.61. Invocation command line was
CONFIG_FILES = $CONFIG_FILES
@ -27836,7 +27836,7 @@ Report bugs to <bug-autoconf@gnu.org>."
_ACEOF
cat >>$CONFIG_STATUS <<_ACEOF
ac_cs_version="\\
libarchive config.status 2.5.0b
libarchive config.status 2.5.1b
configured by $0, generated by GNU Autoconf 2.61,
with options \\"`echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`\\"

View file

@ -4,8 +4,8 @@ dnl First, define all of the version numbers up front.
dnl In particular, this allows the version macro to be used in AC_INIT
dnl These first two version numbers are updated automatically on each release.
m4_define([LIBARCHIVE_VERSION_S],[2.5.0b])
m4_define([LIBARCHIVE_VERSION_N],[2005000])
m4_define([LIBARCHIVE_VERSION_S],[2.5.1b])
m4_define([LIBARCHIVE_VERSION_N],[2005001])
dnl bsdtar versioning tracks libarchive
m4_define([BSDTAR_VERSION_S],LIBARCHIVE_VERSION_S())

View file

@ -112,7 +112,7 @@ cpio_getopt(struct cpio *cpio)
opt = getopt_long(cpio->argc, cpio->argv, cpio_opts,
cpio_longopts, &option_index);
#else
opt = getopt(cpio->argc, cpio->argv, optstring);
opt = getopt(cpio->argc, cpio->argv, cpio_opts);
#endif
/* Support long options through -W longopt=value */

View file

@ -478,7 +478,8 @@ file_to_archive(struct cpio *cpio, const char *srcpath)
static int
entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
{
const char *destpath, *srcpath;
const char *destpath;
char *srcpath;
int fd = -1;
ssize_t bytes_read;
size_t len;
@ -488,7 +489,12 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
/*
* Generate a target path for this entry.
*/
destpath = srcpath = archive_entry_pathname(entry);
p = archive_entry_pathname(entry);
srcpath = malloc(strlen(p) + 1);
if (srcpath == NULL)
cpio_errc(1, ENOMEM, "Can't allocate path buffer");
strcpy(srcpath, p);
destpath = srcpath;
if (cpio->destdir) {
len = strlen(cpio->destdir) + strlen(srcpath) + 8;
if (len >= cpio->pass_destpath_alloc) {
@ -511,8 +517,10 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
}
if (cpio->option_rename)
destpath = cpio_rename(destpath);
if (destpath == NULL)
if (destpath == NULL) {
free(srcpath);
return (0);
}
archive_entry_copy_pathname(entry, destpath);
/* Print out the destination name to the user. */
@ -526,12 +534,12 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
/* Note: link(2) doesn't create parent directories. */
archive_entry_set_hardlink(entry, srcpath);
r = archive_write_header(cpio->archive, entry);
if (r == ARCHIVE_OK)
return (0);
cpio_warnc(archive_errno(cpio->archive),
archive_error_string(cpio->archive));
if (r != ARCHIVE_OK)
cpio_warnc(archive_errno(cpio->archive),
archive_error_string(cpio->archive));
if (r == ARCHIVE_FATAL)
exit(1);
free(srcpath);
return (0);
}
@ -563,8 +571,7 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
if (r == ARCHIVE_FATAL)
exit(1);
if (r >= ARCHIVE_WARN && fd >= 0 && archive_entry_size(entry) > 0) {
fd = open(srcpath, O_RDONLY);
if (r >= ARCHIVE_WARN && fd >= 0) {
bytes_read = read(fd, cpio->buff, cpio->buff_size);
while (bytes_read > 0) {
r = archive_write_data(cpio->archive,
@ -587,6 +594,7 @@ cleanup:
fprintf(stderr,"\n");
if (fd >= 0)
close(fd);
free(srcpath);
return (0);
}

View file

@ -39,7 +39,7 @@
*
* The next few lines are the only differences.
*/
#define PROGRAM "bsdcpio" /* Program being tested. */
#define PROGRAM "bsdcpio" /* Name of program being tested. */
#define ENVBASE "BSDCPIO" /* Prefix for environment variables. */
#undef EXTRA_DUMP /* How to dump extra data */
/* How to generate extra version info. */
@ -55,7 +55,7 @@ __FBSDID("$FreeBSD$");
* test functions.
*/
#undef DEFINE_TEST
#define DEFINE_TEST(name) void name(void);
#define DEFINE_TEST(name) void name(void);
#include "list.h"
/* Interix doesn't define these in a standard header. */
@ -272,6 +272,30 @@ test_assert_equal_int(const char *file, int line,
return (0);
}
static void strdump(const char *p)
{
if (p == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*p != '\0') {
unsigned int c = 0xff & *p++;
switch (c) {
case '\a': fprintf(stderr, "\a"); break;
case '\b': fprintf(stderr, "\b"); break;
case '\n': fprintf(stderr, "\n"); break;
case '\r': fprintf(stderr, "\r"); break;
default:
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else
fprintf(stderr, "\\x%02X", c);
}
}
fprintf(stderr, "\"");
}
/* assertEqualString() displays the values of the two strings. */
int
test_assert_equal_string(const char *file, int line,
@ -294,12 +318,37 @@ test_assert_equal_string(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Strings not equal\n",
file, line);
fprintf(stderr, " %s = \"%s\"\n", e1, v1);
fprintf(stderr, " %s = \"%s\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
strdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
strdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}
static void wcsdump(const wchar_t *w)
{
if (w == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*w != L'\0') {
unsigned int c = *w++;
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else if (c < 256)
fprintf(stderr, "\\x%02X", c);
else if (c < 0x10000)
fprintf(stderr, "\\u%04X", c);
else
fprintf(stderr, "\\U%08X", c);
}
fprintf(stderr, "\"");
}
/* assertEqualWString() displays the values of the two strings. */
int
test_assert_equal_wstring(const char *file, int line,
@ -308,7 +357,17 @@ test_assert_equal_wstring(const char *file, int line,
void *extra)
{
++assertions;
if (wcscmp(v1, v2) == 0) {
if (v1 == NULL) {
if (v2 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (v2 == NULL) {
if (v1 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (wcscmp(v1, v2) == 0) {
msg[0] = '\0';
return (1);
}
@ -317,8 +376,12 @@ test_assert_equal_wstring(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Unicode strings not equal\n",
file, line);
fwprintf(stderr, L" %s = \"%ls\"\n", e1, v1);
fwprintf(stderr, L" %s = \"%ls\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
wcsdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
wcsdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}
@ -587,7 +650,7 @@ slurpfile(size_t * sizep, const char *fmt, ...)
* We reuse it here to define a list of all tests (functions and names).
*/
#undef DEFINE_TEST
#define DEFINE_TEST(n) { n, #n },
#define DEFINE_TEST(n) { n, #n },
struct { void (*func)(void); const char *name; } tests[] = {
#include "list.h"
};
@ -662,7 +725,7 @@ static void usage(const char *program)
exit(1);
}
#define uudecode(c) (((c) - 0x20) & 0x3f)
#define UUDECODE(c) (((c) - 0x20) & 0x3f)
void
extract_reference_file(const char *name)
@ -695,23 +758,23 @@ extract_reference_file(const char *name)
if (memcmp(buff, "end", 3) == 0)
break;
bytes = uudecode(*p++);
bytes = UUDECODE(*p++);
while (bytes > 0) {
int n = 0;
/* Write out 1-3 bytes from that. */
if (bytes > 0) {
n = uudecode(*p++) << 18;
n |= uudecode(*p++) << 12;
n = UUDECODE(*p++) << 18;
n |= UUDECODE(*p++) << 12;
fputc(n >> 16, out);
--bytes;
}
if (bytes > 0) {
n |= uudecode(*p++) << 6;
n |= UUDECODE(*p++) << 6;
fputc((n >> 8) & 0xFF, out);
--bytes;
}
if (bytes > 0) {
n |= uudecode(*p++);
n |= UUDECODE(*p++);
fputc(n & 0xFF, out);
--bytes;
}

View file

@ -54,28 +54,30 @@ DEFINE_TEST(test_option_L)
r = systemf("cat filelist | %s -pd -L --quiet copy-L >copy-L.out 2>copy-L.err", testprog);
assertEqualInt(r, 0);
assertEmptyFile("copy-L.out");
assertEmptyFile("copy-L.err");
assertEqualInt(0, lstat("copy-L/symlink", &st));
failure("-pL should dereference symlinks and turn them into files.");
failure("-pdL should dereference symlinks and turn them into files.");
assert(!S_ISLNK(st.st_mode));
r = systemf("cat filelist | %s -o --quiet >archive.out 2>archive.err", testprog);
failure("Error invoking %s -o --quiet %s", testprog);
failure("Error invoking %s -o --quiet", testprog);
assertEqualInt(r, 0);
assertEqualInt(0, mkdir("unpack", 0755));
r = systemf("cat archive.out | (cd unpack ; %s -i >unpack.out 2>unpack.err)", testprog);
failure("Error invoking %s -i %s", testprog);
failure("Error invoking %s -i", testprog);
assertEqualInt(r, 0);
assertEqualInt(0, lstat("unpack/symlink", &st));
assert(S_ISLNK(st.st_mode));
r = systemf("cat filelist | %s -oL --quiet >archive-L.out 2>archive-L.err", testprog);
failure("Error invoking %s -oL --quiet %s", testprog);
failure("Error invoking %s -oL --quiet", testprog);
assertEqualInt(r, 0);
assertEqualInt(0, mkdir("unpack-L", 0755));
r = systemf("cat archive-L.out | (cd unpack-L ; %s -i >unpack-L.out 2>unpack-L.err)", testprog);
failure("Error invoking %s -i %s < archive-L.out", testprog);
failure("Error invoking %s -i < archive-L.out", testprog);
assertEqualInt(r, 0);
assertEqualInt(0, lstat("unpack-L/symlink", &st));
assert(!S_ISLNK(st.st_mode));

View file

@ -56,6 +56,8 @@ DEFINE_TEST(test_option_a)
{
struct stat st;
int r;
int f;
char buff[64];
/* Sanity check; verify that test_create really works. */
test_create("f0");
@ -63,24 +65,46 @@ DEFINE_TEST(test_option_a)
failure("test_create function is supposed to create a file with atime == 1; if this doesn't work, test_option_a is entirely invalid.");
assertEqualInt(st.st_atime, 1);
/* Copy the file without -a; should change the atime. */
test_create("f1");
r = systemf("echo f1 | %s -pd --quiet copy-no-a > copy-no-a.out 2>copy-no-a.err", testprog);
assertEqualInt(r, 0);
assertEmptyFile("copy-no-a.err");
assertEmptyFile("copy-no-a.out");
assertEqualInt(0, stat("f1", &st));
failure("Copying file without -a should have changed atime. Ignore this if your system does not record atimes.");
assert(st.st_atime != 1);
/* Sanity check; verify that atimes really do get modified. */
f = open("f0", O_RDONLY);
assertEqualInt(1, read(f,buff, 1));
assertEqualInt(0, close(f));
assertEqualInt(0, stat("f0", &st));
if (st.st_atime == 1) {
skipping("Cannot verify -a option\n"
" Your system appears to not support atime.");
}
else
{
/*
* If this disk is mounted noatime, then we can't
* verify correct operation without -a.
*/
/* Archive the file without -a; should change the atime. */
test_create("f2");
r = systemf("echo f2 | %s -o --quiet > archive-no-a.out 2>archive-no-a.err", testprog);
assertEqualInt(r, 0);
assertEmptyFile("copy-no-a.err");
assertEqualInt(0, stat("f2", &st));
failure("Archiving file without -a should have changed atime. Ignore this if your system does not record atimes.");
assert(st.st_atime != 1);
/* Copy the file without -a; should change the atime. */
test_create("f1");
r = systemf("echo f1 | %s -pd --quiet copy-no-a > copy-no-a.out 2>copy-no-a.err", testprog);
assertEqualInt(r, 0);
assertEmptyFile("copy-no-a.err");
assertEmptyFile("copy-no-a.out");
assertEqualInt(0, stat("f1", &st));
failure("Copying file without -a should have changed atime. Ignore this if your system does not record atimes.");
assert(st.st_atime != 1);
/* Archive the file without -a; should change the atime. */
test_create("f2");
r = systemf("echo f2 | %s -o --quiet > archive-no-a.out 2>archive-no-a.err", testprog);
assertEqualInt(r, 0);
assertEmptyFile("copy-no-a.err");
assertEqualInt(0, stat("f2", &st));
failure("Archiving file without -a should have changed atime. Ignore this if your system does not record atimes.");
assert(st.st_atime != 1);
}
/*
* We can, of course, still verify that the atime is unchanged
* when using the -a option.
*/
/* Copy the file with -a; should not change the atime. */
test_create("f3");

View file

@ -137,7 +137,7 @@ Each of the two 16 bit values are stored in machine-native byte order.
.TP
\fInamesize\fP
The number of bytes in the pathname that follows the header.
This count includes the trailing NULL byte.
This count includes the trailing NUL byte.
.TP
\fIfilesize\fP
The size of the file.
@ -149,8 +149,8 @@ above for a description of the storage of four-byte integers.
The pathname immediately follows the fixed header.
If the
\fBnamesize\fP
is odd, an additional NULL byte is added after the pathname.
The file data is then appended, padded with NULL
is odd, an additional NUL byte is added after the pathname.
The file data is then appended, padded with NUL
bytes to an even length.
Hardlinked files are not given special treatment;
the full file contents are included with each copy of the
@ -187,7 +187,7 @@ Unlike the old binary format, there is no additional padding
after the pathname or file contents.
If the files being archived are themselves entirely ASCII, then
the resulting archive will be entirely ASCII, except for the
NULL byte that terminates the name field.
NUL byte that terminates the name field.
.SS New ASCII Format
The "new" ASCII format uses 8-byte hexadecimal fields for
all numbers and separates device numbers into separate fields
@ -220,7 +220,7 @@ The string
\fIcheck\fP
This field is always set to zero by writers and ignored by readers.
See the next section for more details.
The pathname is followed by NULL bytes so that the total size
The pathname is followed by NUL bytes so that the total size
of the fixed header plus pathname is a multiple of four.
Likewise, the file data is padded to a multiple of four bytes.
Note that this format supports only 4 gigabyte files (unlike the

View file

@ -193,7 +193,7 @@ field with several new type values:
.TP
``0''
Regular file.
NULL should be treated as a synonym, for compatibility purposes.
NUL should be treated as a synonym, for compatibility purposes.
.TP
``1''
Hard link.
@ -238,7 +238,7 @@ by readers.
\fImagic\fP
Contains the magic value
``ustar''
followed by a NULL byte to indicate that this is a POSIX standard archive.
followed by a NUL byte to indicate that this is a POSIX standard archive.
Full compliance requires the uname and gname fields be properly set.
.TP
\fIversion\fP
@ -267,7 +267,7 @@ the prefix value and a
\fI/\fP
character to the regular name field to obtain the full pathname.
Note that all unused bytes must be set to
.BR NULL.
.BR NUL.
Field termination is specified slightly differently by POSIX
than by previous implementations.
The
@ -276,14 +276,14 @@ The
and
\fIgname\fP
fields must have a trailing
.BR NULL.
.BR NUL.
The
\fIpathname\fP,
\fIlinkname\fP,
and
\fIprefix\fP
fields must have a trailing
.BR NULL
.BR NUL
unless they fill the entire field.
(In particular, it is possible to store a 256-character pathname if it
happens to have a
@ -291,7 +291,7 @@ happens to have a
as the 156th character.)
POSIX requires numeric fields to be zero-padded in the front, and allows
them to be terminated with either space or
.BR NULL
.BR NUL
characters.
Currently, most tar implementations comply with the ustar
format, occasionally extending it by adding new fields to the

View file

@ -91,7 +91,7 @@ DESCRIPTION
namesize
The number of bytes in the pathname that follows the header.
This count includes the trailing NULL byte.
This count includes the trailing NUL byte.
filesize
The size of the file. Note that this archive format is limited
@ -99,8 +99,8 @@ DESCRIPTION
of the storage of four-byte integers.
The pathname immediately follows the fixed header. If the namesize is
odd, an additional NULL byte is added after the pathname. The file data
is then appended, padded with NULL bytes to an even length.
odd, an additional NUL byte is added after the pathname. The file data
is then appended, padded with NUL bytes to an even length.
Hardlinked files are not given special treatment; the full file contents
are included with each copy of the file.
@ -130,7 +130,7 @@ DESCRIPTION
file body follow the fixed header. Unlike the old binary format, there
is no additional padding after the pathname or file contents. If the
files being archived are themselves entirely ASCII, then the resulting
archive will be entirely ASCII, except for the NULL byte that terminates
archive will be entirely ASCII, except for the NUL byte that terminates
the name field.
New ASCII Format
@ -163,11 +163,11 @@ DESCRIPTION
check This field is always set to zero by writers and ignored by read-
ers. See the next section for more details.
The pathname is followed by NULL bytes so that the total size of the
fixed header plus pathname is a multiple of four. Likewise, the file
data is padded to a multiple of four bytes. Note that this format sup-
ports only 4 gigabyte files (unlike the older ASCII format, which sup-
ports 8 gigabyte files).
The pathname is followed by NUL bytes so that the total size of the fixed
header plus pathname is a multiple of four. Likewise, the file data is
padded to a multiple of four bytes. Note that this format supports only
4 gigabyte files (unlike the older ASCII format, which supports 8 giga-
byte files).
In this format, hardlinked files are handled by setting the filesize to
zero for each entry except the last one that appears in the archive.

View file

@ -143,7 +143,7 @@ DESCRIPTION
typeflag
Type of entry. POSIX extended the earlier linkflag field with
several new type values:
``0'' Regular file. NULL should be treated as a synonym, for
``0'' Regular file. NUL should be treated as a synonym, for
compatibility purposes.
``1'' Hard link.
``2'' Symbolic link.
@ -167,7 +167,7 @@ DESCRIPTION
allocate directory space. For all other types, it should be set
to zero by writers and ignored by readers.
magic Contains the magic value ``ustar'' followed by a NULL byte to
magic Contains the magic value ``ustar'' followed by a NUL byte to
indicate that this is a POSIX standard archive. Full compliance
requires the uname and gname fields be properly set.
@ -191,15 +191,15 @@ DESCRIPTION
a / character to the regular name field to obtain the full path-
name.
Note that all unused bytes must be set to NULL.
Note that all unused bytes must be set to NUL.
Field termination is specified slightly differently by POSIX than by pre-
vious implementations. The magic, uname, and gname fields must have a
trailing NULL. The pathname, linkname, and prefix fields must have a
trailing NULL unless they fill the entire field. (In particular, it is
trailing NUL. The pathname, linkname, and prefix fields must have a
trailing NUL unless they fill the entire field. (In particular, it is
possible to store a 256-character pathname if it happens to have a / as
the 156th character.) POSIX requires numeric fields to be zero-padded in
the front, and allows them to be terminated with either space or NULL
the front, and allows them to be terminated with either space or NUL
characters.
Currently, most tar implementations comply with the ustar format, occa-

View file

@ -29,17 +29,48 @@
#define ARCHIVE_H_INCLUDED
#include <sys/types.h> /* Linux requires this for off_t */
@ARCHIVE_H_INCLUDE_INTTYPES_H@
/* TODO: Conditionalize this include on platforms that don't support it. */
#include <inttypes.h> /* int64_t, etc. */
#include <stdio.h> /* For FILE * */
#ifndef _WIN32
#include <unistd.h> /* For ssize_t and size_t */
/* Get appropriate definitions of standard POSIX-style types. */
/* These should match the types used in 'struct stat' */
#ifdef _WIN32
#define __LA_SSIZE_T long
#define __LA_UID_T unsigned int
#define __LA_GID_T unsigned int
#else
typedef long ssize_t;
typedef unsigned int uid_t;
typedef unsigned int gid_t;
typedef unsigned short mode_t;
#include <unistd.h> /* ssize_t, uid_t, and gid_t */
#define __LA_SSIZE_T ssize_t
#define __LA_UID_T uid_t
#define __LA_GID_T gid_t
#endif
/*
* On Windows, define LIBARCHIVE_STATIC if you're building or using a
* .lib. The default here assumes you're building a DLL. Only
* libarchive source should ever define __LIBARCHIVE_BUILD.
*/
#if ((defined __WIN32__) || (defined _WIN32)) && (!defined LIBARCHIVE_STATIC)
# ifdef __LIBARCHIVE_BUILD
# ifdef __GNUC__
# define __LA_DECL __attribute__((dllexport)) extern
# else
# define __LA_DECL __declspec(dllexport)
# endif
# else
# ifdef __GNUC__
# define __LA_DECL __attribute__((dllimport)) extern
# else
# define __LA_DECL __declspec(dllimport)
# endif
# endif
#else
/* Static libraries or non-Windows needs no special declaration. */
# define __LA_DECL
#endif
#ifdef __cplusplus
extern "C" {
#endif
@ -48,7 +79,9 @@ extern "C" {
* The version number is provided as both a macro and a function.
* The macro identifies the installed header; the function identifies
* the library version (which may not be the same if you're using a
* dynamically-linked version of the library).
* dynamically-linked version of the library). Of course, if the
* header and library are very different, you should expect some
* strangeness. Don't do that.
*/
/*
@ -70,13 +103,13 @@ extern "C" {
* #endif
*/
#define ARCHIVE_VERSION_NUMBER @LIBARCHIVE_VERSION_NUMBER@
int archive_version_number(void);
__LA_DECL int archive_version_number(void);
/*
* Textual name/version of the library, useful for version displays.
*/
#define ARCHIVE_VERSION_STRING "libarchive @LIBARCHIVE_VERSION_STRING@"
const char * archive_version_string(void);
__LA_DECL const char * archive_version_string(void);
#if ARCHIVE_VERSION_NUMBER < 3000000
/*
@ -84,13 +117,13 @@ const char * archive_version_string(void);
* the simpler definitions above.
*/
#define ARCHIVE_VERSION_STAMP ARCHIVE_VERSION_NUMBER
int archive_version_stamp(void);
__LA_DECL int archive_version_stamp(void);
#define ARCHIVE_LIBRARY_VERSION ARCHIVE_VERSION_STRING
const char * archive_version(void);
__LA_DECL const char * archive_version(void);
#define ARCHIVE_API_VERSION (ARCHIVE_VERSION_NUMBER / 1000000)
int archive_api_version(void);
__LA_DECL int archive_api_version(void);
#define ARCHIVE_API_FEATURE ((ARCHIVE_VERSION_NUMBER / 1000) % 1000)
int archive_api_feature(void);
__LA_DECL int archive_api_feature(void);
#endif
#if ARCHIVE_VERSION_NUMBER < 3000000
@ -140,18 +173,18 @@ struct archive_entry;
*/
/* Returns pointer and size of next block of data from archive. */
typedef ssize_t archive_read_callback(struct archive *, void *_client_data,
typedef __LA_SSIZE_T archive_read_callback(struct archive *, void *_client_data,
const void **_buffer);
/* Skips at most request bytes from archive and returns the skipped amount */
#if ARCHIVE_VERSION_NUMBER < 2000000
typedef ssize_t archive_skip_callback(struct archive *, void *_client_data,
typedef __LA_SSIZE_T archive_skip_callback(struct archive *, void *_client_data,
size_t request);
#else
typedef off_t archive_skip_callback(struct archive *, void *_client_data,
off_t request);
#endif
/* Returns size actually written, zero on EOF, -1 on error. */
typedef ssize_t archive_write_callback(struct archive *, void *_client_data,
typedef __LA_SSIZE_T archive_write_callback(struct archive *, void *_client_data,
const void *_buffer, size_t _length);
typedef int archive_open_callback(struct archive *, void *_client_data);
typedef int archive_close_callback(struct archive *, void *_client_data);
@ -219,7 +252,7 @@ typedef int archive_close_callback(struct archive *, void *_client_data);
* data for entries of interest.
* 5) Call archive_read_finish to end processing.
*/
struct archive *archive_read_new(void);
__LA_DECL struct archive *archive_read_new(void);
/*
* The archive_read_support_XXX calls enable auto-detect for this
@ -228,30 +261,30 @@ struct archive *archive_read_new(void);
* support_compression_bzip2(). The "all" functions provide the
* obvious shorthand.
*/
int archive_read_support_compression_all(struct archive *);
int archive_read_support_compression_bzip2(struct archive *);
int archive_read_support_compression_compress(struct archive *);
int archive_read_support_compression_gzip(struct archive *);
int archive_read_support_compression_none(struct archive *);
int archive_read_support_compression_program(struct archive *,
__LA_DECL int archive_read_support_compression_all(struct archive *);
__LA_DECL int archive_read_support_compression_bzip2(struct archive *);
__LA_DECL int archive_read_support_compression_compress(struct archive *);
__LA_DECL int archive_read_support_compression_gzip(struct archive *);
__LA_DECL int archive_read_support_compression_none(struct archive *);
__LA_DECL int archive_read_support_compression_program(struct archive *,
const char *command);
int archive_read_support_format_all(struct archive *);
int archive_read_support_format_ar(struct archive *);
int archive_read_support_format_cpio(struct archive *);
int archive_read_support_format_empty(struct archive *);
int archive_read_support_format_gnutar(struct archive *);
int archive_read_support_format_iso9660(struct archive *);
int archive_read_support_format_mtree(struct archive *);
int archive_read_support_format_tar(struct archive *);
int archive_read_support_format_zip(struct archive *);
__LA_DECL int archive_read_support_format_all(struct archive *);
__LA_DECL int archive_read_support_format_ar(struct archive *);
__LA_DECL int archive_read_support_format_cpio(struct archive *);
__LA_DECL int archive_read_support_format_empty(struct archive *);
__LA_DECL int archive_read_support_format_gnutar(struct archive *);
__LA_DECL int archive_read_support_format_iso9660(struct archive *);
__LA_DECL int archive_read_support_format_mtree(struct archive *);
__LA_DECL int archive_read_support_format_tar(struct archive *);
__LA_DECL int archive_read_support_format_zip(struct archive *);
/* Open the archive using callbacks for archive I/O. */
int archive_read_open(struct archive *, void *_client_data,
__LA_DECL int archive_read_open(struct archive *, void *_client_data,
archive_open_callback *, archive_read_callback *,
archive_close_callback *);
int archive_read_open2(struct archive *, void *_client_data,
__LA_DECL int archive_read_open2(struct archive *, void *_client_data,
archive_open_callback *, archive_read_callback *,
archive_skip_callback *, archive_close_callback *);
@ -261,43 +294,43 @@ int archive_read_open2(struct archive *, void *_client_data,
* accept a block size handle tape blocking correctly.
*/
/* Use this if you know the filename. Note: NULL indicates stdin. */
int archive_read_open_filename(struct archive *,
__LA_DECL int archive_read_open_filename(struct archive *,
const char *_filename, size_t _block_size);
/* archive_read_open_file() is a deprecated synonym for ..._open_filename(). */
int archive_read_open_file(struct archive *,
__LA_DECL int archive_read_open_file(struct archive *,
const char *_filename, size_t _block_size);
/* Read an archive that's stored in memory. */
int archive_read_open_memory(struct archive *,
__LA_DECL int archive_read_open_memory(struct archive *,
void * buff, size_t size);
/* A more involved version that is only used for internal testing. */
int archive_read_open_memory2(struct archive *a, void *buff,
__LA_DECL int archive_read_open_memory2(struct archive *a, void *buff,
size_t size, size_t read_size);
/* Read an archive that's already open, using the file descriptor. */
int archive_read_open_fd(struct archive *, int _fd,
__LA_DECL int archive_read_open_fd(struct archive *, int _fd,
size_t _block_size);
/* Read an archive that's already open, using a FILE *. */
/* Note: DO NOT use this with tape drives. */
int archive_read_open_FILE(struct archive *, FILE *_file);
__LA_DECL int archive_read_open_FILE(struct archive *, FILE *_file);
/* Parses and returns next entry header. */
int archive_read_next_header(struct archive *,
__LA_DECL int archive_read_next_header(struct archive *,
struct archive_entry **);
/*
* Retrieve the byte offset in UNCOMPRESSED data where last-read
* header started.
*/
int64_t archive_read_header_position(struct archive *);
__LA_DECL int64_t archive_read_header_position(struct archive *);
/* Read data from the body of an entry. Similar to read(2). */
ssize_t archive_read_data(struct archive *, void *, size_t);
__LA_DECL __LA_SSIZE_T archive_read_data(struct archive *, void *, size_t);
/*
* A zero-copy version of archive_read_data that also exposes the file offset
* of each returned block. Note that the client has no way to specify
* the desired size of the block. The API does guarantee that offsets will
* be strictly increasing and that returned blocks will not overlap.
*/
int archive_read_data_block(struct archive *a,
__LA_DECL int archive_read_data_block(struct archive *a,
const void **buff, size_t *size, off_t *offset);
/*-
@ -306,10 +339,10 @@ int archive_read_data_block(struct archive *a,
* 'into_buffer': writes data into memory buffer that you provide
* 'into_fd': writes data to specified filedes
*/
int archive_read_data_skip(struct archive *);
int archive_read_data_into_buffer(struct archive *, void *buffer,
ssize_t len);
int archive_read_data_into_fd(struct archive *, int fd);
__LA_DECL int archive_read_data_skip(struct archive *);
__LA_DECL int archive_read_data_into_buffer(struct archive *, void *buffer,
__LA_SSIZE_T len);
__LA_DECL int archive_read_data_into_fd(struct archive *, int fd);
/*-
* Convenience function to recreate the current entry (whose header
@ -353,26 +386,26 @@ int archive_read_data_into_fd(struct archive *, int fd);
/* Default: Overwrite files, even if one on disk is newer. */
#define ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER (2048)
int archive_read_extract(struct archive *, struct archive_entry *,
__LA_DECL int archive_read_extract(struct archive *, struct archive_entry *,
int flags);
void archive_read_extract_set_progress_callback(struct archive *,
__LA_DECL void archive_read_extract_set_progress_callback(struct archive *,
void (*_progress_func)(void *), void *_user_data);
/* Record the dev/ino of a file that will not be written. This is
* generally set to the dev/ino of the archive being read. */
void archive_read_extract_set_skip_file(struct archive *,
__LA_DECL void archive_read_extract_set_skip_file(struct archive *,
dev_t, ino_t);
/* Close the file and release most resources. */
int archive_read_close(struct archive *);
__LA_DECL int archive_read_close(struct archive *);
/* Release all resources and destroy the object. */
/* Note that archive_read_finish will call archive_read_close for you. */
#if ARCHIVE_VERSION_NUMBER >= 2000000
int archive_read_finish(struct archive *);
__LA_DECL int archive_read_finish(struct archive *);
#else
/* Temporarily allow library to compile with either 1.x or 2.0 API. */
/* Erroneously declared to return void in libarchive 1.x */
void archive_read_finish(struct archive *);
__LA_DECL void archive_read_finish(struct archive *);
#endif
/*-
@ -390,75 +423,75 @@ void archive_read_finish(struct archive *);
* 5) archive_write_close to close the output
* 6) archive_write_finish to cleanup the writer and release resources
*/
struct archive *archive_write_new(void);
int archive_write_set_bytes_per_block(struct archive *,
__LA_DECL struct archive *archive_write_new(void);
__LA_DECL int archive_write_set_bytes_per_block(struct archive *,
int bytes_per_block);
int archive_write_get_bytes_per_block(struct archive *);
__LA_DECL int archive_write_get_bytes_per_block(struct archive *);
/* XXX This is badly misnamed; suggestions appreciated. XXX */
int archive_write_set_bytes_in_last_block(struct archive *,
__LA_DECL int archive_write_set_bytes_in_last_block(struct archive *,
int bytes_in_last_block);
int archive_write_get_bytes_in_last_block(struct archive *);
__LA_DECL int archive_write_get_bytes_in_last_block(struct archive *);
/* The dev/ino of a file that won't be archived. This is used
* to avoid recursively adding an archive to itself. */
int archive_write_set_skip_file(struct archive *, dev_t, ino_t);
__LA_DECL int archive_write_set_skip_file(struct archive *, dev_t, ino_t);
int archive_write_set_compression_bzip2(struct archive *);
int archive_write_set_compression_compress(struct archive *);
int archive_write_set_compression_gzip(struct archive *);
int archive_write_set_compression_none(struct archive *);
int archive_write_set_compression_program(struct archive *,
__LA_DECL int archive_write_set_compression_bzip2(struct archive *);
__LA_DECL int archive_write_set_compression_compress(struct archive *);
__LA_DECL int archive_write_set_compression_gzip(struct archive *);
__LA_DECL int archive_write_set_compression_none(struct archive *);
__LA_DECL int archive_write_set_compression_program(struct archive *,
const char *cmd);
/* A convenience function to set the format based on the code or name. */
int archive_write_set_format(struct archive *, int format_code);
int archive_write_set_format_by_name(struct archive *,
__LA_DECL int archive_write_set_format(struct archive *, int format_code);
__LA_DECL int archive_write_set_format_by_name(struct archive *,
const char *name);
/* To minimize link pollution, use one or more of the following. */
int archive_write_set_format_ar_bsd(struct archive *);
int archive_write_set_format_ar_svr4(struct archive *);
int archive_write_set_format_cpio(struct archive *);
int archive_write_set_format_cpio_newc(struct archive *);
__LA_DECL int archive_write_set_format_ar_bsd(struct archive *);
__LA_DECL int archive_write_set_format_ar_svr4(struct archive *);
__LA_DECL int archive_write_set_format_cpio(struct archive *);
__LA_DECL int archive_write_set_format_cpio_newc(struct archive *);
/* TODO: int archive_write_set_format_old_tar(struct archive *); */
int archive_write_set_format_pax(struct archive *);
int archive_write_set_format_pax_restricted(struct archive *);
int archive_write_set_format_shar(struct archive *);
int archive_write_set_format_shar_dump(struct archive *);
int archive_write_set_format_ustar(struct archive *);
int archive_write_open(struct archive *, void *,
__LA_DECL int archive_write_set_format_pax(struct archive *);
__LA_DECL int archive_write_set_format_pax_restricted(struct archive *);
__LA_DECL int archive_write_set_format_shar(struct archive *);
__LA_DECL int archive_write_set_format_shar_dump(struct archive *);
__LA_DECL int archive_write_set_format_ustar(struct archive *);
__LA_DECL int archive_write_open(struct archive *, void *,
archive_open_callback *, archive_write_callback *,
archive_close_callback *);
int archive_write_open_fd(struct archive *, int _fd);
int archive_write_open_filename(struct archive *, const char *_file);
__LA_DECL int archive_write_open_fd(struct archive *, int _fd);
__LA_DECL int archive_write_open_filename(struct archive *, const char *_file);
/* A deprecated synonym for archive_write_open_filename() */
int archive_write_open_file(struct archive *, const char *_file);
int archive_write_open_FILE(struct archive *, FILE *);
__LA_DECL int archive_write_open_file(struct archive *, const char *_file);
__LA_DECL int archive_write_open_FILE(struct archive *, FILE *);
/* _buffSize is the size of the buffer, _used refers to a variable that
* will be updated after each write into the buffer. */
int archive_write_open_memory(struct archive *,
__LA_DECL int archive_write_open_memory(struct archive *,
void *_buffer, size_t _buffSize, size_t *_used);
/*
* Note that the library will truncate writes beyond the size provided
* to archive_write_header or pad if the provided data is short.
*/
int archive_write_header(struct archive *,
__LA_DECL int archive_write_header(struct archive *,
struct archive_entry *);
#if ARCHIVE_VERSION_NUMBER >= 2000000
ssize_t archive_write_data(struct archive *, const void *, size_t);
__LA_DECL __LA_SSIZE_T archive_write_data(struct archive *, const void *, size_t);
#else
/* Temporarily allow library to compile with either 1.x or 2.0 API. */
/* This was erroneously declared to return "int" in libarchive 1.x. */
int archive_write_data(struct archive *, const void *, size_t);
__LA_DECL int archive_write_data(struct archive *, const void *, size_t);
#endif
ssize_t archive_write_data_block(struct archive *, const void *, size_t, off_t);
int archive_write_finish_entry(struct archive *);
int archive_write_close(struct archive *);
__LA_DECL __LA_SSIZE_T archive_write_data_block(struct archive *, const void *, size_t, off_t);
__LA_DECL int archive_write_finish_entry(struct archive *);
__LA_DECL int archive_write_close(struct archive *);
#if ARCHIVE_VERSION_NUMBER >= 2000000
int archive_write_finish(struct archive *);
__LA_DECL int archive_write_finish(struct archive *);
#else
/* Temporarily allow library to compile with either 1.x or 2.0 API. */
/* Return value was incorrect in libarchive 1.x. */
void archive_write_finish(struct archive *);
__LA_DECL void archive_write_finish(struct archive *);
#endif
/*-
@ -475,12 +508,12 @@ void archive_write_finish(struct archive *);
* In particular, you can use this in conjunction with archive_read()
* to pull entries out of an archive and create them on disk.
*/
struct archive *archive_write_disk_new(void);
__LA_DECL struct archive *archive_write_disk_new(void);
/* This file will not be overwritten. */
int archive_write_disk_set_skip_file(struct archive *,
__LA_DECL int archive_write_disk_set_skip_file(struct archive *,
dev_t, ino_t);
/* Set flags to control how the next item gets created. */
int archive_write_disk_set_options(struct archive *,
__LA_DECL int archive_write_disk_set_options(struct archive *,
int flags);
/*
* The lookup functions are given uname/uid (or gname/gid) pairs and
@ -499,42 +532,47 @@ int archive_write_disk_set_options(struct archive *,
* particular, these match the specifications of POSIX "pax" and old
* POSIX "tar".
*/
int archive_write_disk_set_standard_lookup(struct archive *);
__LA_DECL int archive_write_disk_set_standard_lookup(struct archive *);
/*
* If neither the default (naive) nor the standard (big) functions suit
* your needs, you can write your own and register them. Be sure to
* include a cleanup function if you have allocated private data.
*/
int archive_write_disk_set_group_lookup(struct archive *,
void *private_data,
gid_t (*loookup)(void *, const char *gname, gid_t gid),
void (*cleanup)(void *));
int archive_write_disk_set_user_lookup(struct archive *,
void *private_data,
uid_t (*)(void *, const char *uname, uid_t uid),
void (*cleanup)(void *));
__LA_DECL int archive_write_disk_set_group_lookup(struct archive *,
void * /* private_data */,
__LA_GID_T (*)(void *, const char *, __LA_GID_T),
void (* /* cleanup */)(void *));
__LA_DECL int archive_write_disk_set_user_lookup(struct archive *,
void * /* private_data */,
__LA_UID_T (*)(void *, const char *, __LA_UID_T),
void (* /* cleanup */)(void *));
/*
* Accessor functions to read/set various information in
* the struct archive object:
*/
/* Bytes written after compression or read before decompression. */
int64_t archive_position_compressed(struct archive *);
__LA_DECL int64_t archive_position_compressed(struct archive *);
/* Bytes written to compressor or read from decompressor. */
int64_t archive_position_uncompressed(struct archive *);
__LA_DECL int64_t archive_position_uncompressed(struct archive *);
const char *archive_compression_name(struct archive *);
int archive_compression(struct archive *);
int archive_errno(struct archive *);
const char *archive_error_string(struct archive *);
const char *archive_format_name(struct archive *);
int archive_format(struct archive *);
void archive_clear_error(struct archive *);
void archive_set_error(struct archive *, int _err, const char *fmt, ...);
void archive_copy_error(struct archive *dest, struct archive *src);
__LA_DECL const char *archive_compression_name(struct archive *);
__LA_DECL int archive_compression(struct archive *);
__LA_DECL int archive_errno(struct archive *);
__LA_DECL const char *archive_error_string(struct archive *);
__LA_DECL const char *archive_format_name(struct archive *);
__LA_DECL int archive_format(struct archive *);
__LA_DECL void archive_clear_error(struct archive *);
__LA_DECL void archive_set_error(struct archive *, int _err,
const char *fmt, ...);
__LA_DECL void archive_copy_error(struct archive *dest,
struct archive *src);
#ifdef __cplusplus
}
#endif
/* This is meaningless outside of this header. */
#undef __LA_DECL
#endif /* !ARCHIVE_H_INCLUDED */

View file

@ -91,15 +91,17 @@ static void aes_clean(struct aes *);
static void aes_copy(struct aes *dest, struct aes *src);
static const char * aes_get_mbs(struct aes *);
static const wchar_t * aes_get_wcs(struct aes *);
static void aes_set_mbs(struct aes *, const char *mbs);
static void aes_copy_mbs(struct aes *, const char *mbs);
static int aes_set_mbs(struct aes *, const char *mbs);
static int aes_copy_mbs(struct aes *, const char *mbs);
/* static void aes_set_wcs(struct aes *, const wchar_t *wcs); */
static void aes_copy_wcs(struct aes *, const wchar_t *wcs);
static void aes_copy_wcs_len(struct aes *, const wchar_t *wcs, size_t);
static int aes_copy_wcs(struct aes *, const wchar_t *wcs);
static int aes_copy_wcs_len(struct aes *, const wchar_t *wcs, size_t);
static char * ae_fflagstostr(unsigned long bitset, unsigned long bitclear);
static const wchar_t *ae_wcstofflags(const wchar_t *stringp,
unsigned long *setp, unsigned long *clrp);
static const char *ae_strtofflags(const char *stringp,
unsigned long *setp, unsigned long *clrp);
static void append_entry_w(wchar_t **wp, const wchar_t *prefix, int tag,
const wchar_t *wname, int perm, int id);
static void append_id_w(wchar_t **wp, int id);
@ -144,173 +146,187 @@ static size_t wcslen(const wchar_t *s)
#define wmemcpy(a,b,i) (wchar_t *)memcpy((a), (b), (i) * sizeof(wchar_t))
#endif
static void
aes_clean(struct aes *aes)
{
if (aes->aes_mbs_alloc) {
free(aes->aes_mbs_alloc);
aes->aes_mbs_alloc = NULL;
if (aes->aes_wcs) {
free((wchar_t *)(uintptr_t)aes->aes_wcs);
aes->aes_wcs = NULL;
}
if (aes->aes_wcs_alloc) {
free(aes->aes_wcs_alloc);
aes->aes_wcs_alloc = NULL;
}
memset(aes, 0, sizeof(*aes));
archive_string_free(&(aes->aes_mbs));
archive_string_free(&(aes->aes_utf8));
}
static void
aes_copy(struct aes *dest, struct aes *src)
{
*dest = *src;
if (src->aes_mbs != NULL) {
dest->aes_mbs_alloc = strdup(src->aes_mbs);
dest->aes_mbs = dest->aes_mbs_alloc;
if (dest->aes_mbs == NULL)
__archive_errx(1, "No memory for aes_copy()");
}
wchar_t *wp;
archive_string_copy(&(dest->aes_mbs), &(src->aes_mbs));
archive_string_copy(&(dest->aes_utf8), &(src->aes_utf8));
if (src->aes_wcs != NULL) {
dest->aes_wcs_alloc = (wchar_t *)malloc((wcslen(src->aes_wcs) + 1)
wp = (wchar_t *)malloc((wcslen(src->aes_wcs) + 1)
* sizeof(wchar_t));
dest->aes_wcs = dest->aes_wcs_alloc;
if (dest->aes_wcs == NULL)
if (wp == NULL)
__archive_errx(1, "No memory for aes_copy()");
wcscpy(dest->aes_wcs_alloc, src->aes_wcs);
wcscpy(wp, src->aes_wcs);
dest->aes_wcs = wp;
}
}
static const char *
aes_get_utf8(struct aes *aes)
{
if (aes->aes_utf8.s == NULL || aes->aes_utf8.length == 0) {
if (aes->aes_wcs == NULL)
return (NULL);
if (archive_strappend_w_utf8(&(aes->aes_utf8), aes->aes_wcs) == NULL)
return (NULL);
}
return (aes->aes_utf8.s);
}
static const char *
aes_get_mbs(struct aes *aes)
{
if (aes->aes_mbs == NULL && aes->aes_wcs == NULL)
return NULL;
if (aes->aes_mbs == NULL && aes->aes_wcs != NULL) {
/*
* XXX Need to estimate the number of byte in the
* multi-byte form. Assume that, on average, wcs
* chars encode to no more than 3 bytes. There must
* be a better way... XXX
*/
size_t mbs_length = wcslen(aes->aes_wcs) * 3 + 64;
aes->aes_mbs_alloc = (char *)malloc(mbs_length);
aes->aes_mbs = aes->aes_mbs_alloc;
if (aes->aes_mbs == NULL)
__archive_errx(1, "No memory for aes_get_mbs()");
wcstombs(aes->aes_mbs_alloc, aes->aes_wcs, mbs_length - 1);
aes->aes_mbs_alloc[mbs_length - 1] = 0;
}
return (aes->aes_mbs);
/* If we already have an MBS form, return that immediately. */
if (aes->aes_mbs.s != NULL && aes->aes_mbs.length != 0)
return (aes->aes_mbs.s);
/* If there's a WCS form, try converting with the native locale. */
if (aes->aes_wcs != NULL
&& archive_strappend_w_mbs(&(aes->aes_mbs), aes->aes_wcs) != NULL)
return (aes->aes_mbs.s);
/* We'll use UTF-8 for MBS if all else fails. */
return (aes_get_utf8(aes));
}
static const wchar_t *
aes_get_wcs(struct aes *aes)
{
wchar_t *w;
int r;
if (aes->aes_wcs == NULL && aes->aes_mbs == NULL)
return NULL;
if (aes->aes_wcs == NULL && aes->aes_mbs != NULL) {
/* Return WCS form if we already have it. */
if (aes->aes_wcs != NULL)
return (aes->aes_wcs);
if (aes->aes_mbs.s != NULL && aes->aes_mbs.length > 0) {
/* Try converting MBS to WCS using native locale. */
/*
* No single byte will be more than one wide character,
* so this length estimate will always be big enough.
*/
size_t wcs_length = strlen(aes->aes_mbs);
size_t wcs_length = aes->aes_mbs.length;
aes->aes_wcs_alloc
= (wchar_t *)malloc((wcs_length + 1) * sizeof(wchar_t));
aes->aes_wcs = aes->aes_wcs_alloc;
if (aes->aes_wcs == NULL)
w = (wchar_t *)malloc((wcs_length + 1) * sizeof(wchar_t));
if (w == NULL)
__archive_errx(1, "No memory for aes_get_wcs()");
r = mbstowcs(aes->aes_wcs_alloc, aes->aes_mbs, wcs_length);
aes->aes_wcs_alloc[wcs_length] = 0;
if (r == -1) {
/* Conversion failed, don't lie to our clients. */
free(aes->aes_wcs_alloc);
aes->aes_wcs = aes->aes_wcs_alloc = NULL;
}
r = mbstowcs(w, aes->aes_mbs.s, wcs_length);
w[wcs_length] = 0;
if (r > 0)
return (aes->aes_wcs = w);
free(w);
}
return (aes->aes_wcs);
if (aes->aes_utf8.s != NULL && aes->aes_utf8.length > 0) {
/* Try converting UTF8 to WCS. */
aes->aes_wcs = __archive_string_utf8_w(&(aes->aes_utf8));
return (aes->aes_wcs);
}
return (NULL);
}
static void
static int
aes_set_mbs(struct aes *aes, const char *mbs)
{
if (aes->aes_mbs_alloc) {
free(aes->aes_mbs_alloc);
aes->aes_mbs_alloc = NULL;
}
if (aes->aes_wcs_alloc) {
free(aes->aes_wcs_alloc);
aes->aes_wcs_alloc = NULL;
}
aes->aes_mbs = mbs;
aes->aes_wcs = NULL;
return (aes_copy_mbs(aes, mbs));
}
static void
static int
aes_copy_mbs(struct aes *aes, const char *mbs)
{
if (aes->aes_mbs_alloc) {
free(aes->aes_mbs_alloc);
aes->aes_mbs_alloc = NULL;
archive_strcpy(&(aes->aes_mbs), mbs);
archive_string_empty(&(aes->aes_utf8));
if (aes->aes_wcs) {
free((wchar_t *)(uintptr_t)aes->aes_wcs);
aes->aes_wcs = NULL;
}
if (aes->aes_wcs_alloc) {
free(aes->aes_wcs_alloc);
aes->aes_wcs_alloc = NULL;
}
aes->aes_mbs_alloc = (char *)malloc((strlen(mbs) + 1) * sizeof(char));
if (aes->aes_mbs_alloc == NULL)
__archive_errx(1, "No memory for aes_copy_mbs()");
strcpy(aes->aes_mbs_alloc, mbs);
aes->aes_mbs = aes->aes_mbs_alloc;
aes->aes_wcs = NULL;
return (0);
}
#if 0
static void
aes_set_wcs(struct aes *aes, const wchar_t *wcs)
/*
* The 'update' form tries to proactively update all forms of
* this string (WCS and MBS) and returns an error if any of
* them fail. This is used by the 'pax' handler, for instance,
* to detect and report character-conversion failures early while
* still allowing clients to get potentially useful values from
* the more tolerant lazy conversions. (get_mbs and get_wcs will
* strive to give the user something useful, so you can get hopefully
* usable values even if some of the character conversions are failing.)
*/
static int
aes_update_utf8(struct aes *aes, const char *utf8)
{
if (aes->aes_mbs_alloc) {
free(aes->aes_mbs_alloc);
aes->aes_mbs_alloc = NULL;
}
if (aes->aes_wcs_alloc) {
free(aes->aes_wcs_alloc);
aes->aes_wcs_alloc = NULL;
}
aes->aes_mbs = NULL;
aes->aes_wcs = wcs;
}
#endif
/* Save the UTF8 string. */
archive_strcpy(&(aes->aes_utf8), utf8);
static void
/* Empty the mbs and wcs strings. */
archive_string_empty(&(aes->aes_mbs));
if (aes->aes_wcs) {
free((wchar_t *)(uintptr_t)aes->aes_wcs);
aes->aes_wcs = NULL;
}
/* TODO: We should just do a UTF-8 to MBS conversion here.
* That would be faster, use less space, and give the same
* information. */
/* Try converting UTF8 to WCS, return false on failure. */
aes->aes_wcs = __archive_string_utf8_w(&(aes->aes_utf8));
if (aes->aes_wcs == NULL)
return (0);
/* Try converting WCS to MBS, return false on failure. */
if (archive_strappend_w_mbs(&(aes->aes_mbs), aes->aes_wcs) == NULL)
return (0);
/* All conversions succeeded. */
return (1);
}
static int
aes_copy_wcs(struct aes *aes, const wchar_t *wcs)
{
aes_copy_wcs_len(aes, wcs, wcslen(wcs));
return aes_copy_wcs_len(aes, wcs, wcslen(wcs));
}
static void
static int
aes_copy_wcs_len(struct aes *aes, const wchar_t *wcs, size_t len)
{
if (aes->aes_mbs_alloc) {
free(aes->aes_mbs_alloc);
aes->aes_mbs_alloc = NULL;
wchar_t *w;
archive_string_empty(&(aes->aes_mbs));
archive_string_empty(&(aes->aes_utf8));
if (aes->aes_wcs) {
free((wchar_t *)(uintptr_t)aes->aes_wcs);
aes->aes_wcs = NULL;
}
if (aes->aes_wcs_alloc) {
free(aes->aes_wcs_alloc);
aes->aes_wcs_alloc = NULL;
}
aes->aes_mbs = NULL;
aes->aes_wcs_alloc = (wchar_t *)malloc((len + 1) * sizeof(wchar_t));
if (aes->aes_wcs_alloc == NULL)
w = (wchar_t *)malloc((len + 1) * sizeof(wchar_t));
if (w == NULL)
__archive_errx(1, "No memory for aes_copy_wcs()");
wmemcpy(aes->aes_wcs_alloc, wcs, len);
aes->aes_wcs_alloc[len] = L'\0';
aes->aes_wcs = aes->aes_wcs_alloc;
wmemcpy(w, wcs, len);
w[len] = L'\0';
aes->aes_wcs = w;
return (0);
}
/****************************************************************************
*
* Public Interface
*
****************************************************************************/
struct archive_entry *
archive_entry_clear(struct archive_entry *entry)
{
@ -350,6 +366,8 @@ archive_entry_clone(struct archive_entry *entry)
aes_copy(&entry2->ae_hardlink, &entry->ae_hardlink);
aes_copy(&entry2->ae_pathname, &entry->ae_pathname);
aes_copy(&entry2->ae_symlink, &entry->ae_symlink);
entry2->ae_hardlinkset = entry->ae_hardlinkset;
entry2->ae_symlinkset = entry->ae_symlinkset;
aes_copy(&entry2->ae_uname, &entry->ae_uname);
/* Copy ACL data over. */
@ -515,12 +533,16 @@ archive_entry_gname_w(struct archive_entry *entry)
const char *
archive_entry_hardlink(struct archive_entry *entry)
{
if (!entry->ae_hardlinkset)
return (NULL);
return (aes_get_mbs(&entry->ae_hardlink));
}
const wchar_t *
archive_entry_hardlink_w(struct archive_entry *entry)
{
if (!entry->ae_hardlinkset)
return (NULL);
return (aes_get_wcs(&entry->ae_hardlink));
}
@ -603,12 +625,16 @@ archive_entry_size(struct archive_entry *entry)
const char *
archive_entry_symlink(struct archive_entry *entry)
{
if (!entry->ae_symlinkset)
return (NULL);
return (aes_get_mbs(&entry->ae_symlink));
}
const wchar_t *
archive_entry_symlink_w(struct archive_entry *entry)
{
if (!entry->ae_symlinkset)
return (NULL);
return (aes_get_wcs(&entry->ae_symlink));
}
@ -651,6 +677,15 @@ archive_entry_set_fflags(struct archive_entry *entry,
entry->ae_fflags_clear = clear;
}
const char *
archive_entry_copy_fflags_text(struct archive_entry *entry,
const char *flags)
{
aes_copy_mbs(&entry->ae_fflags_text, flags);
return (ae_strtofflags(flags,
&entry->ae_fflags_set, &entry->ae_fflags_clear));
}
const wchar_t *
archive_entry_copy_fflags_text_w(struct archive_entry *entry,
const wchar_t *flags)
@ -685,6 +720,12 @@ archive_entry_copy_gname_w(struct archive_entry *entry, const wchar_t *name)
aes_copy_wcs(&entry->ae_gname, name);
}
int
archive_entry_update_gname_utf8(struct archive_entry *entry, const char *name)
{
return (aes_update_utf8(&entry->ae_gname, name));
}
void
archive_entry_set_ino(struct archive_entry *entry, unsigned long ino)
{
@ -696,18 +737,24 @@ void
archive_entry_set_hardlink(struct archive_entry *entry, const char *target)
{
aes_set_mbs(&entry->ae_hardlink, target);
if (target != NULL)
entry->ae_hardlinkset = 1;
}
void
archive_entry_copy_hardlink(struct archive_entry *entry, const char *target)
{
aes_copy_mbs(&entry->ae_hardlink, target);
if (target != NULL)
entry->ae_hardlinkset = 1;
}
void
archive_entry_copy_hardlink_w(struct archive_entry *entry, const wchar_t *target)
{
aes_copy_wcs(&entry->ae_hardlink, target);
if (target != NULL)
entry->ae_hardlinkset = 1;
}
void
@ -754,8 +801,7 @@ archive_entry_set_devminor(struct archive_entry *entry, dev_t m)
void
archive_entry_set_link(struct archive_entry *entry, const char *target)
{
if (entry->ae_symlink.aes_mbs != NULL ||
entry->ae_symlink.aes_wcs != NULL)
if (entry->ae_symlinkset)
aes_set_mbs(&entry->ae_symlink, target);
else
aes_set_mbs(&entry->ae_hardlink, target);
@ -765,8 +811,7 @@ archive_entry_set_link(struct archive_entry *entry, const char *target)
void
archive_entry_copy_link(struct archive_entry *entry, const char *target)
{
if (entry->ae_symlink.aes_mbs != NULL ||
entry->ae_symlink.aes_wcs != NULL)
if (entry->ae_symlinkset)
aes_copy_mbs(&entry->ae_symlink, target);
else
aes_copy_mbs(&entry->ae_hardlink, target);
@ -776,13 +821,21 @@ archive_entry_copy_link(struct archive_entry *entry, const char *target)
void
archive_entry_copy_link_w(struct archive_entry *entry, const wchar_t *target)
{
if (entry->ae_symlink.aes_mbs != NULL ||
entry->ae_symlink.aes_wcs != NULL)
if (entry->ae_symlinkset)
aes_copy_wcs(&entry->ae_symlink, target);
else
aes_copy_wcs(&entry->ae_hardlink, target);
}
int
archive_entry_update_link_utf8(struct archive_entry *entry, const char *target)
{
if (entry->ae_symlinkset)
return (aes_update_utf8(&entry->ae_symlink, target));
else
return (aes_update_utf8(&entry->ae_hardlink, target));
}
void
archive_entry_set_mode(struct archive_entry *entry, mode_t m)
{
@ -823,6 +876,12 @@ archive_entry_copy_pathname_w(struct archive_entry *entry, const wchar_t *name)
aes_copy_wcs(&entry->ae_pathname, name);
}
int
archive_entry_update_pathname_utf8(struct archive_entry *entry, const char *name)
{
return (aes_update_utf8(&entry->ae_pathname, name));
}
void
archive_entry_set_perm(struct archive_entry *entry, mode_t p)
{
@ -866,18 +925,24 @@ void
archive_entry_set_symlink(struct archive_entry *entry, const char *linkname)
{
aes_set_mbs(&entry->ae_symlink, linkname);
if (linkname != NULL)
entry->ae_symlinkset = 1;
}
void
archive_entry_copy_symlink(struct archive_entry *entry, const char *linkname)
{
aes_copy_mbs(&entry->ae_symlink, linkname);
if (linkname != NULL)
entry->ae_symlinkset = 1;
}
void
archive_entry_copy_symlink_w(struct archive_entry *entry, const wchar_t *linkname)
{
aes_copy_wcs(&entry->ae_symlink, linkname);
if (linkname != NULL)
entry->ae_symlinkset = 1;
}
void
@ -905,6 +970,12 @@ archive_entry_copy_uname_w(struct archive_entry *entry, const wchar_t *name)
aes_copy_wcs(&entry->ae_uname, name);
}
int
archive_entry_update_uname_utf8(struct archive_entry *entry, const char *name)
{
return (aes_update_utf8(&entry->ae_uname, name));
}
/*
* ACL management. The following would, of course, be a lot simpler
* if: 1) the last draft of POSIX.1e were a really thorough and
@ -1744,7 +1815,7 @@ static struct flag {
* Convert file flags to a comma-separated string. If no flags
* are set, return the empty string.
*/
char *
static char *
ae_fflagstostr(unsigned long bitset, unsigned long bitclear)
{
char *string, *dp;
@ -1788,6 +1859,70 @@ ae_fflagstostr(unsigned long bitset, unsigned long bitclear)
return (string);
}
/*
* strtofflags --
* Take string of arguments and return file flags. This
* version works a little differently than strtofflags(3).
* In particular, it always tests every token, skipping any
* unrecognized tokens. It returns a pointer to the first
* unrecognized token, or NULL if every token was recognized.
* This version is also const-correct and does not modify the
* provided string.
*/
static const char *
ae_strtofflags(const char *s, unsigned long *setp, unsigned long *clrp)
{
const char *start, *end;
struct flag *flag;
unsigned long set, clear;
const char *failed;
set = clear = 0;
start = s;
failed = NULL;
/* Find start of first token. */
while (*start == '\t' || *start == ' ' || *start == ',')
start++;
while (*start != '\0') {
/* Locate end of token. */
end = start;
while (*end != '\0' && *end != '\t' &&
*end != ' ' && *end != ',')
end++;
for (flag = flags; flag->wname != NULL; flag++) {
if (memcmp(start, flag->wname, end - start) == 0) {
/* Matched "noXXXX", so reverse the sense. */
clear |= flag->set;
set |= flag->clear;
break;
} else if (memcmp(start, flag->wname + 2, end - start)
== 0) {
/* Matched "XXXX", so don't reverse. */
set |= flag->set;
clear |= flag->clear;
break;
}
}
/* Ignore unknown flag names. */
if (flag->wname == NULL && failed == NULL)
failed = start;
/* Find start of next token. */
start = end;
while (*start == '\t' || *start == ' ' || *start == ',')
start++;
}
if (setp)
*setp = set;
if (clrp)
*clrp = clear;
/* Return location of first failure. */
return (failed);
}
/*
* wcstofflags --
* Take string of arguments and return file flags. This
@ -1798,7 +1933,7 @@ ae_fflagstostr(unsigned long bitset, unsigned long bitclear)
* This version is also const-correct and does not modify the
* provided string.
*/
const wchar_t *
static const wchar_t *
ae_wcstofflags(const wchar_t *s, unsigned long *setp, unsigned long *clrp)
{
const wchar_t *start, *end;

View file

@ -31,14 +31,46 @@
#include <sys/types.h>
#include <stddef.h> /* for wchar_t */
#include <time.h>
#ifndef _WIN32
#include <unistd.h>
/* Get appropriate definitions of standard POSIX-style types. */
/* These should match the types used in 'struct stat' */
#ifdef _WIN32
#define __LA_UID_T unsigned int
#define __LA_GID_T unsigned int
#define __LA_INO_T unsigned int
#define __LA_DEV_T unsigned int
#define __LA_MODE_T unsigned short
#else
typedef unsigned int uid_t;
typedef unsigned int gid_t;
typedef unsigned int ino_t;
typedef unsigned int dev_t;
typedef unsigned short mode_t;
#include <unistd.h>
#define __LA_UID_T uid_t
#define __LA_GID_T gid_t
#define __LA_INO_T ino_t
#define __LA_DEV_T dev_t
#define __LA_MODE_T mode_t
#endif
/*
* On Windows, define LIBARCHIVE_STATIC if you're building or using a
* .lib. The default here assumes you're building a DLL. Only
* libarchive source should ever define __LIBARCHIVE_BUILD.
*/
#if ((defined __WIN32__) || (defined _WIN32)) && (!defined LIBARCHIVE_STATIC)
# ifdef __LIBARCHIVE_BUILD
# ifdef __GNUC__
# define __LA_DECL __attribute__((dllexport)) extern
# else
# define __LA_DECL __declspec(dllexport)
# endif
# else
# ifdef __GNUC__
# define __LA_DECL __attribute__((dllimport)) extern
# else
# define __LA_DECL __declspec(dllimport)
# endif
# endif
#else
/* Static libraries and shared libraries on non-Windows. */
# define __LA_DECL
#endif
#ifdef __cplusplus
@ -97,50 +129,50 @@ struct archive_entry;
* Basic object manipulation
*/
struct archive_entry *archive_entry_clear(struct archive_entry *);
__LA_DECL struct archive_entry *archive_entry_clear(struct archive_entry *);
/* The 'clone' function does a deep copy; all of the strings are copied too. */
struct archive_entry *archive_entry_clone(struct archive_entry *);
void archive_entry_free(struct archive_entry *);
struct archive_entry *archive_entry_new(void);
__LA_DECL struct archive_entry *archive_entry_clone(struct archive_entry *);
__LA_DECL void archive_entry_free(struct archive_entry *);
__LA_DECL struct archive_entry *archive_entry_new(void);
/*
* Retrieve fields from an archive_entry.
*/
time_t archive_entry_atime(struct archive_entry *);
long archive_entry_atime_nsec(struct archive_entry *);
time_t archive_entry_ctime(struct archive_entry *);
long archive_entry_ctime_nsec(struct archive_entry *);
dev_t archive_entry_dev(struct archive_entry *);
dev_t archive_entry_devmajor(struct archive_entry *);
dev_t archive_entry_devminor(struct archive_entry *);
mode_t archive_entry_filetype(struct archive_entry *);
void archive_entry_fflags(struct archive_entry *,
__LA_DECL time_t archive_entry_atime(struct archive_entry *);
__LA_DECL long archive_entry_atime_nsec(struct archive_entry *);
__LA_DECL time_t archive_entry_ctime(struct archive_entry *);
__LA_DECL long archive_entry_ctime_nsec(struct archive_entry *);
__LA_DECL dev_t archive_entry_dev(struct archive_entry *);
__LA_DECL dev_t archive_entry_devmajor(struct archive_entry *);
__LA_DECL dev_t archive_entry_devminor(struct archive_entry *);
__LA_DECL __LA_MODE_T archive_entry_filetype(struct archive_entry *);
__LA_DECL void archive_entry_fflags(struct archive_entry *,
unsigned long * /* set */,
unsigned long * /* clear */);
const char *archive_entry_fflags_text(struct archive_entry *);
gid_t archive_entry_gid(struct archive_entry *);
const char *archive_entry_gname(struct archive_entry *);
const wchar_t *archive_entry_gname_w(struct archive_entry *);
const char *archive_entry_hardlink(struct archive_entry *);
const wchar_t *archive_entry_hardlink_w(struct archive_entry *);
ino_t archive_entry_ino(struct archive_entry *);
mode_t archive_entry_mode(struct archive_entry *);
time_t archive_entry_mtime(struct archive_entry *);
long archive_entry_mtime_nsec(struct archive_entry *);
unsigned int archive_entry_nlink(struct archive_entry *);
const char *archive_entry_pathname(struct archive_entry *);
const wchar_t *archive_entry_pathname_w(struct archive_entry *);
dev_t archive_entry_rdev(struct archive_entry *);
dev_t archive_entry_rdevmajor(struct archive_entry *);
dev_t archive_entry_rdevminor(struct archive_entry *);
int64_t archive_entry_size(struct archive_entry *);
const char *archive_entry_strmode(struct archive_entry *);
const char *archive_entry_symlink(struct archive_entry *);
const wchar_t *archive_entry_symlink_w(struct archive_entry *);
uid_t archive_entry_uid(struct archive_entry *);
const char *archive_entry_uname(struct archive_entry *);
const wchar_t *archive_entry_uname_w(struct archive_entry *);
__LA_DECL const char *archive_entry_fflags_text(struct archive_entry *);
__LA_DECL __LA_GID_T archive_entry_gid(struct archive_entry *);
__LA_DECL const char *archive_entry_gname(struct archive_entry *);
__LA_DECL const wchar_t *archive_entry_gname_w(struct archive_entry *);
__LA_DECL const char *archive_entry_hardlink(struct archive_entry *);
__LA_DECL const wchar_t *archive_entry_hardlink_w(struct archive_entry *);
__LA_DECL __LA_INO_T archive_entry_ino(struct archive_entry *);
__LA_DECL __LA_MODE_T archive_entry_mode(struct archive_entry *);
__LA_DECL time_t archive_entry_mtime(struct archive_entry *);
__LA_DECL long archive_entry_mtime_nsec(struct archive_entry *);
__LA_DECL unsigned int archive_entry_nlink(struct archive_entry *);
__LA_DECL const char *archive_entry_pathname(struct archive_entry *);
__LA_DECL const wchar_t *archive_entry_pathname_w(struct archive_entry *);
__LA_DECL dev_t archive_entry_rdev(struct archive_entry *);
__LA_DECL dev_t archive_entry_rdevmajor(struct archive_entry *);
__LA_DECL dev_t archive_entry_rdevminor(struct archive_entry *);
__LA_DECL int64_t archive_entry_size(struct archive_entry *);
__LA_DECL const char *archive_entry_strmode(struct archive_entry *);
__LA_DECL const char *archive_entry_symlink(struct archive_entry *);
__LA_DECL const wchar_t *archive_entry_symlink_w(struct archive_entry *);
__LA_DECL __LA_UID_T archive_entry_uid(struct archive_entry *);
__LA_DECL const char *archive_entry_uname(struct archive_entry *);
__LA_DECL const wchar_t *archive_entry_uname_w(struct archive_entry *);
/*
* Set fields in an archive_entry.
@ -149,48 +181,53 @@ const wchar_t *archive_entry_uname_w(struct archive_entry *);
* In contrast, 'copy' functions do copy the object pointed to.
*/
void archive_entry_set_atime(struct archive_entry *, time_t, long);
void archive_entry_set_ctime(struct archive_entry *, time_t, long);
void archive_entry_set_dev(struct archive_entry *, dev_t);
void archive_entry_set_devmajor(struct archive_entry *, dev_t);
void archive_entry_set_devminor(struct archive_entry *, dev_t);
void archive_entry_set_filetype(struct archive_entry *, unsigned int);
void archive_entry_set_fflags(struct archive_entry *,
__LA_DECL void archive_entry_set_atime(struct archive_entry *, time_t, long);
__LA_DECL void archive_entry_set_ctime(struct archive_entry *, time_t, long);
__LA_DECL void archive_entry_set_dev(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_devmajor(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_devminor(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_filetype(struct archive_entry *, unsigned int);
__LA_DECL void archive_entry_set_fflags(struct archive_entry *,
unsigned long /* set */, unsigned long /* clear */);
/* Returns pointer to start of first invalid token, or NULL if none. */
/* Note that all recognized tokens are processed, regardless. */
const wchar_t *archive_entry_copy_fflags_text_w(struct archive_entry *,
__LA_DECL const char *archive_entry_copy_fflags_text(struct archive_entry *,
const char *);
__LA_DECL const wchar_t *archive_entry_copy_fflags_text_w(struct archive_entry *,
const wchar_t *);
void archive_entry_set_gid(struct archive_entry *, gid_t);
void archive_entry_set_gname(struct archive_entry *, const char *);
void archive_entry_copy_gname(struct archive_entry *, const char *);
void archive_entry_copy_gname_w(struct archive_entry *, const wchar_t *);
void archive_entry_set_hardlink(struct archive_entry *, const char *);
void archive_entry_copy_hardlink(struct archive_entry *, const char *);
void archive_entry_copy_hardlink_w(struct archive_entry *, const wchar_t *);
void archive_entry_set_ino(struct archive_entry *, unsigned long);
void archive_entry_set_link(struct archive_entry *, const char *);
void archive_entry_copy_link(struct archive_entry *, const char *);
void archive_entry_copy_link_w(struct archive_entry *, const wchar_t *);
void archive_entry_set_mode(struct archive_entry *, mode_t);
void archive_entry_set_mtime(struct archive_entry *, time_t, long);
void archive_entry_set_nlink(struct archive_entry *, unsigned int);
void archive_entry_set_pathname(struct archive_entry *, const char *);
void archive_entry_copy_pathname(struct archive_entry *, const char *);
void archive_entry_copy_pathname_w(struct archive_entry *, const wchar_t *);
void archive_entry_set_perm(struct archive_entry *, mode_t);
void archive_entry_set_rdev(struct archive_entry *, dev_t);
void archive_entry_set_rdevmajor(struct archive_entry *, dev_t);
void archive_entry_set_rdevminor(struct archive_entry *, dev_t);
void archive_entry_set_size(struct archive_entry *, int64_t);
void archive_entry_set_symlink(struct archive_entry *, const char *);
void archive_entry_copy_symlink(struct archive_entry *, const char *);
void archive_entry_copy_symlink_w(struct archive_entry *, const wchar_t *);
void archive_entry_set_uid(struct archive_entry *, uid_t);
void archive_entry_set_uname(struct archive_entry *, const char *);
void archive_entry_copy_uname(struct archive_entry *, const char *);
void archive_entry_copy_uname_w(struct archive_entry *, const wchar_t *);
__LA_DECL void archive_entry_set_gid(struct archive_entry *, __LA_GID_T);
__LA_DECL void archive_entry_set_gname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_gname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_gname_w(struct archive_entry *, const wchar_t *);
__LA_DECL int archive_entry_update_gname_utf8(struct archive_entry *, const char *);
__LA_DECL void archive_entry_set_hardlink(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_hardlink(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_hardlink_w(struct archive_entry *, const wchar_t *);
__LA_DECL void archive_entry_set_ino(struct archive_entry *, unsigned long);
__LA_DECL void archive_entry_set_link(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_link(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_link_w(struct archive_entry *, const wchar_t *);
__LA_DECL int archive_entry_update_link_utf8(struct archive_entry *, const char *);
__LA_DECL void archive_entry_set_mode(struct archive_entry *, __LA_MODE_T);
__LA_DECL void archive_entry_set_mtime(struct archive_entry *, time_t, long);
__LA_DECL void archive_entry_set_nlink(struct archive_entry *, unsigned int);
__LA_DECL void archive_entry_set_pathname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_pathname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_pathname_w(struct archive_entry *, const wchar_t *);
__LA_DECL int archive_entry_update_pathname_utf8(struct archive_entry *, const char *);
__LA_DECL void archive_entry_set_perm(struct archive_entry *, __LA_MODE_T);
__LA_DECL void archive_entry_set_rdev(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_rdevmajor(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_rdevminor(struct archive_entry *, dev_t);
__LA_DECL void archive_entry_set_size(struct archive_entry *, int64_t);
__LA_DECL void archive_entry_set_symlink(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_symlink(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_symlink_w(struct archive_entry *, const wchar_t *);
__LA_DECL void archive_entry_set_uid(struct archive_entry *, __LA_UID_T);
__LA_DECL void archive_entry_set_uname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_uname(struct archive_entry *, const char *);
__LA_DECL void archive_entry_copy_uname_w(struct archive_entry *, const wchar_t *);
__LA_DECL int archive_entry_update_uname_utf8(struct archive_entry *, const char *);
/*
* Routines to bulk copy fields to/from a platform-native "struct
* stat." Libarchive used to just store a struct stat inside of each
@ -200,8 +237,8 @@ void archive_entry_copy_uname_w(struct archive_entry *, const wchar_t *);
*
* TODO: On Linux, provide both stat32 and stat64 versions of these functions.
*/
const struct stat *archive_entry_stat(struct archive_entry *);
void archive_entry_copy_stat(struct archive_entry *, const struct stat *);
__LA_DECL const struct stat *archive_entry_stat(struct archive_entry *);
__LA_DECL void archive_entry_copy_stat(struct archive_entry *, const struct stat *);
/*
* ACL routines. This used to simply store and return text-format ACL
@ -249,11 +286,11 @@ void archive_entry_copy_stat(struct archive_entry *, const struct stat *);
* POSIX.1e) is useful for handling archive formats that combine
* default and access information in a single ACL list.
*/
void archive_entry_acl_clear(struct archive_entry *);
void archive_entry_acl_add_entry(struct archive_entry *,
__LA_DECL void archive_entry_acl_clear(struct archive_entry *);
__LA_DECL void archive_entry_acl_add_entry(struct archive_entry *,
int /* type */, int /* permset */, int /* tag */,
int /* qual */, const char * /* name */);
void archive_entry_acl_add_entry_w(struct archive_entry *,
__LA_DECL void archive_entry_acl_add_entry_w(struct archive_entry *,
int /* type */, int /* permset */, int /* tag */,
int /* qual */, const wchar_t * /* name */);
@ -262,11 +299,11 @@ void archive_entry_acl_add_entry_w(struct archive_entry *,
* "next" entry. The want_type parameter allows you to request only
* access entries or only default entries.
*/
int archive_entry_acl_reset(struct archive_entry *, int /* want_type */);
int archive_entry_acl_next(struct archive_entry *, int /* want_type */,
__LA_DECL int archive_entry_acl_reset(struct archive_entry *, int /* want_type */);
__LA_DECL int archive_entry_acl_next(struct archive_entry *, int /* want_type */,
int * /* type */, int * /* permset */, int * /* tag */,
int * /* qual */, const char ** /* name */);
int archive_entry_acl_next_w(struct archive_entry *, int /* want_type */,
__LA_DECL int archive_entry_acl_next_w(struct archive_entry *, int /* want_type */,
int * /* type */, int * /* permset */, int * /* tag */,
int * /* qual */, const wchar_t ** /* name */);
@ -283,11 +320,11 @@ int archive_entry_acl_next_w(struct archive_entry *, int /* want_type */,
*/
#define ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID 1024
#define ARCHIVE_ENTRY_ACL_STYLE_MARK_DEFAULT 2048
const wchar_t *archive_entry_acl_text_w(struct archive_entry *,
__LA_DECL const wchar_t *archive_entry_acl_text_w(struct archive_entry *,
int /* flags */);
/* Return a count of entries matching 'want_type' */
int archive_entry_acl_count(struct archive_entry *, int /* want_type */);
__LA_DECL int archive_entry_acl_count(struct archive_entry *, int /* want_type */);
/*
* Private ACL parser. This is private because it handles some
@ -302,15 +339,15 @@ int archive_entry_acl_count(struct archive_entry *, int /* want_type */);
* TODO: Move this declaration out of the public header and into
* a private header. Warnings above are silly.
*/
int __archive_entry_acl_parse_w(struct archive_entry *,
__LA_DECL int __archive_entry_acl_parse_w(struct archive_entry *,
const wchar_t *, int /* type */);
/*
* extended attributes
*/
void archive_entry_xattr_clear(struct archive_entry *);
void archive_entry_xattr_add_entry(struct archive_entry *,
__LA_DECL void archive_entry_xattr_clear(struct archive_entry *);
__LA_DECL void archive_entry_xattr_add_entry(struct archive_entry *,
const char * /* name */, const void * /* value */,
size_t /* size */);
@ -319,9 +356,9 @@ void archive_entry_xattr_add_entry(struct archive_entry *,
* "next" entry.
*/
int archive_entry_xattr_count(struct archive_entry *);
int archive_entry_xattr_reset(struct archive_entry *);
int archive_entry_xattr_next(struct archive_entry *,
__LA_DECL int archive_entry_xattr_count(struct archive_entry *);
__LA_DECL int archive_entry_xattr_reset(struct archive_entry *);
__LA_DECL int archive_entry_xattr_next(struct archive_entry *,
const char ** /* name */, const void ** /* value */, size_t *);
/*
@ -352,7 +389,7 @@ int archive_entry_xattr_next(struct archive_entry *,
* Note that archive_entry_size() is reset to zero if the file
* body should not be written to the archive. Pay attention!
*/
struct archive_entry_linkresolver;
__LA_DECL struct archive_entry_linkresolver;
/*
* There are three different strategies for marking hardlinks.
@ -394,15 +431,18 @@ struct archive_entry_linkresolver;
* correctly extract an arbitrary link.
*/
struct archive_entry_linkresolver *archive_entry_linkresolver_new(void);
void archive_entry_linkresolver_set_strategy(
__LA_DECL struct archive_entry_linkresolver *archive_entry_linkresolver_new(void);
__LA_DECL void archive_entry_linkresolver_set_strategy(
struct archive_entry_linkresolver *, int /* format_code */);
void archive_entry_linkresolver_free(struct archive_entry_linkresolver *);
void archive_entry_linkify(struct archive_entry_linkresolver *,
__LA_DECL void archive_entry_linkresolver_free(struct archive_entry_linkresolver *);
__LA_DECL void archive_entry_linkify(struct archive_entry_linkresolver *,
struct archive_entry **, struct archive_entry **);
#ifdef __cplusplus
}
#endif
/* This is meaningless outside of this header. */
#undef __LA_DECL
#endif /* !ARCHIVE_ENTRY_H_INCLUDED */

View file

@ -28,6 +28,8 @@
#ifndef ARCHIVE_ENTRY_PRIVATE_H_INCLUDED
#define ARCHIVE_ENTRY_PRIVATE_H_INCLUDED
#include "archive_string.h"
/*
* Handle wide character (i.e., Unicode) and non-wide character
* strings transparently.
@ -35,10 +37,9 @@
*/
struct aes {
const char *aes_mbs;
char *aes_mbs_alloc;
struct archive_string aes_mbs;
struct archive_string aes_utf8;
const wchar_t *aes_wcs;
wchar_t *aes_wcs_alloc;
};
struct ae_acl {
@ -141,6 +142,8 @@ struct archive_entry {
struct aes ae_pathname; /* Name of entry */
struct aes ae_symlink; /* symlink contents */
struct aes ae_uname; /* Name of owner */
unsigned char ae_hardlinkset;
unsigned char ae_symlinkset;
struct ae_acl *acl_head;
struct ae_acl *acl_p;

View file

@ -36,6 +36,9 @@
#ifndef ARCHIVE_PLATFORM_H_INCLUDED
#define ARCHIVE_PLATFORM_H_INCLUDED
/* archive.h and archive_entry.h require this. */
#define __LIBARCHIVE_BUILD 1
#ifdef _WIN32
#include "config_windows.h"
#include "archive_windows.h"

View file

@ -145,6 +145,8 @@ struct sparse_block {
struct tar {
struct archive_string acl_text;
struct archive_string entry_pathname;
/* For "GNU.sparse.name" and other similar path extensions. */
struct archive_string entry_pathname_override;
struct archive_string entry_linkpath;
struct archive_string entry_uname;
struct archive_string entry_gname;
@ -272,6 +274,7 @@ archive_read_format_tar_cleanup(struct archive_read *a)
gnu_clear_sparse_list(tar);
archive_string_free(&tar->acl_text);
archive_string_free(&tar->entry_pathname);
archive_string_free(&tar->entry_pathname_override);
archive_string_free(&tar->entry_linkpath);
archive_string_free(&tar->entry_uname);
archive_string_free(&tar->entry_gname);
@ -1174,7 +1177,6 @@ pax_header(struct archive_read *a, struct tar *tar,
size_t attr_length, l, line_length;
char *line, *p;
char *key, *value;
wchar_t *wp;
int err, err2;
attr_length = strlen(attr);
@ -1182,6 +1184,7 @@ pax_header(struct archive_read *a, struct tar *tar,
archive_string_empty(&(tar->entry_gname));
archive_string_empty(&(tar->entry_linkpath));
archive_string_empty(&(tar->entry_pathname));
archive_string_empty(&(tar->entry_pathname_override));
archive_string_empty(&(tar->entry_uname));
err = ARCHIVE_OK;
while (attr_length > 0) {
@ -1257,13 +1260,13 @@ pax_header(struct archive_read *a, struct tar *tar,
if (tar->pax_hdrcharset_binary)
archive_entry_copy_gname(entry, value);
else {
wp = utf8_decode(tar, value, strlen(value));
if (wp == NULL) {
archive_entry_copy_gname(entry, value);
if (err > ARCHIVE_WARN)
err = ARCHIVE_WARN;
} else
archive_entry_copy_gname_w(entry, wp);
if (!archive_entry_update_gname_utf8(entry, value)) {
err = ARCHIVE_WARN;
archive_set_error(&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"Gname in pax header can't "
"be converted to current locale.");
}
}
}
if (archive_strlen(&(tar->entry_linkpath)) > 0) {
@ -1271,27 +1274,40 @@ pax_header(struct archive_read *a, struct tar *tar,
if (tar->pax_hdrcharset_binary)
archive_entry_copy_link(entry, value);
else {
wp = utf8_decode(tar, value, strlen(value));
if (wp == NULL) {
archive_entry_copy_link(entry, value);
if (err > ARCHIVE_WARN)
err = ARCHIVE_WARN;
} else
archive_entry_copy_link_w(entry, wp);
if (!archive_entry_update_link_utf8(entry, value)) {
err = ARCHIVE_WARN;
archive_set_error(&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"Linkname in pax header can't "
"be converted to current locale.");
}
}
}
if (archive_strlen(&(tar->entry_pathname)) > 0) {
/*
* Some extensions (such as the GNU sparse file extensions)
* deliberately store a synthetic name under the regular 'path'
* attribute and the real file name under a different attribute.
* Since we're supposed to not care about the order, we
* have no choice but to store all of the various filenames
* we find and figure it all out afterwards. This is the
* figuring out part.
*/
value = NULL;
if (archive_strlen(&(tar->entry_pathname_override)) > 0)
value = tar->entry_pathname_override.s;
else if (archive_strlen(&(tar->entry_pathname)) > 0)
value = tar->entry_pathname.s;
if (value != NULL) {
if (tar->pax_hdrcharset_binary)
archive_entry_copy_pathname(entry, value);
else {
wp = utf8_decode(tar, value, strlen(value));
if (wp == NULL) {
archive_entry_copy_pathname(entry, value);
if (err > ARCHIVE_WARN)
err = ARCHIVE_WARN;
} else
archive_entry_copy_pathname_w(entry, wp);
if (!archive_entry_update_pathname_utf8(entry, value)) {
err = ARCHIVE_WARN;
archive_set_error(&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"Pathname in pax header can't be "
"converted to current locale.");
}
}
}
if (archive_strlen(&(tar->entry_uname)) > 0) {
@ -1299,13 +1315,13 @@ pax_header(struct archive_read *a, struct tar *tar,
if (tar->pax_hdrcharset_binary)
archive_entry_copy_uname(entry, value);
else {
wp = utf8_decode(tar, value, strlen(value));
if (wp == NULL) {
archive_entry_copy_uname(entry, value);
if (err > ARCHIVE_WARN)
err = ARCHIVE_WARN;
} else
archive_entry_copy_uname_w(entry, wp);
if (!archive_entry_update_uname_utf8(entry, value)) {
err = ARCHIVE_WARN;
archive_set_error(&a->archive,
ARCHIVE_ERRNO_FILE_FORMAT,
"Uname in pax header can't "
"be converted to current locale.");
}
}
}
return (err);
@ -1415,11 +1431,13 @@ pax_attribute(struct tar *tar, struct archive_entry *entry,
tar->sparse_gnu_pending = 1;
}
if (strcmp(key, "GNU.sparse.name") == 0) {
wp = utf8_decode(tar, value, strlen(value));
if (wp != NULL)
archive_entry_copy_pathname_w(entry, wp);
else
archive_entry_copy_pathname(entry, value);
/*
* The real filename; when storing sparse
* files, GNU tar puts a synthesized name into
* the regular 'path' attribute in an attempt
* to limit confusion. ;-)
*/
archive_strcpy(&(tar->entry_pathname_override), value);
}
if (strcmp(key, "GNU.sparse.realsize") == 0) {
tar->realsize = tar_atol10(value, strlen(value));
@ -1455,9 +1473,7 @@ pax_attribute(struct tar *tar, struct archive_entry *entry,
archive_entry_set_rdevminor(entry,
tar_atol10(value, strlen(value)));
} else if (strcmp(key, "SCHILY.fflags")==0) {
wp = utf8_decode(tar, value, strlen(value));
/* TODO: if (wp == NULL) */
archive_entry_copy_fflags_text_w(entry, wp);
archive_entry_copy_fflags_text(entry, value);
} else if (strcmp(key, "SCHILY.dev")==0) {
archive_entry_set_dev(entry,
tar_atol10(value, strlen(value)));

View file

@ -37,6 +37,9 @@ __FBSDID("$FreeBSD: src/lib/libarchive/archive_string.c,v 1.11 2007/07/15 19:13:
#ifdef HAVE_STRING_H
#include <string.h>
#endif
#ifdef HAVE_WCHAR_H
#include <wchar.h>
#endif
#include "archive_private.h"
#include "archive_string.h"
@ -55,11 +58,15 @@ __archive_string_append(struct archive_string *as, const char *p, size_t s)
void
__archive_string_copy(struct archive_string *dest, struct archive_string *src)
{
if (__archive_string_ensure(dest, src->length + 1) == NULL)
__archive_errx(1, "Out of memory");
memcpy(dest->s, src->s, src->length);
dest->length = src->length;
dest->s[dest->length] = 0;
if (src->length == 0)
dest->length = 0;
else {
if (__archive_string_ensure(dest, src->length + 1) == NULL)
__archive_errx(1, "Out of memory");
memcpy(dest->s, src->s, src->length);
dest->length = src->length;
dest->s[dest->length] = 0;
}
}
void
@ -67,21 +74,52 @@ __archive_string_free(struct archive_string *as)
{
as->length = 0;
as->buffer_length = 0;
if (as->s != NULL)
if (as->s != NULL) {
free(as->s);
as->s = NULL;
}
}
/* Returns NULL on any allocation failure. */
struct archive_string *
__archive_string_ensure(struct archive_string *as, size_t s)
{
/* If buffer is already big enough, don't reallocate. */
if (as->s && (s <= as->buffer_length))
return (as);
/*
* Growing the buffer at least exponentially ensures that
* append operations are always linear in the number of
* characters appended. Using a smaller growth rate for
* larger buffers reduces memory waste somewhat at the cost of
* a larger constant factor.
*/
if (as->buffer_length < 32)
/* Start with a minimum 32-character buffer. */
as->buffer_length = 32;
while (as->buffer_length < s)
else if (as->buffer_length < 8192)
/* Buffers under 8k are doubled for speed. */
as->buffer_length *= 2;
else {
/* Buffers 8k and over grow by at least 25% each time. */
size_t old_length = as->buffer_length;
as->buffer_length = (as->buffer_length * 5) / 4;
/* Be safe: If size wraps, release buffer and return NULL. */
if (as->buffer_length < old_length) {
free(as->s);
as->s = NULL;
return (NULL);
}
}
/*
* The computation above is a lower limit to how much we'll
* grow the buffer. In any case, we have to grow it enough to
* hold the request.
*/
if (as->buffer_length < s)
as->buffer_length = s;
/* Now we can reallocate the buffer. */
as->s = (char *)realloc(as->s, as->buffer_length);
if (as->s == NULL)
return (NULL);
@ -124,3 +162,206 @@ __archive_strappend_int(struct archive_string *as, int d, int base)
__archive_strappend_char(as, digits[d % base]);
return (as);
}
/*
* Home-grown wcrtomb for UTF-8.
*/
static size_t
my_wcrtomb_utf8(char *p, wchar_t wc, mbstate_t *s)
{
(void)s; /* UNUSED */
if (p == NULL)
return (0);
if (wc <= 0x7f) {
p[0] = (char)wc;
return (1);
}
if (wc <= 0x7ff) {
p[0] = 0xc0 | ((wc >> 6) & 0x1f);
p[1] = 0x80 | (wc & 0x3f);
return (2);
}
if (wc <= 0xffff) {
p[0] = 0xe0 | ((wc >> 12) & 0x0f);
p[1] = 0x80 | ((wc >> 6) & 0x3f);
p[2] = 0x80 | (wc & 0x3f);
return (3);
}
if (wc <= 0x1fffff) {
p[0] = 0xf0 | ((wc >> 18) & 0x07);
p[1] = 0x80 | ((wc >> 12) & 0x3f);
p[2] = 0x80 | ((wc >> 6) & 0x3f);
p[3] = 0x80 | (wc & 0x3f);
return (4);
}
/* Unicode has no codes larger than 0x1fffff. */
/*
* Awkward point: UTF-8 <-> wchar_t conversions
* can actually fail.
*/
return ((size_t)-1);
}
static int
my_wcstombs(struct archive_string *as, const wchar_t *w,
size_t (*func)(char *, wchar_t, mbstate_t *))
{
size_t n;
char *p;
mbstate_t shift_state;
char buff[256];
/*
* Convert one wide char at a time into 'buff', whenever that
* fills, append it to the string.
*/
p = buff;
wcrtomb(NULL, L'\0', &shift_state);
while (*w != L'\0') {
/* Flush the buffer when we have <=16 bytes free. */
/* (No encoding has a single character >16 bytes.) */
if ((size_t)(p - buff) >= (size_t)(sizeof(buff) - 16)) {
*p = '\0';
archive_strcat(as, buff);
p = buff;
}
n = (*func)(p, *w++, &shift_state);
if (n == (size_t)-1)
return (-1);
p += n;
}
*p = '\0';
archive_strcat(as, buff);
return (0);
}
/*
* Translates a wide character string into UTF-8 and appends
* to the archive_string. Note: returns NULL if conversion fails.
*/
struct archive_string *
__archive_strappend_w_utf8(struct archive_string *as, const wchar_t *w)
{
if (my_wcstombs(as, w, my_wcrtomb_utf8))
return (NULL);
return (as);
}
/*
* Translates a wide character string into current locale character set
* and appends to the archive_string. Note: returns NULL if conversion
* fails.
*
* TODO: use my_wcrtomb_utf8 if !HAVE_WCRTOMB (add configure logic first!)
*/
struct archive_string *
__archive_strappend_w_mbs(struct archive_string *as, const wchar_t *w)
{
if (my_wcstombs(as, w, wcrtomb))
return (NULL);
return (as);
}
/*
* Home-grown mbrtowc for UTF-8. Some systems lack UTF-8
* (or even lack mbrtowc()) and we need UTF-8 support for pax
* format. So please don't replace this with a call to the
* standard mbrtowc() function!
*/
static size_t
my_mbrtowc_utf8(wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
{
int ch;
/*
* This argument is here to make the prototype identical to the
* standard mbrtowc(), so I can build generic string processors
* that just accept a pointer to a suitable mbrtowc() function.
*/
(void)ps; /* UNUSED */
/* Standard behavior: a NULL value for 's' just resets shift state. */
if (s == NULL)
return (0);
/* If length argument is zero, don't look at the first character. */
if (n <= 0)
return ((size_t)-2);
/*
* Decode 1-4 bytes depending on the value of the first byte.
*/
ch = (unsigned char)*s;
if (ch == 0) {
return (0); /* Standard: return 0 for end-of-string. */
}
if ((ch & 0x80) == 0) {
*pwc = ch & 0x7f;
return (1);
}
if ((ch & 0xe0) == 0xc0) {
if (n < 2)
return ((size_t)-2);
if ((s[1] & 0xc0) != 0x80) return (size_t)-1;
*pwc = ((ch & 0x1f) << 6) | (s[1] & 0x3f);
return (2);
}
if ((ch & 0xf0) == 0xe0) {
if (n < 3)
return ((size_t)-2);
if ((s[1] & 0xc0) != 0x80) return (size_t)-1;
if ((s[2] & 0xc0) != 0x80) return (size_t)-1;
*pwc = ((ch & 0x0f) << 12)
| ((s[1] & 0x3f) << 6)
| (s[2] & 0x3f);
return (3);
}
if ((ch & 0xf8) == 0xf0) {
if (n < 4)
return ((size_t)-2);
if ((s[1] & 0xc0) != 0x80) return (size_t)-1;
if ((s[2] & 0xc0) != 0x80) return (size_t)-1;
if ((s[3] & 0xc0) != 0x80) return (size_t)-1;
*pwc = ((ch & 0x07) << 18)
| ((s[1] & 0x3f) << 12)
| ((s[2] & 0x3f) << 6)
| (s[3] & 0x3f);
return (4);
}
/* Invalid first byte. */
return ((size_t)-1);
}
/*
* Return a wide-character string by converting this archive_string
* from UTF-8.
*/
wchar_t *
__archive_string_utf8_w(struct archive_string *as)
{
wchar_t *ws, *dest;
const char *src;
size_t n;
int err;
ws = (wchar_t *)malloc((as->length + 1) * sizeof(wchar_t));
if (ws == NULL)
__archive_errx(1, "Out of memory");
err = 0;
dest = ws;
src = as->s;
while (*src != '\0') {
n = my_mbrtowc_utf8(dest, src, 8, NULL);
if (n == 0)
break;
if (n == (size_t)-1 || n == (size_t)-2) {
free(ws);
return (NULL);
}
dest++;
src += n;
}
*dest++ = L'\0';
return (ws);
}

View file

@ -33,6 +33,9 @@
#ifdef HAVE_STRING_H
#include <string.h>
#endif
#ifdef HAVE_WCHAR_H
#include <wchar.h>
#endif
/*
* Basic resizable/reusable string support a la Java's "StringBuffer."
@ -60,16 +63,22 @@ struct archive_string *
__archive_strappend_char(struct archive_string *, char);
#define archive_strappend_char __archive_strappend_char
/* Append a char to an archive_string using UTF8. */
struct archive_string *
__archive_strappend_char_UTF8(struct archive_string *, int);
#define archive_strappend_char_UTF8 __archive_strappend_char_UTF8
/* Append an integer in the specified base (2 <= base <= 16). */
struct archive_string *
__archive_strappend_int(struct archive_string *as, int d, int base);
#define archive_strappend_int __archive_strappend_int
/* Convert a wide-char string to UTF-8 and append the result. */
struct archive_string *
__archive_strappend_w_utf8(struct archive_string *, const wchar_t *);
#define archive_strappend_w_utf8 __archive_strappend_w_utf8
/* Convert a wide-char string to current locale and append the result. */
/* Returns NULL if conversion fails. */
struct archive_string *
__archive_strappend_w_mbs(struct archive_string *, const wchar_t *);
#define archive_strappend_w_mbs __archive_strappend_w_mbs
/* Basic append operation. */
struct archive_string *
__archive_string_append(struct archive_string *as, const char *p, size_t s);
@ -95,7 +104,7 @@ __archive_strncat(struct archive_string *, const char *, size_t);
/* Copy a C string to an archive_string, resizing as necessary. */
#define archive_strcpy(as,p) \
((as)->length = 0, __archive_string_append((as), (p), strlen(p)))
((as)->length = 0, __archive_string_append((as), (p), p == NULL ? 0 : strlen(p)))
/* Copy a C string to an archive_string with limit, resizing as necessary. */
#define archive_strncpy(as,p,l) \
@ -119,4 +128,9 @@ void __archive_string_vsprintf(struct archive_string *, const char *,
void __archive_string_sprintf(struct archive_string *, const char *, ...);
#define archive_string_sprintf __archive_string_sprintf
/* Allocates a fresh buffer and converts as (assumed to be UTF-8) into it.
* Returns NULL if conversion failed in any way. */
wchar_t *__archive_string_utf8_w(struct archive_string *as);
#endif

View file

@ -496,34 +496,48 @@ archive_write_pax_header(struct archive_write *a,
if (hdrcharset != NULL)
add_pax_attr(&(pax->pax_header), "hdrcharset", hdrcharset);
/*
* Determining whether or not the name is too big is ugly
* because of the rules for dividing names between 'name' and
* 'prefix' fields. Here, I pick out the longest possible
* suffix, then test whether the remaining prefix is too long.
*/
if (strlen(path) <= 100) /* Short enough for just 'name' field */
suffix_start = path; /* Record a zero-length prefix */
else
/* Find the largest suffix that fits in 'name' field. */
suffix_start = strchr(path + strlen(path) - 100 - 1, '/');
/*
* If name is too long, or has non-ASCII characters, add
* 'path' to pax extended attrs. (Note that an unconvertible
* name must have non-ASCII characters.)
*/
if (suffix_start == NULL || suffix_start - path > 155
|| path_w == NULL || has_non_ASCII(path_w)) {
if (path_w == NULL || hdrcharset != NULL)
/* Can't do UTF-8, so store it raw. */
add_pax_attr(&(pax->pax_header), "path", path);
else
add_pax_attr_w(&(pax->pax_header), "path", path_w);
archive_entry_set_pathname(entry_main,
build_ustar_entry_name(ustar_entry_name,
path, strlen(path), NULL));
if (path == NULL) {
/* We don't have a narrow version, so we have to store
* the wide version. */
add_pax_attr_w(&(pax->pax_header), "path", path_w);
archive_entry_set_pathname(entry_main, "@WidePath");
need_extension = 1;
} else {
/* We have a narrow path; we first need to decide
* if we can just put it in the ustar header. */
suffix_start = path; /* Start with a zero-length prefix. */
if (strlen(path) > 100) {
/* Find largest suffix that will fit. */
suffix_start = strchr(path + strlen(path) - 100 - 1, '/');
}
/* We can put it in the ustar header if it's all ASCII
* and it's either <= 100 characters or can be split at a
* '/' into a prefix <= 155 chars and a suffix <= 100 chars.
* (Note the strchr() above will return NULL exactly when
* the path can't be split.)
*/
if (has_non_ASCII(path_w) || suffix_start == NULL
|| suffix_start - path > 155) {
/* Path is either too long or has non-ASCII chars. */
if (path_w == NULL || hdrcharset != NULL) {
/* Can't do UTF-8, so store it raw. */
add_pax_attr(&(pax->pax_header), "path", path);
} else {
/* Store UTF-8 */
add_pax_attr_w(&(pax->pax_header),
"path", path_w);
}
archive_entry_set_pathname(entry_main,
build_ustar_entry_name(ustar_entry_name,
path, strlen(path), NULL));
need_extension = 1;
}
}
if (linkpath != NULL) {
@ -1215,6 +1229,8 @@ archive_write_pax_data(struct archive_write *a, const void *buff, size_t s)
static int
has_non_ASCII(const wchar_t *wp)
{
if (wp == NULL)
return (1);
while (*wp != L'\0' && *wp < 128)
wp++;
return (*wp != L'\0');

View file

@ -149,7 +149,7 @@ followed by the least-significant 16 bits.
Each of the two 16 bit values are stored in machine-native byte order.
.It Va namesize
The number of bytes in the pathname that follows the header.
This count includes the trailing NULL byte.
This count includes the trailing NUL byte.
.It Va filesize
The size of the file.
Note that this archive format is limited to
@ -162,8 +162,8 @@ above for a description of the storage of four-byte integers.
The pathname immediately follows the fixed header.
If the
.Cm namesize
is odd, an additional NULL byte is added after the pathname.
The file data is then appended, padded with NULL
is odd, an additional NUL byte is added after the pathname.
The file data is then appended, padded with NUL
bytes to an even length.
.Pp
Hardlinked files are not given special treatment;
@ -202,7 +202,7 @@ Unlike the old binary format, there is no additional padding
after the pathname or file contents.
If the files being archived are themselves entirely ASCII, then
the resulting archive will be entirely ASCII, except for the
NULL byte that terminates the name field.
NUL byte that terminates the name field.
.Ss New ASCII Format
The "new" ASCII format uses 8-byte hexadecimal fields for
all numbers and separates device numbers into separate fields
@ -237,7 +237,7 @@ This field is always set to zero by writers and ignored by readers.
See the next section for more details.
.El
.Pp
The pathname is followed by NULL bytes so that the total size
The pathname is followed by NUL bytes so that the total size
of the fixed header plus pathname is a multiple of four.
Likewise, the file data is padded to a multiple of four bytes.
Note that this format supports only 4 gigabyte files (unlike the
@ -322,4 +322,4 @@ by SCO under their
license.
The character format was adopted as part of
.St -p1003.1-88 .
XXX when did "newc" appear? Who invented it? When did HP come out with their variant? When did Sun introduce ACLs and extended attributes? XXX
XXX when did "newc" appear? Who invented it? When did HP come out with their variant? When did Sun introduce ACLs and extended attributes? XXX

View file

@ -221,7 +221,7 @@ field with several new type values:
.Bl -tag -width indent -compact
.It Dq 0
Regular file.
NULL should be treated as a synonym, for compatibility purposes.
NUL should be treated as a synonym, for compatibility purposes.
.It Dq 1
Hard link.
.It Dq 2
@ -258,7 +258,7 @@ by readers.
.It Va magic
Contains the magic value
.Dq ustar
followed by a NULL byte to indicate that this is a POSIX standard archive.
followed by a NUL byte to indicate that this is a POSIX standard archive.
Full compliance requires the uname and gname fields be properly set.
.It Va version
Version.
@ -285,7 +285,7 @@ character to the regular name field to obtain the full pathname.
.El
.Pp
Note that all unused bytes must be set to
.Dv NULL .
.Dv NUL .
.Pp
Field termination is specified slightly differently by POSIX
than by previous implementations.
@ -295,14 +295,14 @@ The
and
.Va gname
fields must have a trailing
.Dv NULL .
.Dv NUL .
The
.Va pathname ,
.Va linkname ,
and
.Va prefix
fields must have a trailing
.Dv NULL
.Dv NUL
unless they fill the entire field.
(In particular, it is possible to store a 256-character pathname if it
happens to have a
@ -310,7 +310,7 @@ happens to have a
as the 156th character.)
POSIX requires numeric fields to be zero-padded in the front, and allows
them to be terminated with either space or
.Dv NULL
.Dv NUL
characters.
.Pp
Currently, most tar implementations comply with the ustar

View file

@ -9,7 +9,9 @@ DEFINE_TEST(test_empty_write)
DEFINE_TEST(test_entry)
DEFINE_TEST(test_entry_strmode)
DEFINE_TEST(test_link_resolver)
DEFINE_TEST(test_pax_filename_encoding)
DEFINE_TEST(test_pax_filename_encoding_1)
DEFINE_TEST(test_pax_filename_encoding_2)
DEFINE_TEST(test_pax_filename_encoding_3)
DEFINE_TEST(test_read_compress_program)
DEFINE_TEST(test_read_data_large)
DEFINE_TEST(test_read_extract)

View file

@ -271,6 +271,30 @@ test_assert_equal_int(const char *file, int line,
return (0);
}
static void strdump(const char *p)
{
if (p == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*p != '\0') {
unsigned int c = 0xff & *p++;
switch (c) {
case '\a': fprintf(stderr, "\a"); break;
case '\b': fprintf(stderr, "\b"); break;
case '\n': fprintf(stderr, "\n"); break;
case '\r': fprintf(stderr, "\r"); break;
default:
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else
fprintf(stderr, "\\x%02X", c);
}
}
fprintf(stderr, "\"");
}
/* assertEqualString() displays the values of the two strings. */
int
test_assert_equal_string(const char *file, int line,
@ -293,12 +317,37 @@ test_assert_equal_string(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Strings not equal\n",
file, line);
fprintf(stderr, " %s = \"%s\"\n", e1, v1);
fprintf(stderr, " %s = \"%s\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
strdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
strdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}
static void wcsdump(const wchar_t *w)
{
if (w == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*w != L'\0') {
unsigned int c = *w++;
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else if (c < 256)
fprintf(stderr, "\\x%02X", c);
else if (c < 0x10000)
fprintf(stderr, "\\u%04X", c);
else
fprintf(stderr, "\\U%08X", c);
}
fprintf(stderr, "\"");
}
/* assertEqualWString() displays the values of the two strings. */
int
test_assert_equal_wstring(const char *file, int line,
@ -307,7 +356,17 @@ test_assert_equal_wstring(const char *file, int line,
void *extra)
{
++assertions;
if (wcscmp(v1, v2) == 0) {
if (v1 == NULL) {
if (v2 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (v2 == NULL) {
if (v1 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (wcscmp(v1, v2) == 0) {
msg[0] = '\0';
return (1);
}
@ -316,8 +375,12 @@ test_assert_equal_wstring(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Unicode strings not equal\n",
file, line);
fwprintf(stderr, L" %s = \"%ls\"\n", e1, v1);
fwprintf(stderr, L" %s = \"%ls\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
wcsdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
wcsdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}

View file

@ -332,14 +332,10 @@ acl_match(struct acl_t *acl, int type, int permset, int tag, int qual, const cha
return (1);
if (qual != acl->qual)
return (0);
if (name == NULL) {
if (acl->name == NULL || acl->name[0] == '\0')
return (1);
}
if (acl->name == NULL) {
if (name[0] == '\0')
return (1);
}
if (name == NULL)
return (acl->name == NULL || acl->name[0] == '\0');
if (acl->name == NULL)
return (name == NULL || name[0] == '\0');
return (0 == strcmp(name, acl->name));
}

View file

@ -28,7 +28,7 @@ __FBSDID("$FreeBSD: src/lib/libarchive/test/test_archive_api_feature.c,v 1.4 200
DEFINE_TEST(test_archive_api_feature)
{
char buff[128];
char *p;
const char *p;
/* This is the (hopefully) final versioning API. */
assertEqualInt(ARCHIVE_VERSION_NUMBER, archive_version_number());

View file

@ -52,6 +52,8 @@ DEFINE_TEST(test_entry)
const void *xval; /* For xattr tests. */
size_t xsize; /* For xattr tests. */
int c;
wchar_t wc;
long l;
assert((e = archive_entry_new()) != NULL);
@ -146,7 +148,7 @@ DEFINE_TEST(test_entry)
archive_entry_copy_link_w(e, L"link3");
assertEqualString(archive_entry_hardlink(e), NULL);
assertEqualString(archive_entry_symlink(e), "link3");
/* Arbitrarily override hardlink if both hardlink and symlink set. */
/* Arbitrarily override symlink if both hardlink and symlink set. */
archive_entry_set_hardlink(e, "hardlink");
archive_entry_set_symlink(e, "symlink");
archive_entry_set_link(e, "link");
@ -726,8 +728,10 @@ DEFINE_TEST(test_entry)
/*
* Exercise the character-conversion logic, if we can.
*/
failure("Can't exercise charset-conversion logic.");
if (assert(NULL != setlocale(LC_ALL, "de_DE.UTF-8"))) {
if (NULL == setlocale(LC_ALL, "de_DE.UTF-8")) {
skipping("Can't exercise charset-conversion logic without"
" a suitable locale.");
} else {
/* A filename that cannot be converted to wide characters. */
archive_entry_copy_pathname(e, "abc\314\214mno\374xyz");
failure("Converting invalid chars to Unicode should fail.");
@ -756,6 +760,26 @@ DEFINE_TEST(test_entry)
assert(NULL == archive_entry_symlink_w(e));
}
l = 0x12345678L;
wc = (wchar_t)l; /* Wide character too big for UTF-8. */
if (NULL == setlocale(LC_ALL, "C") || (long)wc != l) {
skipping("Testing charset conversion failure requires 32-bit wchar_t and support for \"C\" locale.");
} else {
/*
* Build the string L"xxx\U12345678yyy\u5678zzz" without
* using C99 \u#### syntax, which isn't uniformly
* supported. (GCC 3.4.6, for instance, defaults to
* "c89 plus GNU extensions.")
*/
wcscpy(wbuff, L"xxxAyyyBzzz");
wbuff[3] = 0x12345678;
wbuff[7] = 0x5678;
/* A wide filename that cannot be converted to narrow. */
archive_entry_copy_pathname_w(e, wbuff);
failure("Converting wide characters from Unicode should fail.");
assertEqualString(NULL, archive_entry_pathname(e));
}
/* Release the experimental entry. */
archive_entry_free(e);
}

View file

@ -34,24 +34,20 @@ __FBSDID("$FreeBSD: src/lib/libarchive/test/test_pax_filename_encoding.c,v 1.1 2
* stored and restored correctly, regardless of the encodings.
*/
DEFINE_TEST(test_pax_filename_encoding)
/*
* Read a manually-created archive that has filenames that are
* stored in binary instead of UTF-8 and verify that we get
* the right filename returned and that we get a warning only
* if the header isn't marked as binary.
*/
DEFINE_TEST(test_pax_filename_encoding_1)
{
static const char testname[] = "test_pax_filename_encoding.tar.gz";
char buff[65536];
/*
* \314\214 is a valid 2-byte UTF-8 sequence.
* \374 is invalid in UTF-8.
*/
char filename[] = "abc\314\214mno\374xyz";
char longname[] = "abc\314\214mno\374xyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
;
size_t used;
struct archive *a;
struct archive_entry *entry;
@ -69,8 +65,7 @@ DEFINE_TEST(test_pax_filename_encoding)
* in it, but the header is not marked as hdrcharset=BINARY, so that
* requires a warning.
*/
failure("An invalid UTF8 pathname in a pax archive should be read\n"
" without conversion but with a warning");
failure("Invalid UTF8 in a pax archive pathname should cause a warning");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualString(filename, archive_entry_pathname(entry));
/*
@ -82,15 +77,39 @@ DEFINE_TEST(test_pax_filename_encoding)
assertEqualInt(ARCHIVE_OK, archive_read_next_header(a, &entry));
assertEqualString(filename, archive_entry_pathname(entry));
archive_read_finish(a);
}
/*
* Set the locale and write a pathname containing invalid characters.
* This should work; the underlying implementation should automatically
* fall back to storing the pathname in binary.
*/
DEFINE_TEST(test_pax_filename_encoding_2)
{
char filename[] = "abc\314\214mno\374xyz";
struct archive *a;
struct archive_entry *entry;
char buff[65536];
char longname[] = "abc\314\214mno\374xyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
"/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz"
;
size_t used;
/*
* We need a starting locale which has invalid sequences.
* de_DE.UTF-8 seems to be commonly supported.
*/
/* If it doesn't exist, just warn and return. */
failure("We need a suitable locale for the encoding tests.");
if (!assert(NULL != setlocale(LC_ALL, "de_DE.UTF-8")))
if (NULL == setlocale(LC_ALL, "de_DE.UTF-8")) {
skipping("invalid encoding tests require a suitable locale;"
" de_DE.UTF-8 not available on this system");
return;
}
assert((a = archive_write_new()) != NULL);
assertEqualIntA(a, 0, archive_write_set_format_pax(a));
@ -159,3 +178,120 @@ DEFINE_TEST(test_pax_filename_encoding)
assertEqualInt(0, archive_read_finish(a));
}
/*
* Create an entry starting from a wide-character Unicode pathname,
* read it back into "C" locale, which doesn't support the name.
* TODO: Figure out the "right" behavior here.
*/
DEFINE_TEST(test_pax_filename_encoding_3)
{
wchar_t badname[] = L"xxxAyyyBzzz";
const char badname_utf8[] = "xxx\xE1\x88\xB4yyy\xE5\x99\xB8zzz";
struct archive *a;
struct archive_entry *entry;
char buff[65536];
size_t used;
badname[3] = 0x1234;
badname[7] = 0x5678;
/* If it doesn't exist, just warn and return. */
if (NULL == setlocale(LC_ALL, "C")) {
skipping("Can't set \"C\" locale, so can't exercise "
"certain character-conversion failures");
return;
}
assert((a = archive_write_new()) != NULL);
assertEqualIntA(a, 0, archive_write_set_format_pax(a));
assertEqualIntA(a, 0, archive_write_set_compression_none(a));
assertEqualIntA(a, 0, archive_write_set_bytes_per_block(a, 0));
assertEqualInt(0,
archive_write_open_memory(a, buff, sizeof(buff), &used));
assert((entry = archive_entry_new()) != NULL);
/* Set pathname to non-convertible wide value. */
archive_entry_copy_pathname_w(entry, badname);
archive_entry_set_filetype(entry, AE_IFREG);
assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
archive_entry_free(entry);
assert((entry = archive_entry_new()) != NULL);
archive_entry_copy_pathname_w(entry, L"abc");
/* Set gname to non-convertible wide value. */
archive_entry_copy_gname_w(entry, badname);
archive_entry_set_filetype(entry, AE_IFREG);
assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
archive_entry_free(entry);
assert((entry = archive_entry_new()) != NULL);
archive_entry_copy_pathname_w(entry, L"abc");
/* Set uname to non-convertible wide value. */
archive_entry_copy_uname_w(entry, badname);
archive_entry_set_filetype(entry, AE_IFREG);
assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
archive_entry_free(entry);
assert((entry = archive_entry_new()) != NULL);
archive_entry_copy_pathname_w(entry, L"abc");
/* Set hardlink to non-convertible wide value. */
archive_entry_copy_hardlink_w(entry, badname);
archive_entry_set_filetype(entry, AE_IFREG);
assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
archive_entry_free(entry);
assert((entry = archive_entry_new()) != NULL);
archive_entry_copy_pathname_w(entry, L"abc");
/* Set symlink to non-convertible wide value. */
archive_entry_copy_symlink_w(entry, badname);
archive_entry_set_filetype(entry, AE_IFLNK);
assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
archive_entry_free(entry);
assertEqualInt(0, archive_write_close(a));
assertEqualInt(0, archive_write_finish(a));
/*
* Now read the entries back.
*/
assert((a = archive_read_new()) != NULL);
assertEqualInt(0, archive_read_support_format_tar(a));
assertEqualInt(0, archive_read_open_memory(a, buff, used));
failure("A non-convertible pathname should cause a warning.");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualWString(badname, archive_entry_pathname_w(entry));
failure("If native locale can't convert, we should get UTF-8 back.");
assertEqualString(badname_utf8, archive_entry_pathname(entry));
failure("A non-convertible gname should cause a warning.");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualWString(badname, archive_entry_gname_w(entry));
failure("If native locale can't convert, we should get UTF-8 back.");
assertEqualString(badname_utf8, archive_entry_gname(entry));
failure("A non-convertible uname should cause a warning.");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualWString(badname, archive_entry_uname_w(entry));
failure("If native locale can't convert, we should get UTF-8 back.");
assertEqualString(badname_utf8, archive_entry_uname(entry));
failure("A non-convertible hardlink should cause a warning.");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualWString(badname, archive_entry_hardlink_w(entry));
failure("If native locale can't convert, we should get UTF-8 back.");
assertEqualString(badname_utf8, archive_entry_hardlink(entry));
failure("A non-convertible symlink should cause a warning.");
assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry));
assertEqualWString(badname, archive_entry_symlink_w(entry));
assertEqualWString(NULL, archive_entry_hardlink_w(entry));
failure("If native locale can't convert, we should get UTF-8 back.");
assertEqualString(badname_utf8, archive_entry_symlink(entry));
assertEqualInt(ARCHIVE_EOF, archive_read_next_header(a, &entry));
assertEqualInt(0, archive_read_close(a));
assertEqualInt(0, archive_read_finish(a));
}

View file

@ -24,7 +24,7 @@
*/
#include "bsdtar_platform.h"
__FBSDID("$FreeBSD: src/usr.bin/tar/matching.c,v 1.11 2007/03/11 10:36:42 kientzle Exp $");
__FBSDID("$FreeBSD: src/usr.bin/tar/matching.c,v 1.12 2008/03/18 06:18:49 kientzle Exp $");
#ifdef HAVE_ERRNO_H
#include <errno.h>
@ -157,7 +157,7 @@ excluded(struct bsdtar *bsdtar, const char *pathname)
*/
if (match->matches == 0) {
match->matches++;
matching->inclusions_unmatched_count++;
matching->inclusions_unmatched_count--;
return (0);
}
/*

View file

@ -39,7 +39,7 @@
*
* The next few lines are the only differences.
*/
#define PROGRAM "bsdtar" /* Name of program we're testing. */
#define PROGRAM "bsdtar" /* Name of program being tested. */
#define ENVBASE "BSDTAR" /* Prefix for environment variables. */
#undef EXTRA_DUMP /* How to dump extra data */
/* How to generate extra version info. */
@ -55,7 +55,7 @@ __FBSDID("$FreeBSD$");
* test functions.
*/
#undef DEFINE_TEST
#define DEFINE_TEST(name) void name(void);
#define DEFINE_TEST(name) void name(void);
#include "list.h"
/* Interix doesn't define these in a standard header. */
@ -272,6 +272,30 @@ test_assert_equal_int(const char *file, int line,
return (0);
}
static void strdump(const char *p)
{
if (p == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*p != '\0') {
unsigned int c = 0xff & *p++;
switch (c) {
case '\a': fprintf(stderr, "\a"); break;
case '\b': fprintf(stderr, "\b"); break;
case '\n': fprintf(stderr, "\n"); break;
case '\r': fprintf(stderr, "\r"); break;
default:
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else
fprintf(stderr, "\\x%02X", c);
}
}
fprintf(stderr, "\"");
}
/* assertEqualString() displays the values of the two strings. */
int
test_assert_equal_string(const char *file, int line,
@ -294,12 +318,37 @@ test_assert_equal_string(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Strings not equal\n",
file, line);
fprintf(stderr, " %s = \"%s\"\n", e1, v1);
fprintf(stderr, " %s = \"%s\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
strdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
strdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}
static void wcsdump(const wchar_t *w)
{
if (w == NULL) {
fprintf(stderr, "(null)");
return;
}
fprintf(stderr, "\"");
while (*w != L'\0') {
unsigned int c = *w++;
if (c >= 32 && c < 127)
fprintf(stderr, "%c", c);
else if (c < 256)
fprintf(stderr, "\\x%02X", c);
else if (c < 0x10000)
fprintf(stderr, "\\u%04X", c);
else
fprintf(stderr, "\\U%08X", c);
}
fprintf(stderr, "\"");
}
/* assertEqualWString() displays the values of the two strings. */
int
test_assert_equal_wstring(const char *file, int line,
@ -308,7 +357,17 @@ test_assert_equal_wstring(const char *file, int line,
void *extra)
{
++assertions;
if (wcscmp(v1, v2) == 0) {
if (v1 == NULL) {
if (v2 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (v2 == NULL) {
if (v1 == NULL) {
msg[0] = '\0';
return (1);
}
} else if (wcscmp(v1, v2) == 0) {
msg[0] = '\0';
return (1);
}
@ -317,8 +376,12 @@ test_assert_equal_wstring(const char *file, int line,
return (0);
fprintf(stderr, "%s:%d: Assertion failed: Unicode strings not equal\n",
file, line);
fwprintf(stderr, L" %s = \"%ls\"\n", e1, v1);
fwprintf(stderr, L" %s = \"%ls\"\n", e2, v2);
fprintf(stderr, " %s = ", e1);
wcsdump(v1);
fprintf(stderr, "\n");
fprintf(stderr, " %s = ", e2);
wcsdump(v2);
fprintf(stderr, "\n");
report_failure(extra);
return (0);
}
@ -587,7 +650,7 @@ slurpfile(size_t * sizep, const char *fmt, ...)
* We reuse it here to define a list of all tests (functions and names).
*/
#undef DEFINE_TEST
#define DEFINE_TEST(n) { n, #n },
#define DEFINE_TEST(n) { n, #n },
struct { void (*func)(void); const char *name; } tests[] = {
#include "list.h"
};
@ -662,7 +725,7 @@ static void usage(const char *program)
exit(1);
}
#define uudecode(c) (((c) - 0x20) & 0x3f)
#define UUDECODE(c) (((c) - 0x20) & 0x3f)
void
extract_reference_file(const char *name)
@ -695,23 +758,23 @@ extract_reference_file(const char *name)
if (memcmp(buff, "end", 3) == 0)
break;
bytes = uudecode(*p++);
bytes = UUDECODE(*p++);
while (bytes > 0) {
int n = 0;
/* Write out 1-3 bytes from that. */
if (bytes > 0) {
n = uudecode(*p++) << 18;
n |= uudecode(*p++) << 12;
n = UUDECODE(*p++) << 18;
n |= UUDECODE(*p++) << 12;
fputc(n >> 16, out);
--bytes;
}
if (bytes > 0) {
n |= uudecode(*p++) << 6;
n |= UUDECODE(*p++) << 6;
fputc((n >> 8) & 0xFF, out);
--bytes;
}
if (bytes > 0) {
n |= uudecode(*p++);
n |= UUDECODE(*p++);
fputc(n & 0xFF, out);
--bytes;
}

View file

@ -76,8 +76,8 @@ static void
verify_tree(int limit)
{
struct stat st, st2;
char buff[260];
char buff2[260];
char filename[260];
char contents[260];
int i, r;
int fd;
int len;
@ -88,23 +88,25 @@ verify_tree(int limit)
/* Generate the names we know should be there and verify them. */
for (i = 0; i < 200; i++) {
/* Verify a file named "f_abcdef..." */
buff[0] = 'f';
buff[1] = '_';
buff[i + 2] = 'a' + (i % 26);
buff[i + 3] = '\0';
if (limit != LIMIT_USTAR || strlen(buff) <= 100) {
fd = open(buff, O_RDONLY);
filename[0] = 'f';
filename[1] = '_';
filename[i + 2] = 'a' + (i % 26);
filename[i + 3] = '\0';
if (limit != LIMIT_USTAR || strlen(filename) <= 100) {
fd = open(filename, O_RDONLY);
failure("Couldn't open \"%s\": %s",
buff, strerror(errno));
assert(fd >= 0);
len = read(fd, buff2, i + 10);
close(fd);
assertEqualInt(len, i + 3);
/* Verify contents of 'buff2' */
buff2[len] = '\0';
assertEqualString(buff, buff2);
/* stat() file and get dev/ino for next check */
assertEqualInt(0, lstat(buff, &st));
filename, strerror(errno));
if (assert(fd >= 0)) {
len = read(fd, contents, i + 10);
close(fd);
assertEqualInt(len, i + 3);
/* Verify contents of 'contents' */
contents[len] = '\0';
failure("Each test file contains its own name");
assertEqualString(filename, contents);
/* stat() file and get dev/ino for next check */
assertEqualInt(0, lstat(filename, &st));
}
}
/*
@ -112,18 +114,18 @@ verify_tree(int limit)
* "original/" as part of the name, so the link
* names here can't exceed 91 chars.
*/
if (limit != LIMIT_USTAR || strlen(buff) <= 91) {
if (limit != LIMIT_USTAR || strlen(filename) <= 91) {
/* Verify hardlink "l_abcdef..." */
buff[0] = 'l';
assertEqualInt(0, (r = lstat(buff, &st2)));
filename[0] = 'l';
assertEqualInt(0, (r = lstat(filename, &st2)));
if (r == 0) {
assertEqualInt(st2.st_dev, st.st_dev);
assertEqualInt(st2.st_ino, st.st_ino);
}
/* Verify hardlink "m_abcdef..." */
buff[0] = 'm';
assertEqualInt(0, (r = lstat(buff, &st2)));
filename[0] = 'm';
assertEqualInt(0, (r = lstat(filename, &st2)));
if (r == 0) {
assertEqualInt(st2.st_dev, st.st_dev);
assertEqualInt(st2.st_ino, st.st_ino);
@ -135,26 +137,36 @@ verify_tree(int limit)
* so the limit here is 100 characters.
*/
/* Verify symlink "s_abcdef..." */
buff[0] = 's';
if (limit != LIMIT_USTAR || strlen(buff) <= 100) {
filename[0] = 's';
if (limit != LIMIT_USTAR || strlen(filename) <= 100) {
/* This is a symlink. */
assertEqualInt(0, lstat(buff, &st2));
assert(S_ISLNK(st2.st_mode));
/* This is a symlink to the file above. */
assertEqualInt(0, stat(buff, &st2));
assertEqualInt(st2.st_dev, st.st_dev);
assertEqualInt(st2.st_ino, st.st_ino);
failure("Couldn't stat %s (length %d)",
filename, strlen(filename));
if (assertEqualInt(0, lstat(filename, &st2))) {
assert(S_ISLNK(st2.st_mode));
/* This is a symlink to the file above. */
failure("Couldn't stat %s", filename);
if (assertEqualInt(0, stat(filename, &st2))) {
assertEqualInt(st2.st_dev, st.st_dev);
assertEqualInt(st2.st_ino, st.st_ino);
}
}
}
/* Verify dir "d_abcdef...". */
buff[0] = 'd';
if (limit != LIMIT_USTAR || strlen(buff) <= 100) {
filename[0] = 'd';
if (limit != LIMIT_USTAR || strlen(filename) <= 100) {
/* This is a dir. */
assertEqualInt(0, lstat(buff, &st2));
assert(S_ISDIR(st2.st_mode));
/* TODO: opendir/readdir this directory and
* make sure it's empty.
*/
failure("Couldn't stat %s (length %d)",
filename, strlen(filename));
if (assertEqualInt(0, lstat(filename, &st2))) {
if (assert(S_ISDIR(st2.st_mode))) {
/* TODO: opendir/readdir this
* directory and make sure
* it's empty.
*/
}
}
}
}
@ -164,11 +176,15 @@ verify_tree(int limit)
p = de->d_name;
switch(p[0]) {
case 'l': case 'm':
if (limit == LIMIT_USTAR)
if (limit == LIMIT_USTAR) {
failure("strlen(p) = %d", strlen(p));
assert(strlen(p) <= 91);
}
case 'd': case 'f': case 's':
if (limit == LIMIT_USTAR)
if (limit == LIMIT_USTAR) {
failure("strlen(p)=%d", strlen(p));
assert(strlen(p) <= 100);
}
/* Our files have very particular filename patterns. */
assert(p[1] == '_' && p[2] == 'a');
assert(p[2] == 'a');
@ -239,7 +255,7 @@ copy_ustar(void)
*/
r = systemf("%s cf archive --format=ustar -C .. original >pack.out 2>pack.err",
testprog);
failure("Error invoking \"%s cf\"", testprog);
failure("Error invoking \"%s cf archive --format=ustar\"", testprog);
assertEqualInt(r, 0);
/* Verify that nothing went to stdout. */

View file

@ -1 +1 @@
2005000b
2005001b