15 files changed, 69 insertions, 42 deletions

diff --git a/archival/bzip2.c b/archival/bzip2.c
index f1c84d681..fdb8b9306 100644
--- a/archival/bzip2.c
+++ b/archival/bzip2.c
@@ -88,7 +88,7 @@ IF_DESKTOP(long long) int bz_write(bz_stream *strm, void* rbuf, ssize_t rlen, vo
                         if (n2 != n) {
                                 if (n2 >= 0)
                                         errno = 0; /* prevent bogus error message */
-                                bb_perror_msg(n2 >= 0 ? "short write" : "write error");
+                                bb_perror_msg(n2 >= 0 ? "short write" : bb_msg_write_error);
                                 return -1;
                         }
                 }
@@ -118,7 +118,7 @@ IF_DESKTOP(long long) int FAST_FUNC compressStream(unpack_info_t *info UNUSED_PA
         while (1) {
                 count = full_read(STDIN_FILENO, rbuf, IOBUF_SIZE);
                 if (count < 0) {
-                        bb_perror_msg("read error");
+                        bb_perror_msg(bb_msg_read_error);
                         total = -1;
                         break;
                 }
diff --git a/archival/libunarchive/decompress_unxz.c b/archival/libunarchive/decompress_unxz.c
index 3f9392984..1302e29fb 100644
--- a/archival/libunarchive/decompress_unxz.c
+++ b/archival/libunarchive/decompress_unxz.c
@@ -86,7 +86,7 @@ unpack_xz_stream(int src_fd, int dst_fd)
                 if (rd) {
                         rd = safe_read(src_fd, membuf + insz, rd);
                         if (rd < 0) {
-                                bb_error_msg("read error");
+                                bb_error_msg(bb_msg_read_error);
                                 total = -1;
                                 break;
                         }
diff --git a/archival/libunarchive/decompress_unzip.c b/archival/libunarchive/decompress_unzip.c
index 33e877ec8..bccd0262e 100644
--- a/archival/libunarchive/decompress_unzip.c
+++ b/archival/libunarchive/decompress_unzip.c
@@ -1072,7 +1072,7 @@ static int top_up(STATE_PARAM unsigned n)
                 bytebuffer_offset = 0;
                 bytebuffer_size = full_read(gunzip_src_fd, &bytebuffer[count], bytebuffer_max - count);
                 if ((int)bytebuffer_size < 0) {
-                        bb_error_msg("read error");
+                        bb_error_msg(bb_msg_read_error);
                         return 0;
                 }
                 bytebuffer_size += count;
diff --git a/archival/rpm2cpio.c b/archival/rpm2cpio.c
index 598ec8670..1c67dcc6e 100644
--- a/archival/rpm2cpio.c
+++ b/archival/rpm2cpio.c
@@ -68,26 +68,31 @@ int rpm2cpio_main(int argc UNUSED_PARAM, char **argv)
 #else
         /* BLOAT */
         {
-                unsigned char magic[8];
+                union {
+                        uint8_t b[4];
+                        uint16_t b16[2];
+                        uint32_t b32[1];
+                } magic;
                 IF_DESKTOP(long long) int FAST_FUNC (*unpack)(int src_fd, int dst_fd);
 
-                xread(rpm_fd, &magic, 2);
-                if (magic[0] == 0x1f && magic[1] == 0x8b) {
+                xread(rpm_fd, magic.b16, sizeof(magic.b16));
+                if (magic.b16[0] == GZIP_MAGIC) {
                         unpack = unpack_gz_stream;
                 } else
                 if (ENABLE_FEATURE_SEAMLESS_BZ2
-                 && magic[0] == 'B' && magic[1] == 'Z'
+                 && magic.b16[0] == BZIP2_MAGIC
                 ) {
                         unpack = unpack_bz2_stream;
                 } else
                 if (ENABLE_FEATURE_SEAMLESS_XZ
-                 && magic[0] == 0xfd && magic[1] == '7'
+                 && magic.b16[0] == XZ_MAGIC1
                 ) {
                         /* .xz signature: 0xfd, '7', 'z', 'X', 'Z', 0x00 */
                         /* More info at: http://tukaani.org/xz/xz-file-format.txt */
-                        xread(rpm_fd, magic + 2, 4);
-                        if (strcmp((char*)magic + 2, "zXZ") != 0)
+                        xread(rpm_fd, magic.b32, sizeof(magic.b32));
+                        if (magic.b32[0] != XZ_MAGIC2)
                                 goto no_magic;
+                        /* unpack_xz_stream wants fd at position 0 */
                         xlseek(rpm_fd, -6, SEEK_CUR);
                         unpack = unpack_xz_stream;
                 } else {
diff --git a/coreutils/head.c b/coreutils/head.c
index 75189eda8..cc2837411 100644
--- a/coreutils/head.c
+++ b/coreutils/head.c
@@ -27,7 +27,7 @@ static const struct suffix_mult head_suffixes[] = {
         { "", 0 }
 };
 
-static const char header_fmt_str[] ALIGN1 = "\n==> %s <==\n";
+#define header_fmt_str "\n==> %s <==\n"
 
 int head_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
 int head_main(int argc, char **argv)
diff --git a/coreutils/od_bloaty.c b/coreutils/od_bloaty.c
index cb97ff8c5..4c6b64d5e 100644
--- a/coreutils/od_bloaty.c
+++ b/coreutils/od_bloaty.c
@@ -509,7 +509,7 @@ check_and_close(void)
         }
 
         if (ferror(stdout)) {
-                bb_error_msg("write error");
+                bb_error_msg(bb_msg_write_error);
                 ioerror = 1;
         }
 }
diff --git a/coreutils/tail.c b/coreutils/tail.c
index 8fc22f562..6397702a0 100644
--- a/coreutils/tail.c
+++ b/coreutils/tail.c
@@ -66,7 +66,7 @@ static ssize_t tail_read(int fd, char *buf, size_t count)
         return r;
 }
 
-static const char header_fmt[] ALIGN1 = "\n==> %s <==\n";
+#define header_fmt_str "\n==> %s <==\n"
 
 static unsigned eat_num(const char *p)
 {
@@ -166,7 +166,7 @@ int tail_main(int argc, char **argv)
         tailbuf = xmalloc(tailbufsize);
 
         /* tail the files */
-        fmt = header_fmt + 1; /* skip header leading newline on first output */
+        fmt = header_fmt_str + 1; /* skip header leading newline on first output */
         i = 0;
         do {
                 char *buf;
@@ -181,7 +181,7 @@ int tail_main(int argc, char **argv)
 
                 if (nfiles > header_threshhold) {
                         tail_xprint_header(fmt, argv[i]);
-                        fmt = header_fmt;
+                        fmt = header_fmt_str;
                 }
 
                 if (!from_top) {
@@ -333,7 +333,7 @@ int tail_main(int argc, char **argv)
                         if (ENABLE_FEATURE_FANCY_TAIL && fd < 0)
                                 continue;
                         if (nfiles > header_threshhold) {
-                                fmt = header_fmt;
+                                fmt = header_fmt_str;
                         }
                         while ((nread = tail_read(fd, tailbuf, BUFSIZ)) > 0) {
                                 if (fmt) {
diff --git a/coreutils/test.c b/coreutils/test.c
index 3115ce6e7..0b9322316 100644
--- a/coreutils/test.c
+++ b/coreutils/test.c
@@ -393,7 +393,7 @@ static number_t getn(const char *s)
         if (errno != 0)
                 syntax(s, "out of range");
 
-        if (*(skip_whitespace(p)))
+        if (p == s || *(skip_whitespace(p)) != '\0')
                 syntax(s, "bad number");
 
         return r;
diff --git a/docs/keep_data_small.txt b/docs/keep_data_small.txt
index 079bac3a6..01c0d3c7b 100644
--- a/docs/keep_data_small.txt
+++ b/docs/keep_data_small.txt
@@ -145,6 +145,11 @@ one of above methods is not worth the resulting code obfuscation.
 If you have less than ~300 bytes of global data - don't bother.
 
 
+                Finding non-shared duplicated strings
+
+strings busybox | sort | uniq -c | sort -nr
+
+
                 gcc's data alignment problem
 
 The following attribute added in vi.c:
diff --git a/include/unarchive.h b/include/unarchive.h
index aa7ecec55..8009de282 100644
--- a/include/unarchive.h
+++ b/include/unarchive.h
@@ -4,6 +4,22 @@
 
 PUSH_AND_SET_FUNCTION_VISIBILITY_TO_HIDDEN
 
+enum {
+#if BB_BIG_ENDIAN
+        COMPRESS_MAGIC = 0x1f9d,
+        GZIP_MAGIC = 0x1f8b,
+        BZIP2_MAGIC = ('B'<<8) + 'Z',
+        XZ_MAGIC1 = (0xfd<<8) + '7',
+        XZ_MAGIC2 = ((((('z'<<8) + 'X')<<8) + 'Z')<<8) + 0,
+#else
+        COMPRESS_MAGIC = 0x9d1f,
+        GZIP_MAGIC = 0x8b1f,
+        BZIP2_MAGIC = ('Z'<<8) + 'B',
+        XZ_MAGIC1 = ('7'<<8) + 0xfd,
+        XZ_MAGIC2 = (((((0<<8) + 'Z')<<8) + 'X')<<8) + 'z',
+#endif
+};
+
 typedef struct file_header_t {
         char *name;
         char *link_target;
diff --git a/libbb/read.c b/libbb/read.c
index cd6bbeb13..b1eb3f24b 100644
--- a/libbb/read.c
+++ b/libbb/read.c
@@ -312,7 +312,11 @@ void* FAST_FUNC xmalloc_xopen_read_close(const char *filename, size_t *maxsz_p)
 void FAST_FUNC setup_unzip_on_fd(int fd /*, int fail_if_not_detected*/)
 {
         const int fail_if_not_detected = 1;
-        unsigned char magic[8];
+        union {
+                uint8_t b[4];
+                uint16_t b16[2];
+                uint32_t b32[1];
+        } magic;
         int offset = -2;
 # if BB_MMU
         IF_DESKTOP(long long) int FAST_FUNC (*xformer)(int src_fd, int dst_fd);
@@ -324,9 +328,9 @@ void FAST_FUNC setup_unzip_on_fd(int fd /*, int fail_if_not_detected*/)
 
         /* .gz and .bz2 both have 2-byte signature, and their
          * unpack_XXX_stream wants this header skipped. */
-        xread(fd, magic, 2);
+        xread(fd, magic.b16, sizeof(magic.b16));
         if (ENABLE_FEATURE_SEAMLESS_GZ
-         && magic[0] == 0x1f && magic[1] == 0x8b
+         && magic.b16[0] == GZIP_MAGIC
         ) {
 # if BB_MMU
                 xformer = unpack_gz_stream;
@@ -336,7 +340,7 @@ void FAST_FUNC setup_unzip_on_fd(int fd /*, int fail_if_not_detected*/)
                 goto found_magic;
         }
         if (ENABLE_FEATURE_SEAMLESS_BZ2
-         && magic[0] == 'B' && magic[1] == 'Z'
+         && magic.b16[0] == BZIP2_MAGIC
         ) {
 # if BB_MMU
                 xformer = unpack_bz2_stream;
@@ -346,19 +350,20 @@ void FAST_FUNC setup_unzip_on_fd(int fd /*, int fail_if_not_detected*/)
                 goto found_magic;
         }
         if (ENABLE_FEATURE_SEAMLESS_XZ
-         && magic[0] == 0xfd && magic[1] == '7'
+         && magic.b16[0] == XZ_MAGIC1
         ) {
                 /* .xz signature: 0xfd, '7', 'z', 'X', 'Z', 0x00 */
                 /* More info at: http://tukaani.org/xz/xz-file-format.txt */
                 offset = -6;
-                xread(fd, magic + 2, 4);
-                if (strcmp((char*)magic + 2, "zXZ") == 0) {
+                xread(fd, magic.b32, sizeof(magic.b32));
+                if (magic.b32[0] == XZ_MAGIC2) {
 # if BB_MMU
                         xformer = unpack_xz_stream;
+                        /* unpack_xz_stream wants fd at position 0 */
+                        xlseek(fd, offset, SEEK_CUR);
 # else
                         xformer_prog = "unxz";
 # endif
-                        xlseek(fd, offset, SEEK_CUR);
                         goto found_magic;
                 }
         }
diff --git a/miscutils/less.c b/miscutils/less.c
index 848266212..da2cd07af 100644
--- a/miscutils/less.c
+++ b/miscutils/less.c
@@ -430,7 +430,7 @@ static void read_lines(void)
                                          * immediately */
                                         eof_error = 1;
                                 } else {
-                                        print_statusline("read error");
+                                        print_statusline(bb_msg_read_error);
                                 }
                         }
 #if !ENABLE_FEATURE_LESS_REGEXP
diff --git a/modutils/modutils-24.c b/modutils/modutils-24.c
index 5c1611c29..7f39e253d 100644
--- a/modutils/modutils-24.c
+++ b/modutils/modutils-24.c
@@ -3529,20 +3529,18 @@ static void set_tainted(int fd, const char *m_name,
 /* Check if loading this module will taint the kernel. */
 static void check_tainted_module(struct obj_file *f, const char *m_name)
 {
-        static const char tainted_file[] ALIGN1 = TAINT_FILENAME;
-
         int fd, kernel_has_tainted;
         const char *ptr;
 
         kernel_has_tainted = 1;
-        fd = open(tainted_file, O_RDWR);
+        fd = open(TAINT_FILENAME, O_RDWR);
         if (fd < 0) {
                 if (errno == ENOENT)
                         kernel_has_tainted = 0;
                 else if (errno == EACCES)
                         kernel_has_tainted = 1;
                 else {
-                        perror(tainted_file);
+                        perror(TAINT_FILENAME);
                         kernel_has_tainted = 0;
                 }
         }
diff --git a/shell/ash.c b/shell/ash.c
index 54a831f6d..913f40f12 100644
--- a/shell/ash.c
+++ b/shell/ash.c
@@ -168,9 +168,7 @@ enum { NOPTS = ARRAY_SIZE(optletters_optnames) };
 
 /* ============ Misc data */
 
-static const char homestr[] ALIGN1 = "HOME";
-static const char snlfmt[] ALIGN1 = "%s\n";
-static const char msg_illnum[] ALIGN1 = "Illegal number: %s";
+#define msg_illnum "Illegal number: %s"
 
 /*
  * We enclose jmp_buf in a structure so that we can declare pointers to
@@ -2674,7 +2672,7 @@ cdcmd(int argc UNUSED_PARAM, char **argv UNUSED_PARAM)
         flags = cdopt();
         dest = *argptr;
         if (!dest)
-                dest = bltinlookup(homestr);
+                dest = bltinlookup("HOME");
         else if (LONE_DASH(dest)) {
                 dest = bltinlookup("OLDPWD");
                 flags |= CD_PRINT;
@@ -2721,7 +2719,7 @@ cdcmd(int argc UNUSED_PARAM, char **argv UNUSED_PARAM)
         /* NOTREACHED */
  out:
         if (flags & CD_PRINT)
-                out1fmt(snlfmt, curdir);
+                out1fmt("%s\n", curdir);
         return 0;
 }
 
@@ -2737,7 +2735,7 @@ pwdcmd(int argc UNUSED_PARAM, char **argv UNUSED_PARAM)
                         setpwd(dir, 0);
                 dir = physdir;
         }
-        out1fmt(snlfmt, dir);
+        out1fmt("%s\n", dir);
         return 0;
 }
 
@@ -5915,7 +5913,7 @@ exptilde(char *startp, char *p, int flags)
  done:
         *p = '\0';
         if (*name == '\0') {
-                home = lookupvar(homestr);
+                home = lookupvar("HOME");
         } else {
                 if (ENABLE_PLATFORM_MINGW32)
                         goto lose;
@@ -10322,7 +10320,7 @@ chkmail(void)
                 }
                 if (!mail_var_path_changed && statb.st_mtime != *mtp) {
                         fprintf(
-                                stderr, snlfmt,
+                                stderr, "%s\n",
                                 pathopt ? pathopt : "you have mail"
                         );
                 }
diff --git a/util-linux/fdisk.c b/util-linux/fdisk.c
index 8e133178d..7227a829e 100644
--- a/util-linux/fdisk.c
+++ b/util-linux/fdisk.c
@@ -107,9 +107,9 @@ struct partition {
         unsigned char size4[4];         /* nr of sectors in partition */
 } PACKED;
 
-static const char unable_to_open[] ALIGN1 = "can't open '%s'";
-static const char unable_to_read[] ALIGN1 = "can't read from %s";
-static const char unable_to_seek[] ALIGN1 = "can't seek on %s";
+#define unable_to_open "can't open '%s'"
+#define unable_to_read "can't read from %s"
+#define unable_to_seek "can't seek on %s"
 
 enum label_type {
         LABEL_DOS, LABEL_SUN, LABEL_SGI, LABEL_AIX, LABEL_OSF