Cleanup: split libtar header, move to sub directory

Some of the on-disk format internals are moved to a separate header
and some of the stuff from internal.h is moved to that format header.
C++ guards are added in addtion.

Everything PAX related is moved to pax_header.c, some internal
functions are marked as static.

Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>

15 files changed, 183 insertions, 219 deletions

diff --git a/lib/tar/Makemodule.am b/lib/tar/Makemodule.am
index 414aca1..7f93d13 100644
--- a/lib/tar/Makemodule.am
+++ b/lib/tar/Makemodule.am
@@ -1,10 +1,10 @@
 libtar_a_SOURCES = lib/tar/read_header.c lib/tar/write_header.c
 libtar_a_SOURCES += lib/tar/number.c lib/tar/checksum.c lib/tar/cleanup.c
-libtar_a_SOURCES += lib/tar/read_sparse_map.c lib/tar/read_sparse_map_old.c
-libtar_a_SOURCES += lib/tar/base64.c lib/tar/urldecode.c lib/tar/internal.h
+libtar_a_SOURCES += lib/tar/read_sparse_map_old.c
+libtar_a_SOURCES += lib/tar/internal.h
 libtar_a_SOURCES += lib/tar/padd_file.c lib/tar/record_to_memory.c
 libtar_a_SOURCES += lib/tar/pax_header.c lib/tar/read_sparse_map_new.c
-libtar_a_SOURCES += include/tar.h
+libtar_a_SOURCES += include/tar/tar.h include/tar/format.h
 libtar_a_CFLAGS = $(AM_CFLAGS)
 libtar_a_CPPFLAGS = $(AM_CPPFLAGS)
 
diff --git a/lib/tar/base64.c b/lib/tar/base64.c
deleted file mode 100644
index f88444a..0000000
--- a/lib/tar/base64.c
+++ /dev/null
@@ -1,59 +0,0 @@
-/* SPDX-License-Identifier: GPL-3.0-or-later */
-/*
- * base64.c
- *
- * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
- */
-#include "config.h"
-
-#include "internal.h"
-
-static sqfs_u8 convert(char in)
-{
-	if (isupper(in))
-		return in - 'A';
-	if (islower(in))
-		return in - 'a' + 26;
-	if (isdigit(in))
-		return in - '0' + 52;
-	if (in == '+')
-		return 62;
-	if (in == '/' || in == '-')
-		return 63;
-	return 0;
-}
-
-size_t base64_decode(sqfs_u8 *out, const char *in, size_t len)
-{
-	sqfs_u8 *start = out;
-
-	while (len > 0) {
-		unsigned int diff = 0, value = 0;
-
-		while (diff < 4 && len > 0) {
-			if (*in == '=' || *in == '_' || *in == '\0') {
-				len = 0;
-			} else {
-				value = (value << 6) | convert(*(in++));
-				--len;
-				++diff;
-			}
-		}
-
-		if (diff < 2)
-			break;
-
-		value <<= 6 * (4 - diff);
-
-		switch (diff) {
-		case 4:  out[2] = value & 0xff; /* fall-through */
-		case 3:  out[1] = (value >> 8) & 0xff; /* fall-through */
-		default: out[0] = (value >> 16) & 0xff;
-		}
-
-		out += (diff * 3) / 4;
-	}
-
-	*out = '\0';
-	return out - start;
-}
diff --git a/lib/tar/checksum.c b/lib/tar/checksum.c
index 144ba43..6541373 100644
--- a/lib/tar/checksum.c
+++ b/lib/tar/checksum.c
@@ -6,7 +6,9 @@
  */
 #include "config.h"
 
-#include "internal.h"
+#include "tar/format.h"
+
+#include <stdio.h>
 
 static unsigned int get_checksum(const tar_header_t *hdr)
 {
diff --git a/lib/tar/cleanup.c b/lib/tar/cleanup.c
index 2f814bf..9f33336 100644
--- a/lib/tar/cleanup.c
+++ b/lib/tar/cleanup.c
@@ -7,6 +7,8 @@
 #include "config.h"
 
 #include "internal.h"
+#include <stdlib.h>
+#include <string.h>
 
 void free_sparse_list(sparse_map_t *sparse)
 {
diff --git a/lib/tar/internal.h b/lib/tar/internal.h
index b7c4c34..cfbb938 100644
--- a/lib/tar/internal.h
+++ b/lib/tar/internal.h
@@ -9,13 +9,8 @@
 
 #include "config.h"
 
-#include "tar.h"
-
-#include <string.h>
-#include <stdlib.h>
-#include <limits.h>
-#include <ctype.h>
-#include <stdio.h>
+#include "tar/tar.h"
+#include "tar/format.h"
 
 enum {
 	PAX_SIZE = 0x001,
@@ -38,37 +33,10 @@ enum {
 	ETV_POSIX,
 };
 
-
-#define TAR_MAX_SYMLINK_LEN (65536)
-#define TAR_MAX_PATH_LEN (65536)
-#define TAR_MAX_PAX_LEN (65536)
-#define TAR_MAX_SPARSE_ENT (65536)
-
-
-int read_octal(const char *str, int digits, sqfs_u64 *out);
-
-int read_binary(const char *str, int digits, sqfs_u64 *out);
-
-int read_number(const char *str, int digits, sqfs_u64 *out);
-
-int pax_read_decimal(const char *str, sqfs_u64 *out);
-
-void update_checksum(tar_header_t *hdr);
-
-bool is_checksum_valid(const tar_header_t *hdr);
-
-sparse_map_t *read_sparse_map(const char *line);
-
 sparse_map_t *read_gnu_old_sparse(istream_t *fp, tar_header_t *hdr);
 
 sparse_map_t *read_gnu_new_sparse(istream_t *fp, tar_header_decoded_t *out);
 
-void free_sparse_list(sparse_map_t *sparse);
-
-size_t base64_decode(sqfs_u8 *out, const char *in, size_t len);
-
-void urldecode(char *str);
-
 char *record_to_memory(istream_t *fp, size_t size);
 
 int read_pax_header(istream_t *fp, sqfs_u64 entsize, unsigned int *set_by_pax,
diff --git a/lib/tar/number.c b/lib/tar/number.c
index 50cb658..2f179df 100644
--- a/lib/tar/number.c
+++ b/lib/tar/number.c
@@ -6,7 +6,10 @@
  */
 #include "config.h"
 
-#include "internal.h"
+#include "tar/format.h"
+
+#include <ctype.h>
+#include <stdio.h>
 
 int read_octal(const char *str, int digits, sqfs_u64 *out)
 {
@@ -31,7 +34,7 @@ int read_octal(const char *str, int digits, sqfs_u64 *out)
 	return 0;
 }
 
-int read_binary(const char *str, int digits, sqfs_u64 *out)
+static int read_binary(const char *str, int digits, sqfs_u64 *out)
 {
 	sqfs_u64 x, ov, result = 0;
 	bool first = true;
@@ -74,20 +77,3 @@ int read_number(const char *str, int digits, sqfs_u64 *out)
 
 	return read_octal(str, digits, out);
 }
-
-int pax_read_decimal(const char *str, sqfs_u64 *out)
-{
-	sqfs_u64 result = 0;
-
-	while (*str >= '0' && *str <= '9') {
-		if (result > 0xFFFFFFFFFFFFFFFFUL / 10) {
-			fputs("numeric overflow parsing pax header\n", stderr);
-			return -1;
-		}
-
-		result = (result * 10) + (*(str++) - '0');
-	}
-
-	*out = result;
-	return 0;
-}
diff --git a/lib/tar/padd_file.c b/lib/tar/padd_file.c
index 1173096..6a1ca05 100644
--- a/lib/tar/padd_file.c
+++ b/lib/tar/padd_file.c
@@ -5,7 +5,8 @@
  * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
  */
 #include "config.h"
-#include "tar.h"
+#include "tar/tar.h"
+#include "tar/format.h"
 
 #include <stdlib.h>
 #include <stdio.h>
diff --git a/lib/tar/pax_header.c b/lib/tar/pax_header.c
index d24def0..8a1b077 100644
--- a/lib/tar/pax_header.c
+++ b/lib/tar/pax_header.c
@@ -7,6 +7,105 @@
 #include "config.h"
 
 #include "internal.h"
+#include <ctype.h>
+#include <string.h>
+#include <stdlib.h>
+
+static sqfs_u8 base64_convert(char in)
+{
+	if (isupper(in))
+		return in - 'A';
+	if (islower(in))
+		return in - 'a' + 26;
+	if (isdigit(in))
+		return in - '0' + 52;
+	if (in == '+')
+		return 62;
+	if (in == '/' || in == '-')
+		return 63;
+	return 0;
+}
+
+static int xdigit(int x)
+{
+	if (isupper(x))
+		return x - 'A' + 0x0A;
+	if (islower(x))
+		return x - 'a' + 0x0A;
+	return x - '0';
+}
+
+static size_t base64_decode(sqfs_u8 *out, const char *in, size_t len)
+{
+	sqfs_u8 *start = out;
+
+	while (len > 0) {
+		unsigned int diff = 0, value = 0;
+
+		while (diff < 4 && len > 0) {
+			if (*in == '=' || *in == '_' || *in == '\0') {
+				len = 0;
+			} else {
+				value = (value << 6) | base64_convert(*(in++));
+				--len;
+				++diff;
+			}
+		}
+
+		if (diff < 2)
+			break;
+
+		value <<= 6 * (4 - diff);
+
+		switch (diff) {
+		case 4:  out[2] = value & 0xff; /* fall-through */
+		case 3:  out[1] = (value >> 8) & 0xff; /* fall-through */
+		default: out[0] = (value >> 16) & 0xff;
+		}
+
+		out += (diff * 3) / 4;
+	}
+
+	*out = '\0';
+	return out - start;
+}
+
+static int pax_read_decimal(const char *str, sqfs_u64 *out)
+{
+	sqfs_u64 result = 0;
+
+	while (*str >= '0' && *str <= '9') {
+		if (result > 0xFFFFFFFFFFFFFFFFUL / 10) {
+			fputs("numeric overflow parsing pax header\n", stderr);
+			return -1;
+		}
+
+		result = (result * 10) + (*(str++) - '0');
+	}
+
+	*out = result;
+	return 0;
+}
+
+static void urldecode(char *str)
+{
+	unsigned char *out = (unsigned char *)str;
+	char *in = str;
+	int x;
+
+	while (*in != '\0') {
+		x = *(in++);
+
+		if (x == '%' && isxdigit(in[0]) && isxdigit(in[1])) {
+			x = xdigit(*(in++)) << 4;
+			x |= xdigit(*(in++));
+		}
+
+		*(out++) = x;
+	}
+
+	*out = '\0';
+}
 
 static int pax_uid(tar_header_decoded_t *out, sqfs_u64 id)
 {
@@ -52,6 +151,55 @@ static int pax_slink(tar_header_decoded_t *out, char *path)
 	return 0;
 }
 
+static int pax_sparse_map(tar_header_decoded_t *out, const char *line)
+{
+	sparse_map_t *last = NULL, *list = NULL, *ent = NULL;
+
+	free_sparse_list(out->sparse);
+	out->sparse = NULL;
+
+	do {
+		ent = calloc(1, sizeof(*ent));
+		if (ent == NULL)
+			goto fail_errno;
+
+		if (pax_read_decimal(line, &ent->offset))
+			goto fail_format;
+
+		while (isdigit(*line))
+			++line;
+
+		if (*(line++) != ',')
+			goto fail_format;
+
+		if (pax_read_decimal(line, &ent->count))
+			goto fail_format;
+
+		while (isdigit(*line))
+			++line;
+
+		if (last == NULL) {
+			list = last = ent;
+		} else {
+			last->next = ent;
+			last = ent;
+		}
+	} while (*(line++) == ',');
+
+	out->sparse = list;
+	return 0;
+fail_errno:
+	perror("parsing GNU pax sparse file record");
+	goto fail;
+fail_format:
+	fputs("malformed GNU pax sparse file record\n", stderr);
+	goto fail;
+fail:
+	free_sparse_list(list);
+	free(ent);
+	return -1;
+}
+
 static int pax_xattr_schily(tar_header_decoded_t *out,
 			    tar_xattr_t *xattr)
 {
@@ -76,6 +224,7 @@ enum {
 	PAX_TYPE_SINT = 0,
 	PAX_TYPE_UINT,
 	PAX_TYPE_STRING,
+	PAX_TYPE_CONST_STRING,
 	PAX_TYPE_PREFIXED_XATTR,
 	PAX_TYPE_IGNORE,
 };
@@ -88,6 +237,7 @@ static const struct pax_handler_t {
 		int (*sint)(tar_header_decoded_t *out, sqfs_s64 sval);
 		int (*uint)(tar_header_decoded_t *out, sqfs_u64 uval);
 		int (*str)(tar_header_decoded_t *out, char *str);
+		int (*cstr)(tar_header_decoded_t *out, const char *str);
 		int (*xattr)(tar_header_decoded_t *out, tar_xattr_t *xattr);
 	} cb;
 } pax_fields[] = {
@@ -110,6 +260,8 @@ static const struct pax_handler_t {
 	  { .xattr = pax_xattr_schily } },
 	{ "LIBARCHIVE.xattr", 0, PAX_TYPE_PREFIXED_XATTR,
 	  { .xattr = pax_xattr_libarchive } },
+	{ "GNU.sparse.map", 0, PAX_TYPE_CONST_STRING,
+	  { .cstr = pax_sparse_map } },
 };
 
 static const struct pax_handler_t *find_handler(const char *key)
@@ -183,6 +335,8 @@ static int apply_handler(tar_header_decoded_t *out,
 		if (pax_read_decimal(value, &uval))
 			return -1;
 		return field->cb.uint(out, uval);
+	case PAX_TYPE_CONST_STRING:
+		return field->cb.cstr(out, value);
 	case PAX_TYPE_STRING:
 		copy = strdup(value);
 		if (copy == NULL) {
@@ -264,13 +418,6 @@ int read_pax_header(istream_t *fp, sqfs_u64 entsize, unsigned int *set_by_pax,
 			}
 
 			*set_by_pax |= field->flag;
-		} else if (!strcmp(key, "GNU.sparse.map")) {
-			free_sparse_list(out->sparse);
-			sparse_last = NULL;
-
-			out->sparse = read_sparse_map(value);
-			if (out->sparse == NULL)
-				goto fail;
 		} else if (!strcmp(key, "GNU.sparse.offset")) {
 			if (pax_read_decimal(value, &offset))
 				goto fail;
diff --git a/lib/tar/read_header.c b/lib/tar/read_header.c
index d1790de..ea4873b 100644
--- a/lib/tar/read_header.c
+++ b/lib/tar/read_header.c
@@ -7,6 +7,8 @@
 #include "config.h"
 
 #include "internal.h"
+#include <string.h>
+#include <stdlib.h>
 
 static bool is_zero_block(const tar_header_t *hdr)
 {
diff --git a/lib/tar/read_sparse_map.c b/lib/tar/read_sparse_map.c
deleted file mode 100644
index 0779b96..0000000
--- a/lib/tar/read_sparse_map.c
+++ /dev/null
@@ -1,54 +0,0 @@
-/* SPDX-License-Identifier: GPL-3.0-or-later */
-/*
- * read_sparse_map.c
- *
- * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
- */
-#include "config.h"
-
-#include "internal.h"
-
-sparse_map_t *read_sparse_map(const char *line)
-{
-	sparse_map_t *last = NULL, *list = NULL, *ent = NULL;
-
-	do {
-		ent = calloc(1, sizeof(*ent));
-		if (ent == NULL)
-			goto fail_errno;
-
-		if (pax_read_decimal(line, &ent->offset))
-			goto fail_format;
-
-		while (isdigit(*line))
-			++line;
-
-		if (*(line++) != ',')
-			goto fail_format;
-
-		if (pax_read_decimal(line, &ent->count))
-			goto fail_format;
-
-		while (isdigit(*line))
-			++line;
-
-		if (last == NULL) {
-			list = last = ent;
-		} else {
-			last->next = ent;
-			last = ent;
-		}
-	} while (*(line++) == ',');
-
-	return list;
-fail_errno:
-	perror("parsing GNU pax sparse file record");
-	goto fail;
-fail_format:
-	fputs("malformed GNU pax sparse file record\n", stderr);
-	goto fail;
-fail:
-	free_sparse_list(list);
-	free(ent);
-	return NULL;
-}
diff --git a/lib/tar/read_sparse_map_new.c b/lib/tar/read_sparse_map_new.c
index d9f6c6e..de1b6a4 100644
--- a/lib/tar/read_sparse_map_new.c
+++ b/lib/tar/read_sparse_map_new.c
@@ -5,9 +5,12 @@
  * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
  */
 #include "config.h"
-
 #include "internal.h"
 
+#include <ctype.h>
+#include <string.h>
+#include <stdlib.h>
+
 static int decode(const char *str, size_t len, size_t *out)
 {
 	size_t count = 0;
diff --git a/lib/tar/read_sparse_map_old.c b/lib/tar/read_sparse_map_old.c
index cd7177d..3dd3300 100644
--- a/lib/tar/read_sparse_map_old.c
+++ b/lib/tar/read_sparse_map_old.c
@@ -5,9 +5,11 @@
  * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
  */
 #include "config.h"
-
 #include "internal.h"
 
+#include <ctype.h>
+#include <stdlib.h>
+
 sparse_map_t *read_gnu_old_sparse(istream_t *fp, tar_header_t *hdr)
 {
 	sparse_map_t *list = NULL, *end = NULL, *node;
diff --git a/lib/tar/record_to_memory.c b/lib/tar/record_to_memory.c
index 822608c..ba422de 100644
--- a/lib/tar/record_to_memory.c
+++ b/lib/tar/record_to_memory.c
@@ -6,8 +6,9 @@
  */
 #include "config.h"
 
-#include "tar.h"
+#include "tar/tar.h"
 #include "internal.h"
+#include <stdlib.h>
 
 char *record_to_memory(istream_t *fp, size_t size)
 {
diff --git a/lib/tar/urldecode.c b/lib/tar/urldecode.c
deleted file mode 100644
index 6fac4d3..0000000
--- a/lib/tar/urldecode.c
+++ /dev/null
@@ -1,38 +0,0 @@
-/* SPDX-License-Identifier: GPL-3.0-or-later */
-/*
- * urldecode.c
- *
- * Copyright (C) 2019 David Oberhollenzer <goliath@infraroot.at>
- */
-#include "config.h"
-
-#include "internal.h"
-
-static int xdigit(int x)
-{
-	if (isupper(x))
-		return x - 'A' + 0x0A;
-	if (islower(x))
-		return x - 'a' + 0x0A;
-	return x - '0';
-}
-
-void urldecode(char *str)
-{
-	unsigned char *out = (unsigned char *)str;
-	char *in = str;
-	int x;
-
-	while (*in != '\0') {
-		x = *(in++);
-
-		if (x == '%' && isxdigit(in[0]) && isxdigit(in[1])) {
-			x = xdigit(*(in++)) << 4;
-			x |= xdigit(*(in++));
-		}
-
-		*(out++) = x;
-	}
-
-	*out = '\0';
-}
diff --git a/lib/tar/write_header.c b/lib/tar/write_header.c
index a337b74..b0711b3 100644
--- a/lib/tar/write_header.c
+++ b/lib/tar/write_header.c
@@ -7,6 +7,7 @@
 #include "config.h"
 
 #include "internal.h"
+#include <string.h>
 
 static void write_binary(char *dst, sqfs_u64 value, int digits)
 {