MFH Update fileinfo to libmagic 5.00 and remove dependency on dirent.…

…h on Windows

ext/fileinfo/config.m4

@@ -8,9 +8,10 @@ if test "$PHP_FILEINFO" != "no"; then
 
   libmagic_sources=" \
     libmagic/apprentice.c libmagic/apptype.c libmagic/ascmagic.c \
-    libmagic/compress.c libmagic/fsmagic.c libmagic/funcs.c \
+    libmagic/cdf.c libmagic/cdf_time.c libmagic/compress.c \
+    libmagic/encoding.c libmagic/fsmagic.c libmagic/funcs.c \
     libmagic/is_tar.c libmagic/magic.c libmagic/print.c \
-    libmagic/readelf.c libmagic/softmagic.c"
+    libmagic/readcdf.c libmagic/readelf.c libmagic/softmagic.c"
 
   PHP_NEW_EXTENSION(fileinfo, fileinfo.c $libmagic_sources, $ext_shared,,-I@ext_srcdir@/libmagic)
   PHP_ADD_BUILD_DIR($ext_builddir/libmagic)

ext/fileinfo/config.w32

@@ -5,22 +5,16 @@ ARG_ENABLE("fileinfo", "fileinfo support", "no");
 
 if (PHP_FILEINFO != 'no') {
 
-		if (CHECK_HEADER_ADD_INCLUDE("dirent.h", "CFLAGS_FILEINFO") && 
-		CHECK_LIB("dirent_a.lib", "fileinfo", PHP_FILEINFO)) {
+	LIBMAGIC_SOURCES=" apprentice.c apptype.c ascmagic.c \
+			cdf.c cdf_time.c compress.c \
+			encoding.c fsmagic.c funcs.c \
+			is_tar.c magic.c print.c \
+			readcdf.c readelf.c softmagic.c";
 
-		LIBMAGIC_SOURCES=" apprentice.c apptype.c ascmagic.c \
-			compress.c fsmagic.c funcs.c is_tar.c \
-			magic.c print.c \
-			readelf.c softmagic.c";
-
-		if (VCVERS <1500) {
-			ADD_FLAG('CFLAGS', '/Zm1000');
-		}
-
-		EXTENSION('fileinfo', 'fileinfo.c', true, "/I" + configure_module_dirname + "/libmagic /I" + configure_module_dirname);
-		ADD_SOURCES(configure_module_dirname + '\\libmagic', LIBMAGIC_SOURCES, "fileinfo");
-	} else {
-		WARNING("fileinfo not enabled; libraries and headers not found");
-		PHP_FILEINFO = "no";
+	if (VCVERS < 1500) {
+		ADD_FLAG('CFLAGS', '/Zm1000');
 	}
+
+	EXTENSION('fileinfo', 'fileinfo.c', null, "/I" + configure_module_dirname + "/libmagic /I" + configure_module_dirname);
+	ADD_SOURCES(configure_module_dirname + '\\libmagic', LIBMAGIC_SOURCES, "fileinfo");
 }

ext/fileinfo/fileinfo.c

@@ -229,6 +229,8 @@ PHP_MINIT_FUNCTION(finfo)
 	REGISTER_LONG_CONSTANT("FILEINFO_NONE",			MAGIC_NONE, CONST_CS|CONST_PERSISTENT);
 	REGISTER_LONG_CONSTANT("FILEINFO_SYMLINK",		MAGIC_SYMLINK, CONST_CS|CONST_PERSISTENT);
 	REGISTER_LONG_CONSTANT("FILEINFO_MIME",			MAGIC_MIME, CONST_CS|CONST_PERSISTENT);
+	REGISTER_LONG_CONSTANT("FILEINFO_MIME_TYPE",	MAGIC_MIME_TYPE, CONST_CS|CONST_PERSISTENT);
+	REGISTER_LONG_CONSTANT("FILEINFO_MIME_ENCODING",MAGIC_MIME_ENCODING, CONST_CS|CONST_PERSISTENT);
 /*	REGISTER_LONG_CONSTANT("FILEINFO_COMPRESS",		MAGIC_COMPRESS, CONST_CS|CONST_PERSISTENT); disabled, as it does fork now */
 	REGISTER_LONG_CONSTANT("FILEINFO_DEVICES",		MAGIC_DEVICES, CONST_CS|CONST_PERSISTENT);
 	REGISTER_LONG_CONSTANT("FILEINFO_CONTINUE",		MAGIC_CONTINUE, CONST_CS|CONST_PERSISTENT);

ext/fileinfo/libmagic/apprentice.c

@@ -32,6 +32,11 @@
 #include "php.h"
 
 #include "file.h"
+
+#ifndef	lint
+FILE_RCSID("@(#)$File: apprentice.c,v 1.132 2008/03/28 18:19:30 christos Exp $")
+#endif	/* lint */
+
 #include "magic.h"
 #include "patchlevel.h"
 #include <stdlib.h>
@@ -53,19 +58,10 @@
 #include <assert.h>
 #include <ctype.h>
 #include <fcntl.h>
-#include <sys/stat.h>
-#ifndef PHP_WIN32
-#include <sys/param.h>
-#endif
-#include <sys/types.h>
 #ifndef PHP_WIN32
 #include <dirent.h>
 #endif
 
-#ifndef	lint
-FILE_RCSID("@(#)$File: apprentice.c,v 1.132 2008/03/28 18:19:30 christos Exp $")
-#endif	/* lint */
-
 #define	EATAB {while (isascii((unsigned char) *l) && \
 		      isspace((unsigned char) *l))  ++l;}
 #define LOWCASE(l) (isupper((unsigned char) (l)) ? \
@@ -120,7 +116,7 @@ private void bs1(struct magic *);
 private uint16_t swap2(uint16_t);
 private uint32_t swap4(uint32_t);
 private uint64_t swap8(uint64_t);
-private void mkdbname(const char *, char **, int);
+private char *mkdbname(struct magic_set *, const char *, int);
 private int apprentice_map(struct magic_set *, struct magic **, uint32_t *,
     const char *);
 private int apprentice_compile(struct magic_set *, struct magic **, uint32_t *,
@@ -129,8 +125,8 @@ private int check_format_type(const char *, int);
 private int check_format(struct magic_set *, struct magic *);
 private int get_op(char);
 private int parse_mime(struct magic_set *, struct magic_entry *, const char *);
-private int parse_strength(struct magic_set *, struct magic_entry *,
-    const char *);
+private int parse_strength(struct magic_set *, struct magic_entry *, const char *);
+private int parse_apple(struct magic_set *, struct magic_entry *, const char *);
 
 private size_t maxmagic = 0;
 private size_t magicsize = sizeof(struct magic);
@@ -143,6 +139,7 @@ private struct {
 } bang[] = {
 #define	DECLARE_FIELD(name) { # name, sizeof(# name) - 1, parse_ ## name }
 	DECLARE_FIELD(mime),
+	DECLARE_FIELD(apple),
 	DECLARE_FIELD(strength),
 #undef	DECLARE_FIELD
 	{ NULL, 0, NULL }
@@ -196,6 +193,9 @@ static const struct type_tbl_s {
 	{ XX("double"),		FILE_DOUBLE,		FILE_FMT_DOUBLE },
 	{ XX("bedouble"),	FILE_BEDOUBLE,		FILE_FMT_DOUBLE },
 	{ XX("ledouble"),	FILE_LEDOUBLE,		FILE_FMT_DOUBLE },
+	{ XX("leid3"),		FILE_LEID3,		FILE_FMT_NUM },
+	{ XX("beid3"),		FILE_BEID3,		FILE_FMT_NUM },
+	{ XX("indirect"),	FILE_INDIRECT,		FILE_FMT_NONE },
 	{ XX_NULL,		FILE_INVALID,		FILE_FMT_NONE },
 # undef XX
 # undef XX_NULL
@@ -569,7 +569,8 @@ set_test_type(struct magic *mstart, struct magic *m)
 	case FILE_REGEX:
 	case FILE_SEARCH:
 		/* binary test if pattern is not text */
-		if (file_looks_utf8(m->value.us, m->vallen, NULL, NULL) <= 0)
+		if (file_looks_utf8(m->value.us, (size_t)m->vallen, NULL,
+		    NULL) <= 0)
 			mstart->flag |= BINTEST;
 		break;
 	case FILE_DEFAULT:
@@ -691,6 +692,8 @@ apprentice_load(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp,
 		(void)fprintf(stderr, "%s\n", usg_hdr);
 
 	/* load directory or file */
+        /* FIXME: Read file names and sort them to prevent
+           non-determinism. See Debian bug #488562. */
 	if (php_sys_stat(fn, &st) == 0 && S_ISDIR(st.st_mode)) {
 		dir = opendir(fn);
 		if (dir) {
@@ -850,6 +853,7 @@ file_signextend(struct magic_set *ms, struct magic *m, uint64_t v)
 		case FILE_REGEX:
 		case FILE_SEARCH:
 		case FILE_DEFAULT:
+		case FILE_INDIRECT:
 			break;
 		default:
 			if (ms->flags & MAGIC_CHECK) {
@@ -1152,6 +1156,12 @@ parse(struct magic_set *ms, struct magic_entry **mentryp, uint32_t *nmentryp,
 			case 'G':
 				m->in_type = FILE_BEDOUBLE;
 				break;
+			case 'i':
+				m->in_type = FILE_LEID3;
+				break;
+			case 'I':
+				m->in_type = FILE_BEID3;
+				break;
 			default:
 				if (ms->flags & MAGIC_CHECK)
 					file_magwarn(ms,
@@ -1435,6 +1445,38 @@ parse_strength(struct magic_set *ms, struct magic_entry *me, const char *line)
 	return -1;
 }
 
+/*
+ * Parse an Apple CREATOR/TYPE annotation from magic file and put it into magic[index - 1]
+ */
+private int
+parse_apple(struct magic_set *ms, struct magic_entry *me, const char *line)
+{
+	size_t i;
+	const char *l = line;
+	struct magic *m = &me->mp[me->cont_count == 0 ? 0 : me->cont_count - 1];
+
+	if (m->apple[0] != '\0') {
+		file_magwarn(ms, "Current entry already has a APPLE type `%.8s',"
+		    " new type `%s'", m->mimetype, l);
+		return -1;
+	}	
+
+	EATAB;
+	for (i = 0; *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l))
+	     || strchr("-+/.", *l)) && i < sizeof(m->apple); m->apple[i++] = *l++)
+		continue;
+	if (i == sizeof(m->apple) && *l) {
+		if (ms->flags & MAGIC_CHECK)
+			file_magwarn(ms, "APPLE type `%s' truncated %zu",
+			    line, i);
+	}
+
+	if (i > 0)
+		return 0;
+	else
+		return -1;
+}
+
 /*
  * parse a MIME annotation line from magic file, put into magic[index - 1]
  * if valid
@@ -1453,10 +1495,8 @@ parse_mime(struct magic_set *ms, struct magic_entry *me, const char *line)
 	}	
 
 	EATAB;
-	for (i = 0;
-	     *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l))
-	     || strchr("-+/.", *l)) && i < sizeof(m->mimetype);
-	     m->mimetype[i++] = *l++)
+	for (i = 0; *l && ((isascii((unsigned char)*l) && isalnum((unsigned char)*l))
+	     || strchr("-+/.", *l)) && i < sizeof(m->mimetype); m->mimetype[i++] = *l++)
 		continue;
 	if (i == sizeof(m->mimetype)) {
 		m->desc[sizeof(m->mimetype) - 1] = '\0';
@@ -1990,7 +2030,7 @@ apprentice_map(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp,
 		goto internal_loaded;
 	}
 
-	mkdbname(fn, &dbname, 0);
+	dbname = mkdbname(ms, fn, 0);
 	if (dbname == NULL)
 		goto error2;
 
@@ -2073,7 +2113,7 @@ apprentice_map(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp,
 	}
 
 	if (dbname) {
-		efree(dbname);
+		free(dbname);
 	}
 	return ret;
 
@@ -2090,7 +2130,7 @@ apprentice_map(struct magic_set *ms, struct magic **magicp, uint32_t *nmagicp,
 	}
 error2:
 	if (dbname) {
-		efree(dbname);
+		free(dbname);
 	}
 	return -1;
 }
@@ -2111,7 +2151,7 @@ apprentice_compile(struct magic_set *ms, struct magic **magicp,
 
 	TSRMLS_FETCH();
 
-	mkdbname(fn, &dbname, 1);
+	dbname = mkdbname(ms, fn, 0);
 
 	if (dbname == NULL) {
 		goto out;
@@ -2148,31 +2188,53 @@ apprentice_compile(struct magic_set *ms, struct magic **magicp,
 
 	rv = 0;
 out:
-	efree(dbname);
+	free(dbname);
 	return rv;
 }
 
 private const char ext[] = ".mgc";
 /*
  * make a dbname
  */
-private void
-mkdbname(const char *fn, char **buf, int strip)
+private char *
+mkdbname(struct magic_set *ms, const char *fn, int strip)
 {
-	const char *p;
+	const char *p, *q;
+	char *buf;
+
 	if (strip) {
 		if ((p = strrchr(fn, '/')) != NULL)
 			fn = ++p;
 	}
-	if ((p = strstr(fn, ext)) != NULL && p[sizeof(ext) - 1] == '\0')
-		*buf = strdup(fn);
-	else
-		(void)spprintf(buf, 0, "%s%s", fn, ext);
 
-	if (buf && *buf && strlen(*buf) > MAXPATHLEN) {
-		efree(*buf);
-		*buf = NULL;
+	for (q = fn; *q; q++)
+		continue;
+	/* Look for .mgc */
+	for (p = ext + sizeof(ext) - 1; p >= ext && q >= fn; p--, q--)
+		if (*p != *q)
+			break;
+
+	/* Did not find .mgc, restore q */
+	if (p >= ext)
+		while (*q)
+			q++;
+
+	q++;
+	/* Compatibility with old code that looked in .mime */
+	if (ms->flags & MAGIC_MIME) {
+		asprintf(&buf, "%.*s.mime%s", (int)(q - fn), fn, ext);
+		if (access(buf, R_OK) != -1) {
+			ms->flags &= MAGIC_MIME_TYPE;
+			return buf;
+		}
+		free(buf);
 	}
+	asprintf(&buf, "%.*s%s", (int)(q - fn), fn, ext);
+
+	/* Compatibility with old code that looked in .mime */
+	if (strstr(p, ".mime") != NULL)
+		ms->flags &= MAGIC_MIME_TYPE;
+	return buf;
 }
 
 /*

ext/fileinfo/libmagic/apptype.c

@@ -26,15 +26,14 @@
 
 #include "file.h"
 
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-
 #ifndef	lint
 FILE_RCSID("@(#)$File: apptype.c,v 1.7 2007/01/12 17:38:27 christos Exp $")
 #endif /* lint */
 
+
+#include <stdlib.h>
+#include <string.h>
+
 #ifdef __EMX__
 #include <io.h>
 #define INCL_DOSSESMGR