add some class id matching (from Christoph Biedl)

diff --git a/src/cdf.c b/src/cdf.c
index fb603643e545c1546f9abcbfc13d6f1088a8abf3..a50e84232eb41227de59d8f44d14931e3803fdd2 100644 (file)
--- a/src/cdf.c
+++ b/src/cdf.c
@@ -35,7 +35,7 @@
 #include "file.h"
 
 #ifndef lint
-FILE_RCSID("@(#)$File: cdf.c,v 1.54 2014/02/25 20:52:02 christos Exp $")
+FILE_RCSID("@(#)$File: cdf.c,v 1.55 2014/02/27 23:26:17 christos Exp $")
 #endif
 
 #include <assert.h>
@@ -675,11 +675,13 @@ out:
 
 int
 cdf_read_short_stream(const cdf_info_t *info, const cdf_header_t *h,
-    const cdf_sat_t *sat, const cdf_dir_t *dir, cdf_stream_t *scn)
+    const cdf_sat_t *sat, const cdf_dir_t *dir, cdf_stream_t *scn,
+    const cdf_directory_t **root)
 {
        size_t i;
        const cdf_directory_t *d;
 
+       *root = NULL;
        for (i = 0; i < dir->dir_len; i++)
                if (dir->dir_tab[i].d_type == CDF_DIR_TYPE_ROOT_STORAGE)
                        break;
@@ -688,6 +690,7 @@ cdf_read_short_stream(const cdf_info_t *info, const cdf_header_t *h,
        if (i == dir->dir_len)
                goto out;
        d = &dir->dir_tab[i];
+       *root = d;
 
        /* If the it is not there, just fake it; some docs don't have it */
        if (d->d_stream_first_sector < 0)

diff --git a/src/cdf.h b/src/cdf.h
index 22e1577422726d2d245f547941cdf719115a5c20..177868eb55ee3dcd19e578d239df2850a647b4ab 100644 (file)
--- a/src/cdf.h
+++ b/src/cdf.h
@@ -294,7 +294,8 @@ int cdf_read_dir(const cdf_info_t *, const cdf_header_t *, const cdf_sat_t *,
 int cdf_read_ssat(const cdf_info_t *, const cdf_header_t *, const cdf_sat_t *,
     cdf_sat_t *);
 int cdf_read_short_stream(const cdf_info_t *, const cdf_header_t *,
-    const cdf_sat_t *, const cdf_dir_t *, cdf_stream_t *);
+    const cdf_sat_t *, const cdf_dir_t *, cdf_stream_t *,
+    const cdf_directory_t **);
 int cdf_read_property_info(const cdf_stream_t *, const cdf_header_t *, uint32_t,
     cdf_property_info_t **, size_t *, size_t *);
 int cdf_read_summary_info(const cdf_info_t *, const cdf_header_t *,

diff --git a/src/readcdf.c b/src/readcdf.c
index 4251c4ab0ccd8bf34bf4dea79c1037805651cb2d..b42434ce5aaa77a9df9bc221aa9f457938a1770d 100644 (file)
--- a/src/readcdf.c
+++ b/src/readcdf.c
@@ -26,7 +26,7 @@
 #include "file.h"
 
 #ifndef lint
-FILE_RCSID("@(#)$File: readcdf.c,v 1.38 2014/02/18 11:09:31 kim Exp $")
+FILE_RCSID("@(#)$File: readcdf.c,v 1.39 2014/02/27 23:26:18 christos Exp $")
 #endif
 
 #include <assert.h>
@@ -70,6 +70,34 @@ static const struct nv {
        { NULL,                         NULL,                   },
 };
 
+static const struct cv {
+       uint64_t clsid[2];
+       const char *mime;
+} clsid2mime[] = {
+       {
+               { 0x00000000000c1084LLU, 0x46000000000000c0LLU },
+               "x-msi",
+       }
+}, clsid2desc[] = {
+       {
+               { 0x00000000000c1084LLU, 0x46000000000000c0LLU },
+               "MSI Installer",
+       },
+};
+
+private const char *
+cdf_clsid_to_mime(const uint64_t clsid[2], const struct cv *cv)
+{
+       size_t i;
+       for (i = 0; cv[i].mime != NULL; i++) {
+       printf("%llx %llx %llx %llx\n",
+               clsid[0], cv[i].clsid[0], clsid[1], cv[i].clsid[1]);
+               if (clsid[0] == cv[i].clsid[0] && clsid[1] == cv[i].clsid[1])
+                       return cv[i].mime;
+       }
+       return NULL;
+}
+
 private const char *
 cdf_app_to_mime(const char *vbuf, const struct nv *nv)
 {
@@ -94,7 +122,7 @@ cdf_app_to_mime(const char *vbuf, const struct nv *nv)
 
 private int
 cdf_file_property_info(struct magic_set *ms, const cdf_property_info_t *info,
-    size_t count)
+    size_t count, const uint64_t clsid[2])
 {
         size_t i;
         cdf_timestamp_t tp;
@@ -104,6 +132,9 @@ cdf_file_property_info(struct magic_set *ms, const cdf_property_info_t *info,
         const char *s;
         int len;
 
+        if (!NOTMIME(ms))
+               str = cdf_clsid_to_mime(clsid, clsid2mime);
+
         for (i = 0; i < count; i++) {
                 cdf_print_property_name(buf, sizeof(buf), info[i].pi_id);
                 switch (info[i].pi_type) {
@@ -160,7 +191,7 @@ cdf_file_property_info(struct magic_set *ms, const cdf_property_info_t *info,
                                                     buf, vbuf) == -1)
                                                         return -1;
                                         }
-                                } else if (info[i].pi_id ==
+                                } else if (str == NULL && info[i].pi_id ==
                                    CDF_PROPERTY_NAME_OF_APPLICATION) {
                                        str = cdf_app_to_mime(vbuf, app2mime);
                                }
@@ -207,7 +238,7 @@ cdf_file_property_info(struct magic_set *ms, const cdf_property_info_t *info,
 
 private int
 cdf_file_summary_info(struct magic_set *ms, const cdf_header_t *h,
-    const cdf_stream_t *sst)
+    const cdf_stream_t *sst, const uint64_t clsid[2])
 {
         cdf_summary_info_header_t si;
         cdf_property_info_t *info;
@@ -218,6 +249,8 @@ cdf_file_summary_info(struct magic_set *ms, const cdf_header_t *h,
                 return -1;
 
         if (NOTMIME(ms)) {
+               const char *str;
+
                 if (file_printf(ms, "Composite Document File V2 Document")
                    == -1)
                         return -1;
@@ -245,14 +278,32 @@ cdf_file_summary_info(struct magic_set *ms, const cdf_header_t *h,
                                 return -2;
                         break;
                 }
+               str = cdf_clsid_to_mime(clsid, clsid2desc);
+               if (str)
+                        if (file_printf(ms, ", %s", str) == -1)
+                               return -2;
         }
 
-        m = cdf_file_property_info(ms, info, count);
+        m = cdf_file_property_info(ms, info, count, clsid);
         free(info);
 
         return m == -1 ? -2 : m;
 }
 
+#ifdef notdef
+private char *
+format_clsid(char *buf, size_t len, const uint64_t uuid[2]) {
+       snprintf(buf, len, "%.8" PRIx64 "-%.4" PRIx64 "-%.4" PRIx64 "-%.4" 
+           PRIx64 "-%.12" PRIx64,
+           (uuid[0] >> 32) & (uint64_t)0x000000000ffffffffLLU,
+           (uuid[0] >> 16) & (uint64_t)0x0000000000000ffffLLU,
+           (uuid[0] >>  0) & (uint64_t)0x0000000000000ffffLLU, 
+           (uuid[1] >> 48) & (uint64_t)0x0000000000000ffffLLU,
+           (uuid[1] >>  0) & (uint64_t)0x0000fffffffffffffLLU);
+       return buf;
+}
+#endif
+
 protected int
 file_trycdf(struct magic_set *ms, int fd, const unsigned char *buf,
     size_t nbytes)
@@ -298,13 +349,26 @@ file_trycdf(struct magic_set *ms, int fd, const unsigned char *buf,
                 goto out2;
         }
 
-        if ((i = cdf_read_short_stream(&info, &h, &sat, &dir, &sst)) == -1) {
+        const cdf_directory_t *root_storage;
+        if ((i = cdf_read_short_stream(&info, &h, &sat, &dir, &sst,
+           &root_storage)) == -1) {
                 expn = "Cannot read short stream";
                 goto out3;
         }
 #ifdef CDF_DEBUG
         cdf_dump_dir(&info, &h, &sat, &ssat, &sst, &dir);
 #endif
+#ifdef notdef
+       if (root_storage) {
+               if (NOTMIME(ms)) {
+                       char clsbuf[128];
+                       if (file_printf(ms, "CLSID %s, ",
+                           format_clsid(clsbuf, sizeof(clsbuf),
+                           root_storage->d_storage_uuid)) == -1)
+                               return -1;
+               }
+       }
+#endif
 
         if ((i = cdf_read_summary_info(&info, &h, &sat, &ssat, &sst, &dir,
             &scn)) == -1) {
@@ -319,23 +383,22 @@ file_trycdf(struct magic_set *ms, int fd, const unsigned char *buf,
 #ifdef CDF_DEBUG
         cdf_dump_summary_info(&h, &scn);
 #endif
-        if ((i = cdf_file_summary_info(ms, &h, &scn)) < 0)
+        if ((i = cdf_file_summary_info(ms, &h, &scn,
+           root_storage->d_storage_uuid)) < 0)
                 expn = "Can't expand summary_info";
+
        if (i == 0) {
                const char *str = NULL;
                cdf_directory_t *d;
                char name[__arraycount(d->d_name)];
                size_t j, k;
-               for (j = 0; j < dir.dir_len; j++) {
+
+               for (j = 0; str == NULL && j < dir.dir_len; j++) {
                        d = &dir.dir_tab[j];
                        for (k = 0; k < sizeof(name); k++)
                                name[k] = (char)cdf_tole2(d->d_name[k]);
-                       if (NOTMIME(ms))
-                               str = cdf_app_to_mime(name, name2desc);
-                       else
-                               str = cdf_app_to_mime(name, name2mime);
-                       if (str != NULL)
-                               break;
+                       str = cdf_app_to_mime(name,
+                           NOTMIME(ms) ? name2desc : name2mime);
                }
                if (NOTMIME(ms)) {
                        if (str != NULL) {