support for generic text normalizer (CL_TYPE_SCRIPT)

git-svn: trunk@3584

ChangeLog

@@ -1,3 +1,8 @@
+Mon Feb  4 23:20:12 EET 2008 (edwin)
+------------------------------------
+  * libclamav/scanners, filetypes, dconf:
+	support for generic text normalizer (CL_TYPE_SCRIPT)
+
 Mon Feb  4 23:06:34 EET 2008 (edwin)
 ---------------------------------
   * libclamav/textnorm.[ch]: generic text normalizer (bb #241)

libclamav/dconf.c

@@ -86,6 +86,7 @@ static struct dconf_module modules[] = {
     { "DOCUMENT",   "HTML",	    DOC_CONF_HTML,	    1 },
     { "DOCUMENT",   "RTF",	    DOC_CONF_RTF,	    1 },
     { "DOCUMENT",   "PDF",	    DOC_CONF_PDF,	    1 },
+    { "DOCUMENT",   "SCRIPT",	    DOC_CONF_SCRIPT,	    1 },
 
     { "MAIL",	    "MBOX",	    MAIL_CONF_MBOX,	    1 },
     { "MAIL",	    "TNEF",	    MAIL_CONF_TNEF,	    1 },

libclamav/dconf.h

@@ -74,6 +74,7 @@ struct cli_dconf {
 #define DOC_CONF_HTML	    0x1
 #define DOC_CONF_RTF	    0x2
 #define DOC_CONF_PDF	    0x4
+#define DOC_CONF_SCRIPT	    0x8
 
 /* Mail flags */
 #define MAIL_CONF_MBOX	    0x1

libclamav/filetypes.c

@@ -76,6 +76,7 @@ static const struct ftmap_s {
     { "CL_TYPE_PDF",		CL_TYPE_PDF		},
     { "CL_TYPE_UUENCODED",	CL_TYPE_UUENCODED	},
     { "CL_TYPE_HTML_UTF16",	CL_TYPE_HTML_UTF16	},
+    { "CL_TYPE_SCRIPT",         CL_TYPE_SCRIPT          },
     { "CL_TYPE_RTF",		CL_TYPE_RTF		},
     { "CL_TYPE_HTML",		CL_TYPE_HTML		},
     { "CL_TYPE_MAIL",		CL_TYPE_MAIL		},

libclamav/filetypes.h

@@ -63,6 +63,7 @@ typedef enum {
     CL_TYPE_CRYPTFF,
     CL_TYPE_PDF,
     CL_TYPE_UUENCODED,
+    CL_TYPE_SCRIPT,
     CL_TYPE_HTML_UTF16,
     CL_TYPE_RTF,
 

libclamav/scanners.c

@@ -85,6 +85,7 @@
 #include "unarj.h"
 #include "nulsft.h"
 #include "autoit.h"
+#include "textnorm.h"
 #include <zlib.h>
 #include "unzip.h"
 
@@ -1064,6 +1065,73 @@ static int cli_scanhtml(int desc, cli_ctx *ctx)
     return ret;
 }
 
+static int cli_scanscript(int desc, cli_ctx *ctx)
+{
+	unsigned char buff[FILEBUFF];
+	unsigned char normalized[SCANBUFF];
+	struct text_norm_state state;
+	struct stat sb;
+	char *tmpname = NULL;
+	int ofd = -1, ret;
+	ssize_t nread;
+
+	cli_dbgmsg("in cli_scantext()\n");
+
+	if(fstat(desc, &sb) == -1) {
+		cli_errmsg("cli_scanscript: fstat() failed for descriptor %d\n", desc);
+		return CL_EIO;
+	}
+
+	/* don't normalize files that are too large */
+	if(sb.st_size > 10485760) {
+		cli_dbgmsg("cli_scanscript: exiting (file larger than 10 MB)\n");
+		return CL_CLEAN;
+	}
+
+	/* dump to disk only if explicitly asked to,
+	 * otherwise we can process just in-memory */
+	if(cli_leavetemps_flag) {
+		if((ret = cli_gentempfd(NULL, &tmpname, &ofd))) {
+			cli_dbgmsg("cli_scanscript: Can't generate temporary file/descriptor\n");
+			return ret;
+		}
+	}
+
+	text_normalize_init(&state, normalized, sizeof(normalized));
+	ret = CL_CLEAN;
+
+	do {
+		nread = cli_readn(desc, buff, sizeof(buff));
+		if(nread <= 0 || state.out_pos + nread > state.out_len) {
+			/* flush if error/EOF, or too little buffer space left */
+			if((ofd != -1) && (write(ofd, state.out, state.out_pos) == -1)) {
+				cli_errmsg("cli_scanscript: can't write to file %s\n",tmpname);
+				close(ofd);
+				ofd = -1;
+				/* we can continue to scan in memory */
+			}
+			/* when we flush the buffer also scan */
+			if(cli_scanbuff(state.out, state.out_pos, ctx->virname, ctx->engine, CL_TYPE_TEXT_ASCII) == CL_VIRUS) {
+				ret = CL_VIRUS;
+				break;
+			}
+			text_normalize_reset(&state);
+		}
+		if(nread > 0 && (text_normalize_buffer(&state, buff, nread)) != nread) {
+			cli_dbgmsg("cli_scanscript: short read during normalizing\n");
+		}
+		/* used a do {}while() here, since we need to flush our buffers at the end,
+		 * and using while(){} loop would mean code duplication */
+	} while (nread > 0);
+
+	if(cli_leavetemps_flag) {
+		free(tmpname);
+		close(ofd);
+	}
+
+	return ret;
+}
+
 static int cli_scanhtml_utf16(int desc, cli_ctx *ctx)
 {
 	char *tempname, buff[512], *decoded;
@@ -1838,6 +1906,11 @@ int cli_magic_scandesc(int desc, cli_ctx *ctx)
 		ret = cli_scanhtml_utf16(desc, ctx);
 	    break;
 
+	case CL_TYPE_SCRIPT:
+	    if(DCONF_DOC & DOC_CONF_SCRIPT)
+	        ret = cli_scanscript(desc, ctx);
+	    break;
+
 	case CL_TYPE_RTF:
 	    if(DCONF_DOC & DOC_CONF_RTF)
 		ret = cli_scanrtf(desc, ctx);