search/index: fixed a number of bad conversions to properly deal with text documents bigger than 2GB

src/utils/md5.cpp

@@ -368,7 +368,7 @@ string& MD5HexScan(const string& xdigest, string& digest)
 class FileScanMd5 : public FileScanDo {
 public:
     FileScanMd5(string& d) : digest(d) {}
-    virtual bool init(unsigned int size, string *)
+    virtual bool init(size_t size, string *)
     {
 	MD5Init(&ctx);
 	return true;

src/utils/readfile.cpp

@@ -78,7 +78,7 @@ class FileToString : public FileScanDo {
 public:
     FileToString(string& data) : m_data(data) {}
     string& m_data;
-    bool init(unsigned int size, string *reason) {
+    bool init(size_t size, string *reason) {
 	if (size > 0)
 	    m_data.reserve(size); 
 	return true;
@@ -135,10 +135,10 @@ bool file_scan(const string &fn, FileScanDo* doer, off_t startoffs,
 	noclosing = false;
     }
 
-    if (st.st_size > 0) {
-	doer->init(st.st_size+1, reason);
-    } else if (cnttoread) {
+    if (cnttoread != (size_t)-1 && cnttoread) {
 	doer->init(cnttoread+1, reason);
+    } else if (st.st_size > 0) {
+	doer->init(st.st_size+1, reason);
     } else {
 	doer->init(0, reason);
     }

src/utils/readfile.h

@@ -29,7 +29,7 @@ using std::string;
 class FileScanDo {
 public:
     virtual ~FileScanDo() {}
-    virtual bool init(unsigned int size, string *reason) = 0;
+    virtual bool init(size_t size, string *reason) = 0;
     virtual bool data(const char *buf, int cnt, string* reason) = 0;
 };
 bool file_scan(const string &filename, FileScanDo* doer, string *reason = 0);

src/utils/smallut.cpp

@@ -14,7 +14,11 @@
  *   Free Software Foundation, Inc.,
  *   59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
  */
+
 #ifndef TEST_SMALLUT
+#ifdef HAVE_CONFIG_H
+#include "autoconfig.h"
+#endif
 #include <stdio.h>
 #include <stdlib.h>
 #include <time.h>
@@ -556,7 +560,7 @@ bool pcSubst(const string& in, string& out, map<string, string>& subs)
 }
 
 // Convert byte count into unit (KB/MB...) appropriate for display
-string displayableBytes(long size)
+string displayableBytes(off_t size)
 {
     char sizebuf[30];
     const char * unit = " B ";
@@ -568,7 +572,7 @@ string displayableBytes(long size)
 	unit = " MB ";
 	size /= (1024*1024);
     }
-    sprintf(sizebuf, "%ld%s", size, unit);
+    sprintf(sizebuf, OFFTPC "%s", size, unit);
     return string(sizebuf);
 }
 

src/utils/smallut.h

@@ -16,6 +16,7 @@
  */
 #ifndef _SMALLUT_H_INCLUDED_
 #define _SMALLUT_H_INCLUDED_
+
 #include <stdlib.h>
 
 #include <string>
@@ -125,7 +126,7 @@ extern string truncate_to_word(const string &input, string::size_type maxlen);
 extern void utf8truncate(string &s, int maxlen);
 
 /** Convert byte count into unit (KB/MB...) appropriate for display */
-string displayableBytes(long size);
+string displayableBytes(off_t size);
 
 /** Break big string into lines */
 string breakIntoLines(const string& in, unsigned int ll = 100,