Merge pull request #16488 from vpisarev:filestorage_longlines

trying to fix handling file storages with extremely long lines

* trying to fix handling of file storages with extremely long lines: https://github.com/opencv/opencv/issues/11061

* * fixed errorneous pointer access in JSON parser.
* it's now crash-test time! temporarily set the initial parser buffer size to just 40 bytes. let's run all the test and check if the buffer is always correctly resized and handled

* fixed pointer use in JSON parser; added the proper test to catch this case

* fixed the test to make it more challenging. generate test json with
*
**
***
etc. shape

modules/core/src/persistence.cpp

@@ -606,12 +606,11 @@ public:
                     int last_occurrence = -1;
                     xml_buf_size = MIN(xml_buf_size, int(file_size));
                     fseek( file, -xml_buf_size, SEEK_END );
-                    std::vector<char> xml_buf_(xml_buf_size+2);
                     // find the last occurrence of </opencv_storage>
                     for(;;)
                     {
                         int line_offset = (int)ftell( file );
-                        const char* ptr0 = this->gets(&xml_buf_[0], xml_buf_size );
+                        const char* ptr0 = this->gets(xml_buf_size);
                         const char* ptr = NULL;
                         if( !ptr0 )
                             break;
@@ -693,8 +692,8 @@ public:
         }
         else
         {
-            const size_t buf_size0 = 1 << 20;
-            size_t buf_size = buf_size0;
+            const size_t buf_size0 = 40;
+            buffer.resize(buf_size0);
             if( mem_mode )
             {
                 strbuf = (char*)filename_or_buf;
@@ -704,8 +703,7 @@ public:
             const char* yaml_signature = "%YAML";
             const char* json_signature = "{";
             const char* xml_signature  = "<?xml";
-            char buf[16];
-            this->gets( buf, sizeof(buf)-2 );
+            char* buf = this->gets(16);
             char* bufPtr = cv_skip_BOM(buf);
             size_t bufOffset = bufPtr - buf;
 
@@ -720,23 +718,8 @@ public:
             else
                 CV_Error(CV_BADARG_ERR, "Unsupported file storage format");
 
-            if( !isGZ )
-            {
-                if( !mem_mode )
-                {
-                    fseek( file, 0, SEEK_END );
-                    buf_size = ftell( file );
-                }
-                else
-                    buf_size = strbufsize;
-                buf_size = MIN( buf_size, buf_size0 );
-                buf_size = MAX( buf_size, (size_t)(CV_FS_MAX_LEN*6 + 1024) );
-            }
             rewind();
             strbufpos = bufOffset;
-
-            buffer.reserve(buf_size + 256);
-            buffer.resize(buf_size);
             bufofs = 0;
 
             try
@@ -809,56 +792,70 @@ public:
             CV_Error( CV_StsError, "The storage is not opened" );
     }
 
-    char* gets( char* str, int maxCount )
+    char* getsFromFile( char* buf, int count )
+    {
+        if( file )
+            return fgets( buf, count, file );
+    #if USE_ZLIB
+        if( gzfile )
+            return gzgets( gzfile, buf, count );
+    #endif
+        CV_Error(CV_StsError, "The storage is not opened");
+    }
+
+    char* gets( size_t maxCount )
     {
         if( strbuf )
         {
             size_t i = strbufpos, len = strbufsize;
-            int j = 0;
             const char* instr = strbuf;
-            while( i < len && j < maxCount-1 )
+            for( ; i < len; i++ )
+            {
+                char c = instr[i];
+                if( c == '\0' || c == '\n' )
                 {
-                char c = instr[i++];
-                if( c == '\0' )
-                    break;
-                str[j++] = c;
                     if( c == '\n' )
+                        i++;
                     break;
                 }
-            str[j++] = '\0';
-            strbufpos = i;
-            if (maxCount > 256 && !(flags & cv::FileStorage::BASE64))
-                CV_Assert(j < maxCount - 1 && "OpenCV persistence doesn't support very long lines");
-            return j > 1 ? str : 0;
-        }
-        if( file )
-        {
-            char* ptr = fgets( str, maxCount, file );
-            if (ptr && maxCount > 256 && !(flags & cv::FileStorage::BASE64))
-            {
-                size_t sz = strnlen(ptr, maxCount);
-                CV_Assert(sz < (size_t)(maxCount - 1) && "OpenCV persistence doesn't support very long lines");
             }
-            return ptr;
+            size_t count = i - strbufpos;
+            if( maxCount == 0 || maxCount > count )
+                maxCount = count;
+            buffer.resize(std::max(buffer.size(), maxCount + 8));
+            memcpy(&buffer[0], instr + strbufpos, maxCount);
+            buffer[maxCount] = '\0';
+            strbufpos = i;
+            return maxCount > 0 ? &buffer[0] : 0;
         }
-#if USE_ZLIB
-        if( gzfile )
-        {
-            char* ptr = gzgets( gzfile, str, maxCount );
-            if (ptr && maxCount > 256 && !(flags & cv::FileStorage::BASE64))
+
+        const size_t MAX_BLOCK_SIZE = INT_MAX/2; // hopefully, that will be enough
+        if( maxCount == 0 )
+            maxCount = MAX_BLOCK_SIZE;
+        else
+            CV_Assert(maxCount < MAX_BLOCK_SIZE);
+        size_t ofs = 0;
+
+        for(;;)
         {
-                size_t sz = strnlen(ptr, maxCount);
-                CV_Assert(sz < (size_t)(maxCount - 1) && "OpenCV persistence doesn't support very long lines");
-            }
-            return ptr;
+            int count = (int)std::min(buffer.size() - ofs - 16, maxCount);
+            char* ptr = getsFromFile( &buffer[ofs], count+1 );
+            if( !ptr )
+                break;
+            int delta = (int)strlen(ptr);
+            ofs += delta;
+            maxCount -= delta;
+            if( ptr[delta-1] == '\n' || maxCount == 0 )
+                break;
+            if( delta == count )
+                buffer.resize((size_t)(buffer.size()*1.5));
         }
-#endif
-        CV_Error(CV_StsError, "The storage is not opened");
+        return ofs > 0 ? &buffer[0] : 0;
     }
 
     char* gets()
     {
-        char* ptr = this->gets(bufferStart(), (int)(bufferEnd() - bufferStart()));
+        char* ptr = this->gets(0);
         if( !ptr )
         {
             ptr = bufferStart();  // FIXIT Why do we need this hack? What is about other parsers JSON/YAML?
@@ -868,10 +865,12 @@ public:
         }
         else
         {
-            FileStorage_API* fs = this;
-            int l = (int)strlen(ptr);
+            size_t l = strlen(ptr);
             if( l > 0 && ptr[l-1] != '\n' && ptr[l-1] != '\r' && !eof() )
-                CV_PARSE_ERROR_CPP("Too long string or a last string w/o newline");
+            {
+                ptr[l] = '\n';
+                ptr[l+1] = '\0';
+            }
         }
         lineno++;
         return ptr;

modules/core/src/persistence_json.cpp

@@ -411,7 +411,10 @@ public:
         if( *ptr != '"' )
             CV_PARSE_ERROR_CPP( "Key must end with \'\"\'" );
 
-        const char * end = ptr;
+        if( ptr == beg )
+            CV_PARSE_ERROR_CPP( "Key is empty" );
+        value_placeholder = fs->addNode(collection, std::string(beg, (size_t)(ptr - beg)), FileNode::NONE);
+
         ptr++;
         ptr = skipSpaces( ptr );
         if( !ptr || !*ptr )
@@ -420,11 +423,6 @@ public:
         if( *ptr != ':' )
             CV_PARSE_ERROR_CPP( "Missing \':\' between key and value" );
 
-        /* [beg, end) */
-        if( end <= beg )
-            CV_PARSE_ERROR_CPP( "Key is empty" );
-
-        value_placeholder = fs->addNode(collection, std::string(beg, (size_t)(end - beg)), FileNode::NONE);
         return ++ptr;
     }
 
@@ -768,8 +766,6 @@ public:
             CV_PARSE_ERROR_CPP( "left-brace of top level is missing" );
         }
 
-        if( !ptr || !*ptr )
-            CV_PARSE_ERROR_CPP( "Unexpected End-Of-File" );
         return true;
     }
 

modules/core/test/test_io.cpp

@@ -1669,4 +1669,46 @@ TEST(Core_InputOutput, FileStorage_YAML_parse_multiple_documents)
     ASSERT_EQ(0, std::remove(filename.c_str()));
 }
 
+TEST(Core_InputOutput, FileStorage_JSON_VeryLongLines)
+{
+    for( int iter = 0; iter < 2; iter++ )
+    {
+        std::string temp_path = cv::tempfile("temp.json");
+        {
+        std::ofstream ofs(temp_path);
+        ofs << "{     ";
+        int prev_len = 0, start = 0;
+        for (int i = 0; i < 52500; i++)
+        {
+            std::string str = cv::format("\"KEY%d\"", i);
+            ofs << str;
+            if(iter == 1 && i - start > prev_len)
+            {
+                // build a stairway with increasing text row width
+                ofs << "\n";
+                prev_len = i - start;
+                start = i;
+            }
+            str = cv::format(": \"VALUE%d\", ", i);
+            ofs << str;
+        }
+        ofs << "}";
+        }
+
+        {
+        cv::FileStorage fs(temp_path, cv::FileStorage::READ);
+        char key[16], val0[16];
+        std::string val;
+        for(int i = 0; i < 52500; i += 100)
+        {
+            sprintf(key, "KEY%d", i);
+            sprintf(val0, "VALUE%d", i);
+            fs[key] >> val;
+            ASSERT_EQ(val, val0);
+        }
+        }
+        remove(temp_path.c_str());
+    }
+}
+
 }} // namespace