I am rather new to handle C callbacks in C++. I made a sqlite wrapper c++ class, which just calls sqlite3_exec().
static int callback(void *NotUsed, int argc, char **argv, char **azColName){
SqliteAccessor* sqlite = static_cast<SqliteAccessor*> NotUsed;
if(argc > 0) {
m_table_exists = true;
}
return 0;
}
class SqliteAccessor{
public:
bool has_table(const string dbName, const string tblName)
{
bool hasTable = false;
string sql;
sql = "SELECT " + quote_string(tblName) + "FROM " + quote_string(dbName)
+ "WHERE type = 'table' AND name = " + quote_string(tblName) + ";";
char *zErrMsg = 0;
int rc = sqlite3_exec(m_db, sql.c_str(), callback, (void*) this, &zErrMsg);
if( rc != SQLITE_OK ){
printf("SQL error: %s", zErrMsg);
sqlite3_free(zErrMsg);
}
return hasTable;
}
};
int caller(){
SqliteAccessor sqlite;
// to check if table exist
if (sqlite->has_table()){
// will above work or
// I should do with an extra call to query the changed state?
}
}
Now, I am quite confused how the caller can get the result from sqlite wrapper. I think, the caller cannot have the result by simply calling has_table()
, because the result is returned from callback. So shall the caller get the result by making another call, e.g. call sqlite->get_result_has_table()
?
Then this implies for every callback, I need to make a state in class SqliteAccessor, which will be very cumbersome. How to design the class to make it nice to use by caller?
Aucun commentaire:
Enregistrer un commentaire