• Connects to MySQL on localhost
• Auto-Discovers all tables from a db of your choosing
• Creates models in C for those tables
• Creates another C application which:
  • Connects to MySQL on localhost
  • loads all the content of your tables into ram (using the previously created C models)
  • provides a basic web interface to access that ram-cached content

Say you have a database, called ttc In that database you have 1 table called routes

MYSQL> describe routes;
+-------------+--------------+------+-----+---------+----------------+ 
| Field       | Type         | Null | Key | Default | Extra          | 
+-------------+--------------+------+-----+---------+----------------+ 
| id          | int(11)      | NO   | PRI | NULL    | auto_increment | 
| created_at  | datetime     | YES  |     | NULL    |                | 
| updated_at  | datetime     | YES  |     | NULL    |                | 
| name        | varchar(100) | YES  |     | NULL    |                | 
| uri         | varchar(100) | YES  |     | NULL    |                | 
+-------------+--------------+------+-----+---------+----------------+ 
6 rows in set (0.01 sec)

This application will produce the following C model:

struct Route {
  int id;
  char created_at[20];
  char updated_at[20];
  char name[100];
  char uri[100];
};

struct Route *routes = 0;
int num_routes = 0;
int load_routes();

When load_routes() is called, it will load all the MySQL content into those models. On which you could go routes[2].name to access the name of route 2.

Once ./ttc_web_test is created and executed you can do a web request to https://localhost:4444/routes?id=2 and it will respond with something like (assuming that was what was in your database):

{ 
  "id":2, 
  "created_at":"2007-09-25 19:45:37", 
  "updated_at":"2010-06-15 16:52:20", 
  "name":"Bloor-Danforth Subway", 
  "uri":"bloor-danforth_subway" 
}

This isn't meant for data that changes at any decent frequency - you have to restart the application to reload your data. The web interface itself isn't terribly useful, I hope to add more example usage. Also, another note: keep your IDs small and together - this app allocates space from 0 to the MAX(id)+1, which means if your table has only 1 row but it's ID is 900k, it will allocate space for 900k rows for that single row. This is working as intended