From: Neil Van Dyke (neil at neilvandyke.org) Date: Tue Dec 30 15:09:54 EST 2014 |
|
Consider implementing it from scratch, in an application-specific way. Implementing the writing is pretty simple and straightforward (easier than the reading). Just be careful that your boundary string does not occur in any part (which is a weakness of the protocol, IMHO): you can either sanity-check all your encoded parts, or you can cross your fingers and make a big random string. Neil V.
Posted on the users mailing list. |
|