You asked variations of this questions multiple times. What is missing from the answers you got so far? Regarding this question I'll paste the help on TCP Read, which has a section on message protocols:
bytes to read is the number of bytes to read. You must wire a value greater than 0 in order to read data from a network connection. The default is 0. Use one of the following techniques to handle messages that might vary in size:
- Send messages that are preceded by a fixed size header that describes the message. For example, it might contain a command integer that identifies what kind of message follows and a length integer that identifies how much more data is in the message. Both the server and client receive messages by issuing a read function of eight bytes (assuming each is a four byte integer), converting them into the two integers, and using the length integer to determine the number of bytes to pass to a second read function for the remainder of the message. Once this second read is complete, each side loops back to the read function of the eight byte header. This technique is the most flexible, but it requires two reads to receive each message. In practice, the second read usually completes immediately if the message is written with a single write function.
- Make each message a fixed size. When the content of a message is smaller than the fixed size you specify, pad the message to the fixed size. This technique is marginally more efficient because only a single read is required to receive a message at the expense of sending unnecessary data sometimes.
- Send messages that are strictly ASCII in content, where each message is terminated by a carriage return and linefeed pair of characters. The read function has a mode input that, when passed CRLF, causes it to read until seeing a carriage return and linefeed sequence. This technique becomes more complicated when message data can possibly contain CRLF sequences, but it is quite common among many internet protocols, including POP3, FTP, and HTTP.
In one of your previous posts, you chose the first option: prepending your message with a fixed size header (consisting only of the number of bytes to follow). An easy way to do that is getting the length of your message (in string form) and typecasting that to a string. Send the length, send the message, done.
The output of "string length" is a signed 32 bit integer. 32 bit is 4 bytes. So on the receiving end, you read 4 bytes, convert (typecast) them to an i32, and read that many bytes to get your message. Done.
Wrap sending and receiving in Sub VIs. Copy them to your projects if necessary.
