# Quick question on reading from TCPSockets

**URL:** <https://www.ruby-forum.com/t/quick-question-on-reading-from-tcpsockets/171899>\
**Category:** Ruby\
**Created:** [July 11, 2009, 12:14am UTC](https://www.ruby-forum.com/t/quick-question-on-reading-from-tcpsockets/171899 "2009-07-11T00:14:55Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![snow-templar](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/977dab/32.png) [@snow-templar](https://www.ruby-forum.com/u/snow-templar)\
**Post date:** [July 11, 2009, 12:14am UTC](https://www.ruby-forum.com/t/quick-question-on-reading-from-tcpsockets/171899/1 "2009-07-11T00:14:55Z")

</div>

So in my program, I am reading a series of ASCII strings from the  
TCPSocket. My code is letting everything through just fine. The  
problem is in the verification of the data. You see, the point of this  
protocol I am using is I am receiving a series of strings where each  
string has a 4 letter command and then parameters after it. The command  
and parameters are separated by commas. So my main worry here is most  
of the examples of TCPSocket networking I’ve seen use the #read method  
which requires a known amount of bytes to be read and I am worried that  
I will cut one of the commands part way through and cause problems.

I thought about using #readlines but I wasn’t sure how it would treat a  
split message. If I get a message like

> GHWD,12,67585,234,4767,234232,757  
> HGYT,0,0  
> PJIY,9876,234,56432,23

but the full last command was actually

> PJIY,9876,234,56432,234565,234,7774

then that messes everything up. Is there a way to read data delimited  
by a newline and if the last line read does not end with a newline to  
have it placed back in a buffer of some sort to get the next set of  
messages appended to it?

---

<div class="post-metadata">

**Author:** ![snow-templar](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/977dab/32.png) [@snow-templar](https://www.ruby-forum.com/u/snow-templar)\
**Post date:** [July 11, 2009, 1:15am UTC](https://www.ruby-forum.com/t/quick-question-on-reading-from-tcpsockets/171899/2 "2009-07-11T01:15:56Z")

</div>

> then that messes everything up. Is there a way to read data delimited  
> by a newline and if the last line read does not end with a newline to  
> have it placed back in a buffer of some sort to get the next set of  
> messages appended to it?

Yeah you’d have to buffer it and process it only when a newline comes  
in.  
=r

---

<div class="post-metadata">

**Author:** ![snow-templar](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/977dab/32.png) [@snow-templar](https://www.ruby-forum.com/u/snow-templar)\
**Post date:** [July 11, 2009, 9:55am UTC](https://www.ruby-forum.com/t/quick-question-on-reading-from-tcpsockets/171899/3 "2009-07-11T09:55:29Z")

</div>

On 11.07.2009 00:15, Greg C. wrote:

> I thought about using #readlines but I wasn’t sure how it would treat a  
> then that messes everything up. Is there a way to read data delimited  
> by a newline and if the last line read does not end with a newline to  
> have it placed back in a buffer of some sort to get the next set of  
> messages appended to it?

You can simply use #gets or #each (same as #each\_line). Note that #each  
and #each\_line accept separators so in case you have something different  
as separator character for records (commands) you can still use that  
approach. You need to be aware though that the separator character is  
preserved:

socket.each “\n” do |command|  
command.chomp!  
cmd, \*args = command.split ‘,’  
end

Kind regards

robert