console printing causing reboots

I've found that large XML files cancrash the Roku when you try to parse them. Are you sure it's the print, and not the parse that's causing the crash?

yeah, its the print. commented it out and everything's working. I first encountered this a week or so ago, when creating a url to do the api request, as a test I printed it to the console before doing the 'get' and boom, rebooted. took a while to figure out, once I did, I just left off the http:// when I printed to the console and no more reboots. Today, I encountered the same thing where I had a mid(stringval,7) one char too short leaving the / in the url. fixed that and then it rebooted when it printed the xml to the screen (also a test so I would know that it was working. commented that out and the app (such as it is so far) is working.

- Joel

I routinely print urls with 'http://' and I've never seen this problem.

-JT

Hesre are the offending lines of code that caused the original problem:

xfer.setURL(api+resource+user+uresource+ckey+limit+offset)
print api+resource+user+uresource+ckey+limit+offset
reboot happens after the print above

where api= "http://api.company.com"
resource= "/files"
user= "/username"
uresource= "/files"
ckey = "consumerkey=myconsumerkeystring"
limit= "&limit=10"
offset= "&offset=0"


replacing this code with this:

xfer.setURL(api+resource+user+uresource+ckey+limit+offset)
print resource+user+uresource+ckey+limit+offset

fixed the problem

I can't reproduce it. It prints fine for me.

-JT

good!

None the less, on all three occasions where I encountered this, commenting out the print statement or altering it in some way eliminated the reboot. in the above example, my consumerkeystring is a long string of numbers and letters (oauth stuff), which is appended to every api request I make, so it is possible it has something to do with that string in combo with the http://

- Joel

I can't recreate with this code snippet...

We believe our box should never crash and would try to fix any reproducible crash. So if you can narrow down a more reproducible test case, I'll take a look.

--Kevin