Forum Discussion

roquoonewbie's avatar
13 years ago

Found a pretty big bug in the new Firmware/UI

Very easy to reproduce.

1) Enter a channel (any channel...say Netflix)
2) Using the External Control Point API, issue a command to take the user to the Channel Store, to a specific channel

For instance...take the user to the Hulu Plus screen in the channel store by doing a POST to:

http://192.168.0.29:8060/launch/11?contentID=2285

About 70-80% of the time, this causes the box to lock up and (very slowly) reboot.

I have confirmed this on 2 boxes (an LT and an XD). Both were fine wrt this prior to today's firmware update.

Roku...can you reproduce and comment on an ETA for the fix? Having a total hard reboot from a simple command is really bad.

14 Replies

  • I can't reproduce it, although I'm not running that exact firmware. Can you PM your serial number to me? Also, do you have Hulu Plus (the target in the Channel Store) installed when you do this?

    --Mark
  • I tried and cannot reproduce the issue on Roku2 XS fw5.1.443, with or without Hulu installed.

    What you are sending does not seem valid HTTP/1.1 to me - have you tried "POST ... HTTP/1.0"? Not that it excuses box crashing.
  • So it turns out that the box only crashes if a dev app is installed while the external request is made.

    Deleting a dev app "fixed" it. Adding the dev app back caused it to crash again (even though the dev app was in no way involved with the external control point request).

    So this is not a "mass" issue, but definitely a bug Roku introduced with the new Channel Store UI in the latest firmware.

    And, yes, the POST info was a valid HTTP 1.1 POST. I even used IE to do it with the same result.

    And the box was in all cases replying back with a 200 OK response, and began to execute the launch of the Channel Store app.

    It just crashed while painting the Channel Store screen (thumbnails were missing but most data was visible at the time of crash).
  • "roquoonewbie" wrote:
    So it turns out that the box only crashes if a dev app is installed while the external request is made.

    Deleting a dev app "fixed" it. Adding the dev app back caused it to crash again (even though the dev app was in no way involved with the external control point request).

    I always have a dev app installed, including when I did the testing of the issue earlier in this thread. Are you seeing it with any dev app, or just a specific one?