Repository navigation
pauses during playback should cause fail up-front without waiting for OF to be run #218
Description
We have a number of examples of pauses in playback causing failures of the duration observation that are only noticed when the OF is run - e.g. cta-wave/device-observation-framework#118 (comment).
In at least some of these, the player reported status 'waiting' - presumably some network issues. Finding this out only when the OF is run is very unfriendly. If the app knows data isn't arriving fast enough then it should report failure up-front so it's obvious to the person running the test during the test run.
Activity
jpiesing commentedon Dec 3, 2025
This was briefly discussed in the HbbTV Testing Group meeting today. There was a preference for underflow to be detected without having to download a recording and run the OF.
louaybassbouss commentedon Dec 4, 2025
A solution is to count the number of transitions the playback goes in
waitingstate and add new assertion which should fail if the number of unexpected transitions to waiting >0. We need to double check if there are expected waiting state and not count them (not sure which tests but I can imagine that playback over gaps or low latency playback may fall into this category).jpiesing commentedon Dec 4, 2025
Perhaps it could be a configuration option, either per-test or globally that can be over-ridden per test? I agree some tests may need to tolerate this.