[v1.x] seekRange isn't increasing while playing a live stream

[atanaschristov]

First of all we are using quite an old version of shaka player. Our last update was on Nov 4, 2015 commit: fc0f19c

Here is the following issue we are struggling with:

When a live stream is playing, although timeShiftBufferDepth increases every time the manifest gets reloaded, it seems that the seekRange isn't affected by it at all. It stays the same since the player was initialized.

So the question is:
Is it wrong to expect the seekRange to increase while the player is playing a live stream?
And if no, is it fixed in one of the following versions?

Thanks in advance

[joeyparrish]

In general, the seek range should end at the live edge and begin timeShiftBufferDepth before that.

It's been 9 months since that revision, so it's tough to say if your specific issue has been fixed. A lot has changed, including a complete rewrite for v2.

The best thing you can do is check to see if your issue can be reproduced on our latest release or the nightly build. You can use the "(custom asset)" selection to input your own manifest and license server URLs.

I notice that the commit you referenced is from the v1.5 branch. If you are not able to upgrade to v2, you can also try to reproduce in our most recent v1 release, v1.6.5. Please note that we are not actively maintaining v1 right now, so please consider v2 if you can.

[atanaschristov]

Thanks for your answer @joeyparrish

It seems that the issue exists in v1.6.5

I have deep dived in it and found out the following:

1. the manifest is loaded and parsed regularly, So I could trace that the shaka receives and is aware of the new timeShiftBufferDepth values
2. it seems that although the mpd in general seems to get updated, the mpd instance in LiveSegmentIndex where the seek range is calculated(getSeekRangeInternal) is not updated. It always contains the initial timeShiftBufferDepth

I suspect(not sure though) that in v2.0.0-beta issue #295 would fix the issue of LiveSegmentIndex to use the latest updated mpd object. Am I correct?

So is v2.0.0-beta3 considered stable and if not when is expected to have v2.0.0 stable version

Again thanks a lot for your help

Update:
It seems in v2.0.0-beta3 the seekRange changes according to timeShiftBufferDepth. So the issue seems to be fixed.

[joeyparrish]

If you want to submit a PR to v1.6.x to backport the fix to v1, I would be happy to review it. Otherwise, I suggest upgrading to v2.

v2.0.0-beta3 is stable, but not yet complete. We currently have 6 open enhancement issues in our v2.0.0 milestone.

[atanaschristov]

We've decided to go with shaka v2.0.0. If I have time I may try to create a PR for v1.6.x.

Thanks a lot and keep up the good work. Cheers.

[joeyparrish]

Closing due to inactivity.