You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have observed JS Heap increasing at a rate of approximately 1MB every 3 minutes when a new manifest is loaded when player.attachSource(<new manifest URL>) is called every 15 seconds when subtitles are enabled. We observe this in dash.js 4.7.4 and are yet to try earlier versions of dash.js
When subtitles are disabled, JS heap increases only a couple of MB per hour when attachSource() is called every 15 seconds.
This leads us to hypothesise that cues are not being properly cleaned up when a new manifest is loaded by the player.
We can see that the number of cues continues to grow across calls to attachSource()
We have observed JS Heap increasing at a rate of approximately 1MB every 3 minutes when a new manifest is loaded when
player.attachSource(<new manifest URL>)
is called every 15 seconds when subtitles are enabled. We observe this in dash.js 4.7.4 and are yet to try earlier versions of dash.jsWhen subtitles are disabled, JS heap increases only a couple of MB per hour when
attachSource()
is called every 15 seconds.This leads us to hypothesise that cues are not being properly cleaned up when a new manifest is loaded by the player.
We can see that the number of cues continues to grow across calls to
attachSource()
The fix merged in #4305 resolved a memory leak when soak testing a stream for 24 hours without calling
attachSource()
Not sure if this is related, but we also wanted to check the logic that has been implemented for the previous PR to discard outdated cues.
https://github.com/Dash-Industry-Forum/dash.js/pull/4305/files#diff-033321ac6c729cfa349f6ca06067679dc6919fe8fe1023ec6fd990c83e98ac9aR829
Should the expression be
return cue.startTime <= currentTime && cue.endTime >= currentTime
to determine if cue is active?The text was updated successfully, but these errors were encountered: