script: Update media to include fix for AudioBufferSourceNode.loop#41075
script: Update media to include fix for AudioBufferSourceNode.loop#41075jdm merged 2 commits intoservo:mainfrom
media to include fix for AudioBufferSourceNode.loop#41075Conversation
|
🔨 Triggering try run (#19963896814) for Linux (WPT) |
|
Test results for linux-wpt from try job (#19963896814): Flaky unexpected result (30)
Stable unexpected results that are known to be intermittent (22)
Stable unexpected results (8)
|
|
|
|
Looks like this is still a draft and doesn't include the changes, but please include a description of the fix made in in the commit description and title. The motivation is this: if we end up having the fix
Thank you! |
This is still a draft, and I am still figuring things out. Do you want me to close the draft PR for now? |
Sorry, you don't have to close the draft. It was just a reminder. |
c17193a to
658c7ae
Compare
|
🔨 Triggering try run (#20048459851) for Linux (WPT) |
|
Test results for linux-wpt from try job (#20048459851): Flaky unexpected result (40)
Stable unexpected results that are known to be intermittent (19)
|
|
✨ Try run (#20048459851) succeeded. |
658c7ae to
8205aba
Compare
8205aba to
224ca77
Compare
|
why I can not reopen this PR? |
Previously: the moment we set bufferSource.loop = true, the node hit the early “extreme edge case” check, decided “I refuse to output”, triggered onended, and returned silence. So the loop never got a chance to run. Now: Early bailout is removed(no idea why was there in first place). The per tick offset is guaranteed to be a sensible value with respect to the loop range (wrapped if extreme). The normal mixing code runs, updating pos each tick and respecting the loop region Signed-off-by: Taym Haddadi <[email protected]>
never mind, I removed my commits |
|
Why are there no changes to test results? |
I am also confused, I think there is not test for this, I will have a look |
|
🔨 Triggering try run (#20065100291) for Linux (WPT) |
media to include fix for WebAudio bufferSource.loop
|
Test results for linux-wpt from try job (#20065100291): Flaky unexpected result (36)
Stable unexpected results that are known to be intermittent (20)
|
|
✨ Try run (#20065100291) succeeded. |
media to include fix for WebAudio bufferSource.loopmedia to include fix for AudioBufferSourceNode.loop
|
Updated the title: |
Signed-off-by: Taym Haddadi <[email protected]>
@jdm this case was not covered by test, I added new WPT test, can you have a quick look? test is passing now and was failing for servo before this changes. |
|
🤖 Opened new upstream WPT pull request (web-platform-tests/wpt#56621) with upstreamable changes. |
Fix in servo/media#476:
Previously:
the moment we set bufferSource.loop = true, the node hit the early “extreme edge case” check, decided “I refuse to output”, triggered onended, and returned silence. So the loop never got a chance to run.
Now:
Early bailout is removed(no idea why was there in first place).
The per tick offset is guaranteed to be a sensible value with respect to the loop range (wrapped if extreme).
The normal mixing code runs, updating pos each tick and respecting the loop region
Testing: I tested this manually, it would be nice to add a test for it, and I’m surprised there isn’t one yet.
Fixes: #41073