Skip to content

Conversation

@stronk-dev
Copy link
Contributor

@stronk-dev stronk-dev commented Aug 19, 2025

This PR:

  • Checks for huge or negative timestamp jumps in the filter and falls back to the default frame duration in such cases. This prevents the encoder from getting stuck in a loop processing specific inputs.
    Note: technically the check for negative timestamps should be enough, but included the other one as well. We could also omit it.
  • Test that confirms whether one of the problematic input segments no longer triggers infinite outputs.

@thomshutt thomshutt requested a review from j0sh September 3, 2025 13:04
@stronk-dev
Copy link
Contributor Author

Note: Navigare reports he has used this build and it resulted in memory leaks. I don't see an apparent cause for this, but a point to keep in mind for the review. Might be missing something (like maybe it is something stupid like how I probe the input size)

@stronk-dev stronk-dev force-pushed the md/bugfix/infinite_outputs branch from 3771207 to d17b16b Compare September 3, 2025 19:32
@stronk-dev
Copy link
Contributor Author

Closed, in favour of #438 which targets the root cause

@stronk-dev stronk-dev closed this Sep 4, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant