View Change

1 comment:

• Patchset:

  • Patch Set #3:

    @pespin: can you explain why calling `osmo_fd_unregister()` after `close()` is wrong? AFAIU, because […]

    The problem is not "closing" the socket per se, but modifying the "fd" field of an osmo_fd which is registered (here setting it to -1).
    That was already not expected by the implementation of the API before, but it was simply not checked and assumed to be done correctly by the user of te osmo_fd API.
    Since recently we dediced to enforce this behavior through an assert instead of having unexpected behaviors.

To view, visit change 31804. To unsubscribe, or for help writing mail filters, visit settings.