[gst-devel] it seems seek event depends on element in the queue behind the seek?

[Ákos Maróy]

I'm looking more into the seeking issue I'm having.

I've made some more experiments, and I've come up with some results I
didn't expect myself.

One is, that it seems that the success of a seek event depends on the
element 'behind' the seek event - something I don't really understand.

consider the following queue:

filesrc ! mad ! alsasink


where you'd send a seek event to the src pad of mad. in that case, it
seems to work (though I have still questions, see my previous e-mail).
my imagination would be, that since this is an upstream event, only
elements 'before' mad are affected (namely filesrc and mad).

now, if I replace alsasink with a fakesink:

filesrc ! mad ! fakesink


the same seek fails.

this is very strange, as the direction of the seek event should be mad
-> filesrc, as it is an upstream event. how come an element that
supposedly doesn't even get this event affects the outcome?


Akos