[gst-devel] debugging system rewrite

[Benjamin Otte]

On 10 Jun 2003, Ronald Bultje wrote:

> Hey Benjamin,
>
> Anyway, question being: do we really need a GST_ERROR/WARNING category?
> What's wrong with g_error()/g_warning() (programming errors) or
> gst_element_error() (user error)?
>
We have a lot of cases where we use g_warning when there is interesting
stuff happening, where the g_warning is wrong and we have to use something
else. Examples are the clock request diff, xruns on the soundcard,
unexpected timestamps in continuous streams and so on. Stuff that is not
supposed to happen but is no programmer's error if it happens. Stuff that
might be important if it happens.
GST_ERROR is about user errors mostly. But even errors that are presented
to the user should be logged, so that you know the user continued after an
error without having to guess when he doesn't tell you.

Benjamin

> No more comments, sounds good to me. ;).
>
> Ronald
>
> --
> Ronald Bultje <rbultje at ronald.bitfreak.net>
>
>