There are several reasons for doing that:
1. It documents the code for the reader and helps find
inconsistencies and bugs.
2. For rej_perms, it guarantees the change will be done
even if the output reference can be created by several
code paths.
3. It can be used to predict cases where a copy will,
or will not happen and optimize buffer allocation
(for example not request a rare direct-rendering buffer
from a device sink if it will be copied anyway).
Note that a filter is still allowed to manage the permissions
on its own without using these fields.
The code currently use cur_buf as the target of the copy,
but cur_buf can be cleared by the filter if it has given
the reference away or stored it elsewhere as soon as start_frame.
The code still relies on the fact that the reference is not
destroyed until end_frame. All filters currently follow that condition.
An av_assert1() is added to check it; it should at least cause
very visible errors in valgrind.
Semantic for the function ff_null_start_frame() was changed in
07bad27810, and it has now the same behavior of
ff_null_start_frame_keep_ref(), thus it makes no sense to keep both of
them.
Use the same values of the video output link.
Avoid the need to override the default_start_frame() with an ad-hoc
start_frame() callback.
In particular, fix the super2xsai filter which was setting the
input w/h values in the output.
Its only used through the start_frame pointer and thus cannot be inlined easily.
It also appears to break compilation with some unidentified compiler on darwin.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
git blame:
132 Stefano Sabatini
77 Vitor Sessak
49 Michael Niedermayer
24 Anton Khirnov
22 S.N. Hemanth Meenakshisundaram
13 Bobby Bingham
7 Luca Barbato
2 Nicolas George
2 Alex Converse
1 Diego Elio Pettenò
Initial commit not traced as this file was split out.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Nicolas George
Martin Storsjö
Stefano Sabatini
Loren Merritt
Anton Khirnov
Paul B Mahol
Michael Niedermayer
Diego Biurrun