In commit bf0964b6 a check for pad is activated was not carried.
This leads to attempt to pull while in push mode when force_caps
is set. In this case without the attached check even when activated
in pull mode we activate back to push mode.
This is from comment in previous code , case number eight:
8. if the sink pad is activated, we are in pull mode. succeed.
- otherwise activate both pads in push mode and succeed.
Putting it back fixes playback of webm in webkit+gstreamer 1.0 .
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=676003
Make it so that one can specify a buffer for get/pull_range where the downstream
element should write into. When passing NULL, upstream should allocate a buffer,
like in 0.10.
We also need to change the probes a little because before the pull probe, there
could already be a buffer passed. This then allows us to use the same PROBE
macro for before and after pulling.
While we're at the probes, make the query probe more powerful by handling the
GST_PAD_PROBE_DROP return value. Returning _DROP from a query probe will now
return TRUE upstream and will not forward the probe to the peer or handler.
Also handle _DROP for get/pull_range properly by not dispatching to the
peer/handler or by generating EOS when the probe returns DROP and no buffer.
Make filesrc handle the non-NULL buffer passed in the get_range function and
skip the allocation in that case, writing directly into the downstream provided
buffer.
Update tests because now we need to make sure to not pass a random value in the
buffer pointer to get/pull_range
Add the pad mode to the activate function so that we can reuse the same function
for all activation modes. This makes the core logic smaller and allows for some
elements to make their activation code easier. It would allow us to add more
scheduling modes later without having to add more activate functions.
Turns some boolean arguments in the scheduling query to flags, which are easier
to extend and makes the code easier to read.
Make extra methods for configuring and querying the supported scheduling modes.
This should make it easier to add new modes later.
when we are flushing, don't store the event on the pad but simply return FALSE.
Don't deactivate the srcpad, we need it to be active in order to push the
caps. Downstream can change the scheduling mode of an active pad.
This reverts commit cf4fbc005c.
This change did not improve the situation for bindings because
queries are usually created, then directly passed to a function
and not stored elsewhere, and the writability problem with
miniobjects usually happens with buffers or caps instead.
This reverts commit 9ef1346b1f.
Way to much for one commit and I'm not sure we want to get rid of the pad caps
just like that. It's nice to have the buffer and its type in onw nice bundle
without having to drag the complete context with it.
Once we switch to normal mode, we're not typefinding anymore and thus the caps
will not change. Therefore can avoid the object lock in the data-flow path.
The locking was added in order to fix bug #608877.
Typefind might mess up pads modes (pull/push) if a
downstream element is plugged and its pads activated
in 'step 2' of typefind pads activation.
This happens because the following steps don't check
if we already emitted typefound due to upstream setting
caps on buffers being pulled in the typefind helpers.
Avoid that by checking if typefound is already emmited.
Fixes#608036
For the reason outlined at the beginning of gst_private.h (inline
functions in glib may need the g_log_domain variable). Also include
gst_private.h before using any G_OS_* defines, esp. in plugin loader.
There's not much point in using GST_DEBUG_FUNCPTR with GObject
virtual functions such as get_property, set_propery, finalize and
dispose, since they'll never be used by anyone anyway. Saves a
few bytes and possibly a tenth of a polar bear.
Original commit message from CVS:
* plugins/elements/gstidentity.c:
Doc typo. Use return value of parent_class->event.
* plugins/elements/gsttypefindelement.c:
Chain up at the end for consistency.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_src_event),
(gst_type_find_element_handle_event):
Use faster and safer _pad_push_event().
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_handle_event),
(gst_type_find_element_send_cached_events),
(gst_type_find_element_change_state):
Clean up on FLUSH_STOP and not FLUSH_START. Forward both events
immediately.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_handle_src_query), (stop_typefinding),
(gst_type_find_element_handle_event),
(gst_type_find_element_send_cached_events),
(gst_type_find_element_change_state):
Forward FLUSH_START events immediately and clean up instead of
caching them.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_class_init),
(gst_type_find_element_set_property),
(gst_type_find_element_get_property),
(gst_type_find_element_activate):
* plugins/elements/gsttypefindelement.h:
Cleanup properties.
Fix pad leak when peer query fails.
We can still typefind when the peer returns -1.
Add property to force caps and bypass typefinding. This will be used in
uridecodebin.
API::force-caps
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_handle_event),
(gst_type_find_element_activate):
Post special error message if we can't determine the type of a stream
because it's empty.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_have_type), (gst_type_find_element_init),
(gst_type_find_element_setcaps), (gst_type_find_element_chain):
Move detecting the input caps of the sinkpad to the setcaps function.
This allows us to update the output caps when we receive new input caps
instead of always using the first detected caps.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_chain_do_typefinding),
(gst_type_find_element_change_state):
Don't leak found caps in chain function (no idea why that never
showed up as a leak anywhere).
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_handle_event):
reverted wrong change and reflowed code to avoid others falling into
this trap
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_handle_event):
If we get EOS before any data is accumulated, don't use
uninitialised local variables.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_chain):
More debug info; when skipping typefinding, send cached
events in all cases.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_chain):
Do not typefind content if the buffers already have caps.
Neccesary for icydemux (#333657), and the right thing to do anyway.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c: (stop_typefinding),
(gst_type_find_element_handle_event),
(gst_type_find_element_send_cached_events),
(gst_type_find_element_change_state):
* plugins/elements/gsttypefindelement.h:
When typefinding is done in push mode, we should cache
events we receive during typefinding instead of just
dropping them (e.g. newsegment, custom events from
dvdreadsrc etc.) and then send them out once we've
determined the type of the stream (and decodebin
has had a chance to plug in a decoder/demuxer).
Original commit message from CVS:
* plugins/elements/gsttypefindelement.h:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_src_event), (start_typefinding),
(stop_typefinding), (gst_type_find_element_handle_event),
(gst_type_find_element_chain),
(gst_type_find_element_chain_do_typefinding):
Use gst_type_find_helper_for_buffer() for chain-based
typefinding.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_class_init),
(gst_type_find_element_set_property),
(gst_type_find_element_get_property):
Deprecate "maximum" property (not only was it only taken into
account for typefinding in push-mode anyway, it also was never
actually possible to set it in the first place because the
property was registered with the numeric property ID for the
"minimum" property). Register "maximum" property correctly,
for the sake of future copy'n'pasters. Remove some cruft
from property get/set functions.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_activate):
Use gst_type_find_helper_get_range() here, so we
can honour the min-probability property and also emit
the signal with the correct probability of the found caps.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_handle_event):
When we get an EOS event and have not found a type yet
(most likely because we had not yet accumulated
TYPE_FIND_MIN_SIZE of data yet), try to determine the
type given the data we have so far. Fixes typefinding
for very short streams again, most notably quicktime
redirections as used on Apple's trailer site (#331701).
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_activate):
Use gst_pad_check_pull_range() before _activate_pull()
to avoid unnecessary open/close (see #331690).
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c: (find_peek):
We can do peeks at non-zero offsets, as long as they
fall within the buffer we have.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_chain):
When typefinding is unsuccessful in the chain function, don't
error out immediately. Only error out with NO_CAPS_FOUND if
the amount of data is at least MAX_TYPEFIND_SIZE bytes,
otherwise simply wait for more data so we can try typefinding
again with more data later. Also, don't attempt to typefind
if we have less than MIN_TYPEFIND_SIZE data available. Overall,
this should improve typefinding from network sources where the
size of the first buffer can be somewhat random.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_activate):
Post TYPE_NOT_FOUND error message when typefinding
is unsuccessful in the activate function as well.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c:
(gst_type_find_element_class_init):
'have-type' signal needs to be G_SIGNAL_RUN_FIRST, as it is the
object method handler that sets the caps on the pad and we want
that to happen before we emit the signal (fixes e.g. feeding a
plain text file to decodebin).
Original commit message from CVS:
* gst/gstpad.c: (gst_pad_activate_pull), (gst_pad_link_prepare):
Don't allow activation of a srcpad in pull_range if it has no
getrange function.
Change some debug statements to be a little clearer
* plugins/elements/gsttypefindelement.c:
(gst_type_find_handle_src_query):
Check that we have a peer before executing queries thereupon.
* tests/examples/metadata/read-metadata.c: (message_loop):
Use gst_bus_pop instead of gst_bus_poll when we just want it to
immediately return us any available message with 0 timeout.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c: (stop_typefinding):
Catch the special case where we are operating chain-based,
but the downstream peer pad has no chain function. Emit a
custom error message in this case instead of letting the
core generate one implying that this is some sort of core
bug. It's not, it just means that whatever got plugged
into the pipeline downstream when we announced the type
can only operate pull-based, while our source can only
operate push-based (e.g. http://foo/bar.mov ! qtdemux ! ...)
Error string has not been marked for translation yet, as
it probably needs some more work first.
(gst_type_find_element_get_best_possibility):
Add helper function to find the best of all available
found possibilities that qualify given the min. threshold.
(gst_type_find_element_handle_event):
Fix the case where we get an EOS while still in TYPEFIND
mode (we want to chose the best of all possible types,
not just the first type that happens to be in our unsorted
list of possible types).
(gst_type_find_element_chain):
Make sure we return GST_FLOW_ERROR when we errored out
in stop_typefinding(); also, don't just find the best of
all found type entries and then use the last examined
type entry, but actually use the best entry.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c: (free_entry),
(gst_type_find_element_chain):
Now that we're not leaking factories, make sure we keep references
to them while we need them.
Original commit message from CVS:
* plugins/elements/gsttypefindelement.c: (find_peek),
(gst_type_find_element_chain):
Fix leaking element factories in typefinding.
Fix problem where we forgot about a probable type on non-seekable
files, and thus later mis-typefound it.
Original commit message from CVS:
* check/Makefile.am:
Enable valgrind check that should work fine now.
* gst/elements/gsttypefindelement.c: (gst_type_find_element_init):
* gst/gstqueue.c: (gst_queue_init):
Fix memleaks in pad allocation.