Fixes negotiation taking a ridiculous amount of
time (multiple 10s of seconds on a core2) when
there are duplicate entries in lists.
Could have a negative performance impact on other
scenarios because we now have to iterate the
dest list to avoid duplicates, but we don't
have a lot of lists any more these days, and
they tend to be small anyway. The negatives
are hopefully countered by the positive effects
of reducing the list length early on in the
process. And in any case, it's the right thing
to do.
Based on patch by Andre Moreira Magalhaes.
https://bugzilla.gnome.org/show_bug.cgi?id=684981
MIME-type -> media types
Fix up the manual in various places with the 1.0 way of doing things
such as probes, static elements, scheduling, ...
Add porting from 0.10 to 1.0 chapter.
Add probe example to build.
Remove some docs for remove components such as GstMixer and
GstPropertyProbe, XML...
Also add test to make sure that if a pad probe is removed while it's
callback is running, the cleanup_hook isn't called again if it
returns GST_PAD_PROBE_REMOVE
define GLIB_DISABLE_DEPRECATION_WARNINGS earlier so that it is defined before
the glib headers are loaded or else we trip over the GValueArray deprecations in
gst-inspect.c.
Our check would make sure that GLib segfaults when
someone tries to instantiate an abstract type, which
is an extremely useful thing to check for.
In newer GLibs this is fixed and we get an abort with
a g_error() now it seems, so let's just remove this
check entirely.
This is because we need to be able to signal different TOCs
to downstream elements such as muxers and the application,
and because we need to send both types as events (because
the sink should post the TOC messages for the app in the
end, just like tag messages are now posted by the sinks),
and hence need to make TOC events multi-sticky.
https://bugzilla.gnome.org/show_bug.cgi?id=678742
This specifies if a given taglist applies to the complete
medium or only this specific stream. By default a taglist
has a stream scope.
Fixes bug #677619.
Add an offset field that is used to track at what position the segment was
updated. This is used to set the running time to 0 when we do a flushing
seek that doesn't update the position.
See https://bugzilla.gnome.org/show_bug.cgi?id=680306
Remove the user_data from the alloc vmethod. Subclasses that implement a new
alloc function can also implement their own vmethod to pass extra arguments. We
can then also require that custom allocators implement an alloc function so that
gst_allocator_alloc() always works.
Stop querying the duration once an element return unknown and return unknown
as a final result. This avoid eventually cutting off a stream too early.
Add a tests to docuement the behavior.
Make GstAllocator a GstObject instead of a GstMiniObject, like bufferpool.
Make a new gstallocator.c file. Make a GstAllocator subclass for the default
allocator.
Clear the initial floating ref in the init function for
busses and clocks. These objects can be set on multiple
elements, so there's no clear parent-child relationship
here. Ideally we'd just not make them derive from
GInitiallyUnowned at all, but since we want to keep
using GstObject features for debugging, we'll just do
it like this.
This should also fix some problems with bindings, which
seem to get confused when they get floating refs from
non-constructor functions (or functions annotated to
have a 'transfer full' return type). This works now:
from gi.repository import GObject, Gst
GObject.threads_init()
Gst.init(None)
pipeline=Gst.Pipeline()
bus = pipeline.get_bus()
pipeline.set_state(Gst.State.NULL)
del pipeline;
https://bugzilla.gnome.org/show_bug.cgi?id=679286https://bugzilla.gnome.org/show_bug.cgi?id=657202
This re-uses existing code and makes sure we properly serialise
and deserialise datetimes where not all fields are set (thus
fixing some warnings when serialising such datetimes).
Expose the GstAllocation structure and provide an _init function. This makes it
easier to make 'subclasses' of the allocator that contain more info.
It also allows us to expose the flags on the allocator miniobject.
Make a flag to note that the allocator uses a custom alloc function.
Add a method that memory implementations can call to initialize the standard
GstMemory structure.
Move the parent handling in the _free handler.
Rearrange some internal function parameters so that the order is consistent.
Add more memory examples
Move the locking methods from GstMemory to GstMiniObject.
Add a miniobject flag to enable LOCKABLE objects. LOCKABLE objects can
use the lock/unlock API to control the access to the object.
Add a minobject flag that allows you to lock an object in readonly mode.
Modify the _is_writable() method to check the shared counter for LOCKABLE
objects. This allows us to control writability separately from the refcount for
LOCKABLE objects.
We added a minimum length of three letters originally so we would
fail to recognise DOS/Windows-style filenames as valid URIs (as we
should). Two should be just fine as well.
Add option to avoid build binaries. When building for platforms like
android, you might want to not link any "final" binary, mostly because
it requires special link flags or other parts of code that aren't
in the C library.
https://bugzilla.gnome.org/show_bug.cgi?id=677621
Now that TOCs are refcounted and have a GType, we can just
stuff a ref of the TOC directly into the various toc
event/message/query structures and get rid of lots of
cracktastic GstStructure <-> GstToc serialisation and
deserialisation code. We lose some TOC sanity checking
in the process, but that should really be done when
it's being created anyway.
Let's keep it simple for now:
gst_toc_setter_reset_toc() -> gst_toc_setter_reset()
gst_toc_setter_get_toc_copy() -> removed
gst_toc_setter_get_toc() -> returns a ref now
gst_toc_setter_get_toc_entry_copy() -> removed,
use TOC functions instead
gst_toc_setter_get_toc_entry() -> removed,
use TOC functions instead
gst_toc_setter_add_toc_entry() -> removed,
to avoid problems with (refcount-dependent)
writability of TOC; use TOC functions instead
When the bin does an upward state change, try to avoid doing a downward state
change on the child and vice versa.
Add some more unit tests for this fix.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=621833
It causes the timestamp to go wrong, should not cause much of a performance
increase and in the cases where it is faster, it is broken in 0.10 as well.
We should try to review this when rewriting the adapter for 0.11 memory
features.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=674791
Rename the _get_value_array() functions to _get_g_value_array() and reintroduce
the former to operate on plain unboxed c datatypes (like in 0.10). The _g_value
variants are for bindings while the _value ones are more suited to processing
in elements.
Add a new LOCKED flag to the metadata. Refuse removing LOCKED metadata from
a buffer.
Mark the metadata from the bufferpool LOCKED.
Add unit test for LOCKED flag
Remove GST_MAJORMINOR and replace it by GST_API_VERSION
Also set GST_VERSION_{MAJOR,MINOR,MICRO,NANO} explicitely
now.
All versions are at 1.0.0 now for the release soon but
API/ABI can still change until the 1.0.0 release.
Next release versions until 1.0.0 will be 0.10.9X and
these will be release candidates. GST_VERSION_* will
nonetheless stay at 1.0.0.0.
gst_buffer_take_memory -> gst_buffer_insert_memory because insert is what the
method does.
Make all methods deal with ranges so that we can replace, merge, remove and map
a certain subset of the memory in a buffer. With the new methods we can make
some code nicer and reuse more code. Being able to deal with a subset of the
buffer memory allows us to optimize more cases later (most notably RTP headers
and payload that could be in different memory objects).
Make some more convenient macros that call the more generic range methods.
Add gst_buffer_append() which appends the memory blocks from one buffer to
another. Remove the old inefficient _merge() and _join() methods which forced a
premature memcpy in most cases.
Remove the _is_span() and _span() methods they are not needed anymore now that
we can _append(). Merging and spanning will be delayed until mapping or maybe
not at all when the element can deal with the different memory blocks.
Make it possible to wrap all kinds of memory by exposing all properties to
gst_buffer_new_wrapped_full(). This makes it possible to also create writable
memory without a free function or memory with extra padding.
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
Group the extra allocation parameters in a GstAllocationParams structure to make
it easier to deal with them and so that we can extend them later if needed.
Make gst_buffer_new_allocate() take the GstAllocationParams for added
functionality.
Add boxed type for GstAllocationParams.
Change gst_allocator_alloc() so that we can also spicify flags and padding.
Add 2 new flags to mark the memory 0 prefixed/padded. This allows us to
remove some resizes in the base classes.
When allocating memory, memset prefix and padding with 0 when the flags tell
us to.
On resize, clear the zero padding flags if we can't guarantee the memory is
still 0 filled.
Update tests.
Rename _do_simplify() to _simplify(). The name was introduced as a replacement
method for a deprecated method but we can now rename it again.
Fix some docs.
Make gst_caps_do_simplify() take ownership of the input caps and produce a
simplified output caps. This removes the requirement of having writable input
caps and the method can make the caps writable only when needed.
Rework some caps operations so they don't rely on writable caps but instead take
ownership of the input caps and do _make_writable() only when needed.
Remove some const from caps functions, it does not make much sense for
refcounted objects and does not allow us to return a refcount to the const input
caps.
Rework the base classes fixate vmethods to not operate on the caps in-place.
All this saves us around 30% of caps and structure copy and new operations.
Sticky events pushed on an unlinked pad should be stored on the pad. When the
pad is then linked and an event is pushed, the event should be merged with the
already existing sticky events and then the sticky events should be pushed in
the order that they were originally pushed.
Should reproduce 'GStreamer-WARNING **: wrong STREAM_LOCK count 0'
warnings (with make pipelines/seek.torture or pipelines/seek.forever
anyway, since it appears to be racy).
https://bugzilla.gnome.org/show_bug.cgi?id=670846
Because gst_pad_get_pad_template_caps() returns ANY when there is no template,
the query caps function should also return ANY when there is no template (and no
pad current caps) instead of EMPTY.
Split out the registration of the metadata API and its implementation. Make a
GType for each metadata API. This allows us to store extra information with the
API type such as the tags.
Change the buffer API so that we can get the metadata using the API GType.
Change the query API so that we use the metadata API GType in the allocation
query instead of a string.
Update netaddress and unit tests
Account for rounding errors in some places, and that two nows are
not always entirely identical, so allow some leeway when comparing
microseconds and seconds. Ran into this too often, esp. when the
system is under load.
Flesh out the transform method. Add a type and extra info to the transform
function so that implementation can transform the metadata.
Remove the copy function and replace with the more generic transform.
Make it possible to configure a GDestroyNotify and user_data for
gst_memory_new_wrapped() this allows for more flexible wrapping of foreign
memory blocks.
Rename _is_writable() with _is_exclusive because the writability does not depend
on the amount of references to the memory object anymore.
Add accessor macros for the memory flags.
Rename the GstBuffer _peek_memory() method to _get_memory() and return a
reference to the memory now that we can do this without affecting writability
of the memory object. Make it possible to also make this function merge the
buffer memory.
Add methods to replace memory in a buffer. Make some convience macros for the
buffer memory functions.
Fix unit tests.
Fix annoying gst_type_find_register() function signature. A simple
string with comma-separated extensions works just as well and saves
lines of code, casts, relocations and ultimately kittens.
This is now bindings firendly as _new is just a classic c convenience and all
the work is done in a constructor. As a side effect _new never fails.
Fix the tests.
We introduced our own when GLib didn't want to add a GType
for GError. But now that there is one, we can use GLib's
unconditionally and remove our version.
Make the memory object simply manage the data pointer and the maxsize and move
the offset and size handling to common functionality.
Use the READONLY flag to set a readonly lock.
Remove the data and size fields from the unmap method. We need an explicit
resize operation instead of using the unmap function.
Make internal helper lock and unlock functions.
Update unit test and users of the old API.
Which we had to add because GLib didn't have it
back in the day. Port everything to plain old
G_TYPE_DATE, which is also a boxed type. Ideally
we'd just use GDateTime for everything, but it
doesn't support not setting some of the fields
unfortuntely (which would be very useful for
tag handling in general, if we could express
2012-01 for example).
https://bugzilla.gnome.org/show_bug.cgi?id=666351
Count how many mappings are currently active and also with what access pattern.
Update the design doc with restrictions on the access patterns for nested
mappings.
Check if nested mappings obey the access mode restrictions of the design doc.
Add various unit tests to check the desired behaviour.
Make an unmap call with a different data pointer than the map call update the
offset field. This allows for both offset and size adjustements in the unmap
call.
The controller object was once copied from buzztards unit tests. Change
TestMonoSource to TestObj as it is not a full fledged element. Split the tests
into a core and library test suite.
Add a GstControlBinding class. This is a preparation for making the
controlsources generate double valued control curves and do the gparamspec
mapping in the control binding. Now the API in GstObject is again mostly
for convenience.
Move most of the code to a GstTimedValueControlSource. Split out the trigger
'interpolation mode' to a new control source class. Move tests and examples to
new api. Update docs.
Fixes#610338
Conflicts:
gst/gstindexfactory.c
libs/gst/base/gstbasetransform.c
plugins/elements/gstfakesink.c
plugins/elements/gstfakesrc.c
plugins/elements/gstidentity.c
plugins/elements/gstinputselector.c
plugins/elements/gstoutputselector.c
Note: did not merge any of the basetransform changes from 0.10.
Rename the last-buffer property to last-sample and make it return the new
GstSample type so that we can include caps and timing info in one nice bundle.
If we are dealing with a blocking probe, only then check if one the
blocking flags of the hook matches.
Add some more debug.
Make the pad unit test less racy.
Calling set_caps at that point is not useful in 0.10 (FIXME comment!), and in
0.11 it is totally pointless: the caps event doesn't stick to a flushing pad.
This test currently fails, but is there to ensure we fix this issue
and keep it fixed, since it completely breaks delayed negotiation
use-cases.
This behaviour started breaking since
dd65aae9a1
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.
Remove the getcaps function on the pad and use the CAPS query for
the same effect.
Add PROXY_CAPS to the pad flags. This instructs the default caps event and query
handlers to pass on the CAPS related queries and events. This simplifies a lot
of elements that passtrough caps negotiation.
Make two utility functions to proxy caps queries and aggregate the result. Needs
to use the pad forward function instead later.
Make the _query_peer_ utility functions use the gst_pad_peer_query() function to
make sure the probes are emited properly.
Make that implizit with attaching/detaching controlsources. This is a lot easier
and has less invalid state (controlled property without control source).
This make the controller even more lightweight (no extra object, no extra lock,
less indirections). For object that don't use the controller the only 'overhead'
is a 3 unused fields in the gst_object structure.
The fixate caps function was not used externally and we have vmethods in the
base classes where it is needed.
Update some docs.
simplify some fixate functions in the base classes. Also pass the untruncated
caps to the vmethod.
Make a new GstPadProbeInfo structure and pass this in the probe callback. This
allows us to add more things later and also allow the callback to replace or
modify the passed object.
Make a separate cookie to detect chancges in the list of probes and keeping
track of what hooks have been invoked yet.
Remove the requirement to have probes on srcpads in push mode and sinkpads in
pull mode.
Add some more debug.
Keep track of what callbacks got executed. If no callback is called and we are a
blocking pad, let the item pass. This allows you to block pads on selected
items only.
Explicitly have an UPSTREAM and DOWNSTREAM PadProbeType. This allows you to only
block the pad on upstream or downstream items.
Add convenience macros to only block on downstream/upstream items.
The internal proxy pad target is simply a cache of the internal proxy pad
peer. This patch uses the well implement GstPad peer handling to obtain the
target. This fixes issues with target not being set in both direction when
two ghostpads are linked together (empty bin).
https://bugzilla.gnome.org/show_bug.cgi?id=658517
Better now than later in the cycle. These might come in handy:
sed -i -e 's/GstProbeReturn/GstPadProbeReturn/g' `git grep GstProbeReturn | sed -e 's/:.*//' | sort -u`
sed -i -e 's/GST_PROBE_/GST_PAD_PROBE_/g' `git grep GST_PROBE_ | sed -e 's/:.*//' | sort -u`
sed -i -e 's/GstProbeType/GstPadProbeType/g' `git grep GstProbeType | sed -e 's/:.*//' | sort -u`
It's only used internally anyway and the helper struct
has namespace issues.
API: deprecated gst_plugin_feature_type_name_filter()
API: deprecated GstTypeNameData
Returning a newly allocated string makes no sense. It's unexpected for a
getter, and also this behaves differently in 0.10, so it would make future
merges harder.
Except for these two places here in core which were updated for the new
semantic, the return value is getting leaked all over the place.
There's no code that uses it other than multiqueue, so make it private
to multiqueue for now. That way we can also do optimisations that
require API/ABI breaks. If anyone ever wants to use it, we can still
make it public again.
Add more debug.
Alow resize to 0 bytes.
Do clipping correctly.
Add more unit tests. Also add a failing test: when we resize to 0 and then
try to resize back to the original size it fails because the memory was
removed.
Allow for negative offsets when doing memory copy and share.
Add fast path in the _get_sizes() function.
Fix resize for negative offset and expanding the buffer.
Add some unit tests.
It was a bit too clever, and didn't really work as an API,
confusing people to no end. Better implement specific methods
whether an interface is usable/available/ready on the interface
itself, or even add GError arguments, rather than try to have
per-instance interfaces.
Add an index to gst_buffer_take_memory() so that we can also insert memory at a
certain offset. This is mostly interesting to prepend a header memory block to
the buffer.
Add a boolean to the flush_stop event to make it possible to implement flushes
that don't reset_time.
Make basesink post async_done with the reset_time property from the flush stop
event.
Fix some unit tests
Require the memory implementations to implement a share operation. This allows
us to remove the fallback share implementation which uses a different allocator
implementation and complicates things too much.
Update design doc a bit.
See #651514 for details. It's apparently impossible to write code
that avoids both type punning warnings with old g_atomic headers and
assertions in the new. Thus, macros and a version check.
Keep track of installed number of probes to shortcut emission.
Allow NULL callbacks, this is useful for blocking probes.
Improve probe selection based on the mask, an empty mask for the data or the
scheduling flags equals that all probes match.
Add some more debug info.
Don't check the flushing flag in the probe callback handler, this needs to be
done before calling the handler.
Fix blocking probes.
Fix unit tests
Make the PadBlock callback take a GstBlockType parameter to handle the different
kind of stages in the pad block. This provides for more backwards compatibility
in the pad block API.
Separate blocking and unblocking into different methods, only blocking can do a
callback, unblock is always immediately. Also removed synchronous blocking, it
can always be implemented with a callback.
Make pad block call the callback as soon as the pad is not in use. This makes it
possible to make sure that when the callback is called, no activity is happening
on the pad and that no activity will ever happen until the pad is unblocked
again. This makes pad blocking work when there is no dataflow or after EOS and
greatly helps dynamic pipelines.
Move the probe handling right where we wait on the pad block. The two are
related but not the same and the probe can eventually influence the pad
blocking as we'll se later.
Fix up some broken unit tests or tests that fail with the new behaviour.
..and as a result gst_caps_is_equal() and others.
This now only checks if for every subset structure there is
a superset structure in the superset caps. Previously we were
subtracting one from another, creating completely new caps
and then even simplified them.
The new implemention now is about 1.27 times faster and doesn't
break the -base unit tests are anything anymore.
Change the sticky event array so that it contains a pending and an active event.
Events on the sinkpad are copied to the pending array and after the eventfunc
returned TRUE, moved to the active event. This allows us to queue new events
like when we do per-pad offsets without removing the currently active event.
Remove the active argument from the gst_pad_get_sticky_event() method, the
pending events are not something we want to expose.
Update the design docs with some clear rules for how sticky events are
handled.
Reimplement the sticky tags, use a small structure to hold the event and its
current state (active or inactive).
Events on sinkpads only become active when the event function returned success
for the event.
When linking, only update events that are different.
Avoid making a copy of the event array, use the object lock to protect the event
array and release it only to call the event function. This will need to check
if something changed, later.
Disable a test in the unit test, it can't work yet.
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.
Improve GstSegment, rename some fields. The idea is to have the GstSegment
structure represent the timing structure of the buffers as they are generated by
the source or demuxer element.
gst_segment_set_seek() -> gst_segment_do_seek()
Rename the NEWSEGMENT event to SEGMENT.
Make parsing of the SEGMENT event into a GstSegment structure.
Pass a GstSegment structure when making a new SEGMENT event. This allows us to
pass the timing info directly to the next element. No accumulation is needed in
the receiving element, all the info is inside the element.
Remove gst_segment_set_newsegment(): This function as used to accumulate
segments received from upstream, which is now not needed anymore because the
segment event contains the complete timing information.
API: gst_mini_object_weak_ref()
API: gst_mini_object_weak_unref()
Add weak referencing functionality to GstMiniObject, which
allows to get notifications when an mini object is destroyed
but doesn't increase the real refcount. This is mostly
useful for bindings.
Fixes bug #609473.
This allows to get the internal pad of ghostpads and
proxypads without using gst_pad_iterate_internal_links()
and is much more convenient.
The internal pad of a ghostpad is the pad of the opposite direction
that is used to link to the ghostpad target.
This prevents adding duplicates over and over again to the resulting
caps if they already describe the new intersection result.
While this changes intersection from O(n*m) to O(n^2*m), it results in
smaller caps, which in the end will decrease further processing times.
For example in an audioconvert ! audioconvert ! audioconvert pipeline,
when forwarding the downstream caps preference in basetransform
(see e26da72de25a91c3eaad9f7c8b2f53ba888a0394) this results in
16 instead of 191 caps structures.
This prevents adding duplicates over and over again to the resulting
caps if they already describe the new intersection result.
While this changes intersection from O(n*m) to O(n^2*m), it results in
smaller caps, which in the end will decrease further processing times.
For example in an audioconvert ! audioconvert ! audioconvert pipeline,
when forwarding the downstream caps preference in basetransform
(see e26da72de25a91c3eaad9f7c8b2f53ba888a0394) this results in
16 instead of 191 caps structures.
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.
Remove pad_alloc and all references. This can now be done more efficiently and
more flexible with the ALLOCATION query and the bufferpool objects. There is no
reverse negotiation yet but that will be done with an event later.
Currently fails (in normal circumstances) because we create a
socket pair for each bin's bus and exhaust the number of available
file descriptors.
https://bugzilla.gnome.org/show_bug.cgi?id=646624
