It's usually not a problem if a query fails if there's no peer,
especially as it will happen during pad linking (caps query)
quite often and spams the logs.
If we try to push sticky events but a probe dropped them, we don't mark
the event as received and mark the pad as PENDING_EVENTS. This ensures
that we resend the event the next time. For this we need to let the
custom flow return from the probe trickle up to
gst_pad_push_event_unchecked() so that we can differentiate between
OK and DROPPED probe returns.
First check that we can actually register the implementation before
making a GstMetaInfo. If we can't register we would otherwise end
up with an undefined type and an invalid GstMetaInfo.
It's possible that type registration fails because another metadata
with the same implementation name was already registered.
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
This happened when glib was not using system printf, and caused the
internal gstreamer printf extensions to be used for all %p printfs,
causing crashes.
https://bugzilla.gnome.org/show_bug.cgi?id=684970
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
Facilitate GstBuffer -> GstSample transition for some tags,
could be hard to catch otherwise when creating tags, since
it'll only be apparent later when someone tries to read the
tags.
We usually first create the stream_id for the stream_start event and then add
the pad to the element. This means that this functions should work when there
are no pads on the element yet.
Recheck for sticky events after doing a pad block because the pad block could
have caused a relink and then we need to resend the events to the newly linked
pad.
Fixes things like switching of visualisations.
The duration should be re-queried via a query using the
normal path, we don't want applications to use the value
from the message itself, since it might no match what a
duration query done from the sink upstream might yield.
Also disables duration caching in GstBin. It should be
added back again at some point.
When making a copy of the memory allocated from the default memory allocator,
make sure the new copy has the same alignment as the original memory.
See https://bugzilla.gnome.org/show_bug.cgi?id=680796
Elements such as the GstIirEqualizerNBands would so far not store the properties
of their children. Now we also grab the properties of child elements and try to
restore them.
Not so useful: just adds/reads stuff from an internal GList without
actually doing anything with those paths, so remove for now:
gst_registry_add_path
gst_registry_get_path_list
https://bugzilla.gnome.org/show_bug.cgi?id=608841
No longer accept any old GObjects. This makes things nicer for
bindings. If a utility function that handles both nicely
is deemed worthwhile, we can still add one to gstutils.
https://bugzilla.gnome.org/show_bug.cgi?id=681681
Add an alternative version of gst_pad_check_reconfigure that doesn't
clear the reconfigure flag.
Useful for increasing error resilience without duplicating the
reconfigure code in pad task functions.
API: gst_pad_needs_reconfigure
https://bugzilla.gnome.org/show_bug.cgi?id=681198
GObject Introspection does not support macros.
This is needed for bindings. We can still add back
macros or inline functions again later if we think
it's worth it.
https://bugzilla.gnome.org/show_bug.cgi?id=678301
When max_buffers > 0 and the pool is empty, actually try to allocate more
buffers up to the max_buffers limit.
We need to add a counter for this to count how many buffers we allocated and
check this against the max_buffers limit.
Reorganise and clean up some code.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=681153
The order of returned pads wasn't specified before, so let's specify
it and use an order which might prove the most useful : the order in
which pads were added to the element.
If someone changes the order, make sure users of those iterators from
now on don't rely on that order !
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.
So we can serialise/deserialise taglists inside structures,
which used to work automagically before because GstTagList
was just a typedef to GstStructure (same for the GType),
but now that it's a separate GType we need to register
explicit functions for this.
Helps with GDP stuff in pipelines/streamheader tests.
Remove the estimated-total field, this should not be part of the buffering
message.
Set the default value of buffering-left to 0 when the percent is 100.
In the proxy_query_caps function, also filter against the filter in the query.
We don't need to filter against the filter in the query anymore in the default
caps query function because we already did this in the proxy_query_caps.
Define a 0 and -1 step amount. They used to almost do the same thing but now, 0
cancels/stops the current step and -1 keeps on stepping until the end of the
segment.
See https://bugzilla.gnome.org/show_bug.cgi?id=679378
This reverts commit dd9fedb41f.
This is not the right place to escape the \, we should only escape the spaces to
keep the arguments together that were provided as one group (with quotes on the
shell).
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.
The ghostpad code directly activates/deactivates the child code by
calling gst_pad_activate_mode, rather than gst_pad_set_active, so
make sure to clear the flags in gst_pad_activate_mode(), which should
catch all cases.
Make gst_query_add_allocation_meta() take a copy of the passed caps instead of
taking ownership. This makes it easier for the caller in most cases because it
doesn't have to make a copy and deal with NULL values.
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).
We still don't do that in _to_iso8601_string() though, since
this will probably mostly be used in tags, where it doesn't
matter so much and the microsecond argument might not be
well-received by some tag readers.
When we fail to parse the number of seconds, reset the value to -1
instead of passing some error value as seconds. Also, we can still
try to parse timezone information.
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.
Make it possible to add API specific flags to the ALLOCATION query. This makes
it possible to also check what kinds of subfeatures of the metadata API are
supported.
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.
The NO_SHARE flag does not influence the exclusiveness of the buffer initially
but only if a _share operation can be done. Otherwise, we would not be able to
WRITE map a buffer memory because it would have a share count of at least 2.
We implement the locking in gst_memory_map with the lock flags, make matching
flags the same number so that we can use the map flags directly as lock flags.
Expose the internally used methods for locking and unlocking the object. Pass
the access mode to the unlock function for extra checks and because we need it
for the EXCLUSIVE locks.
Make some new defines to specify the desired locking.
Add a new EXCLUSIVE lock mode which will increment the shared counter. Objects
with a shared counter > 1 will not be lockable in WRITE mode.
Improve parallel installability in setups like jhbuild by
providing versioned variants of some environment variables:
GST_REGISTRY_1_0
GST_PLUGIN_PATH_1_0
GST_PLUGIN_SYSTEM_PATH_1_0
GST_PLUGIN_SCANNER_1_0
will now be checked before checking the unversioned ones.
https://bugzilla.gnome.org/show_bug.cgi?id=679407
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.
Make GstSeekFlag to GstSegmentFlag conversion explicit, and
set only those seek flags in the segment flags which are
mapped. This makes sure we don't have extraneous flags
littering our segment flag field, which also fixes the
debug printing/serialisation of segment events in the
debug log.
Visual C++ does not have isnan(), so add fallback to
math-compat.h (could use _isnan() in this case, but
this makes it work for all cases where isnan is missing).
https://bugzilla.gnome.org/show_bug.cgi?id=679112
Make a gst_buffer_append_region() function that allows you to append a memory
region from one buffer to another. This is a more general version of
gst_buffer_append().
If we have a file called Foo\Bar.ogg, there is no way to pass
that filename properly to filesrc in gst_parse_launch(), since
gst_parse_unescape() will just unescape \x to x.
Not cherry-picking this into 0.10 since there are apparently
apps that work around this problem and which would break if
we fixed it there too.
https://bugzilla.gnome.org/show_bug.cgi?id=673319
Take into account that not all fields might be valid (though they
are valid in the GDateTime structure). But we should just return
unordered if the set fields don't match. Also, don't check
microseconds when comparing datetimes, since we don't serialise
those by default if they're available. This ensures date times are
still regarded as equal after serialising+deserialising.
Some tag parsers and writers use same datetime format based on ISO 8601.
We can reduce some code by creating some general functions for it.
API: gst_date_time_to_iso8601_string()
API: gst_date_time_new_from_iso8601_string()
https://bugzilla.gnome.org/show_bug.cgi?id=678031
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
So mini objects don't have to poke into the GstMiniObject part
of the structure. Saves lines of code, and seems slightly cleaner.
We don't have proper OO hierarchies or methods here after all.
These changes are to clean up syntax issues such as missing colons,
missing spaces, etc., and minor issues such as argument names in
headers not matching the implementation and/or documentation.
When 2 weak refs are added, the array is not resized big enough.
Simplify the weak ref handling code.
Free memory when we remove all weak refs.
Allow installing the same weak ref multiple times, like in gobject.
The size field is used by subclasses to store the total allocated size of the
memory for this miniobject. Because miniobject doesn't really do anything with
this field we can move it to the subclasses.
Add the running-time of the buffer that caused the async operation to complete
to the async-done message.
Update bin to handle the new async-done message.
Use the new RESET_TIME message to reset the start-time of the pipeline to the
requested time.
Make basesink request a new running-time when the flush-stop message tells it to
insteasd of waiting for preroll.
Add a new message to reset the pipeline running_time. Currently reseting the
pipeline can only be requested in the async_done message which means that the
pipeline needs to be prerolled. It is better to move this to a separate message.
Remove constructors we don't want:
gst_date_time_new_ymd_h() because we don't want to
support hour-only for now;
gst_date_time_new_ymd_hm() because we don't want to
add constructors with time info where the caller doesn't
have to think about what timezone the time is in.
Lots of compulsive clean-up. Docs fixes. Replace
has_minute() and has_hour() with has_time().
In order to deserialise and re-serialise dates and date times
from tags properly, we need to be able to express partial
dates (e.g. YYYY or YYYY-MM) and date times.
We only support partial date times where all the more
significant fields above the first unset field are set
(e.g. YYYY-00-DD is not supported).
Calling _get_foo() when foo is not set is not allowed
any more, callers need to check which fields are set
first.
https://bugzilla.gnome.org/show_bug.cgi?id=677757
Make the gst_bin_remove_func more like the add_func. Check if the element we try
to remove from the bin has the bin as the parent and set the parent flag to NULL
immediately, this allows us to avoid concurrent remove operations without using
the UNPARENTING element flag. After we unparented the element from the bin, we
update the bin state and remove the element from the list. Finally we unlink
all the pads.
This avoids a race condition where the element could still claim to have the
bin as the parent while the bin didn't have a pointer to the element anymore.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=647759
In the dispose handler we first need to release all the request pads and then
remove the remaining pads. This is because it is possible that releasing the
request pad might also cleanly remove some of the other dynamic pads, like
what rtpsession does.
https://bugzilla.gnome.org/show_bug.cgi?id=677436
Context: Latency configuration should not be
messed up because of not-linked pads. In general,
one return FALSE on latency distribution causes
the "overall" pipeline latency configuration to
fail. This shows up as noise in logs (warning).
Conflicts:
gst/gstpad.c
Otherwise a pipeline where one sticky event fails to be sent will
never forward EOS events downstream. This can cause pipelines to
wait forever for EOS on errors.
The linking behaviour of external variables that are not initialized
in the compilation unit where they are defined is undefined. On OS X
this causes a linking failure when statically linking GStreamer.
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
