GST_VALUE_HOLDS_... macros may cause -Waddress warning
on gcc if GValue is allocated on stack:
gstvalue.h:145:46: warning: the comparison will always
evaluate as ‘true’ for the address of ‘v’ will never
be NULL [-Waddress]
#define GST_VALUE_HOLDS_CAPS(x) ((x) != NULL &&
G_VALUE_TYPE(x) == _gst_caps_type)
Fixes#653
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/738>
Preserve the previous behaviour where:
name, val="5";
passed to gst_structure_from_string would have resulted in an int value,
rather than a string, despite the quote marks.
This will be changed to being interpreted as a string in the future, but
for the time being we will issue a warning about this to give users time
to fix their code to no longer rely on this bug.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/303>
Allow a string in gst_string_unwrap to include unescaped characters that
are not in GST_STRING_IS_ASCII. This extra leniency allows
gst_structure_from_string to, e.g., receive
name, val=(string)"string with space";
Note that many gst tests, and potentially users, exploited this behaviour
by giving
name, val="string with space";
i.e. without the (string) type specifier. This was allowed before
because, without a type specifier, the string was passed to
_priv_gst_value_parse_string with unescape set to TRUE, *rather* than
being sent to gst_string_unwrap. This caused a difference in behaviour
between strings that are or are not preceded by (string). E.g.
name, val=(string)"string with space";
would fail, whilst
name, val="string with space";
would not. And
name, val=(string)"\316\261";
would produce a val="α", whereas
name, val=(string)"\316\261";
would produce a val="316261" (a bug).
The current behaviour is to treat both of these cases the same, which is
desirable. But in order to not break potentially common usage of this
discrepancy (it was in our own tests), the best option is to make string
parsing less strict in general.
New behaviour would be for
name, val=(string)"string with space";
to pass and give val="string with space", and
name, val="\316\261";
would produce a val="α".
Also changed deserializing string test to expect successes where
previously a failure was expected.
In a similar way, this also effected the deserializing of GstStructure,
GstCaps, GstTagList and GstCapsFeatures. So, now
name, val=(structure)"sub-name, sub-val=(string)\"a: \\316\\261\";";
will also pass and give sub-val="a: α". Note that the quote marks
and backslash still need to be escaped for the sub-structure, but other
characters need not be.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/303>
No longer call _priv_gst_value_parse_string with unescape set to TRUE
before passing a value to gst_value_deserialize in
_priv_gst_value_parse_value. This latter function is called by
gst_structure_from_string and gst_caps_from_string.
When gst_structure_to_string and gst_caps_to_string are called, no
escaping is performed after calling gst_value_serialize. Therefore, by
unescaping the value string, we were introducing an additional operation
that was not performed by the original *_to_string functions. In
particular, this has meant that the derialization functions for many
non-basic types are incomplete reverses of the corresponding
serialization function (i.e., if you pipe the output of the
serialization function into the deserialization function it could fail)
because they have to compensate for this additional escaping operation,
when really this should be the domain of the deserialization functions
instead.
Correspondingly changed a few deserialization functions.
Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/issues/452
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/303>
This reduces the chance of the main thread getting starved while trying
to shut down the test, potentially causing a timeout.
Even on an idle 96-processor system this reduces the duration of the
systemclock tests from ~8s to ~3s.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/734>
Until now we were enforcing that only 1 signal GSource was attached
the bus but we could attach as many GSource with `gst_bus_create_watch`
as we wanted... but in the end only 1 GSource will ever be dispatched for
a given `GstMessage` leading to totally broken behavior.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/718>
The GIR parser does not want any empty line after the function or macro
name line.
Fixes the following warning:
[309/4246] Generating Gst-1.0.gir with a custom command
../subprojects/gstreamer/gst/gstelement.h:57: Warning: Gst: "@element" parameter unexpected at this location:
* @element: The element name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gstelement.h:84: Warning: Gst: "@e" parameter unexpected at this location:
* @e: The element name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gstelement.h:106: Warning: Gst: "@e" parameter unexpected at this location:
* @e: The element name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gstdeviceprovider.h:32: Warning: Gst: "@d_p" parameter unexpected at this location:
* @d_p: The device provider name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gstdynamictypefactory.h:28: Warning: Gst: "@t_n" parameter unexpected at this location:
* @t_n: The dynamic type name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gsttypefind.h:34: Warning: Gst: "@type_find" parameter unexpected at this location:
* @type_find: The type find name in lower case, with words separated by '_'.
^
../subprojects/gstreamer/gst/gsttypefind.h:61: Warning: Gst: "@t_f" parameter unexpected at this location:
* @t_f: The type find name in lower case, with words separated by '_'.
^
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/719>
This introduces a more human friendly syntax to specify nested
structures It does so by using 2 different markers for opening and
closing them instead of abusing quotes which lead to requiring an insane
amount of escaping to match nesting levels.
The brackets (`[` and `]`) have been chosen as they avoid complex
constructions with curly brackets (or lower/higher than signs) where you
could have structures embedded inside arrays (which also use curly
brackets), ie. `s, array=(structure){{struct}}` should be parsed as an
array of structures, but the cast seems to imply something different. We
do not have this issue with brackets as they are currently used for
ranges, which can only be casted to numeric types.
This commit does not make use of that new syntax for serialization as
that would break backward compatibility, so it is basically a 'sugar'
syntax for humans. A notice has been explicitly made in the
documentation to let the user know about it.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/532>
The various wait implementation have a latency ranging from 50 to 500+
microseconds. While this is not a major issue when dealing with a low number of
waits per second (for ex: video), it does introduce a non-negligeable jitter for
synchronization of higher packet rate systems.
The `clock_nanosleep` syscall does offer a lower-latency waiting system but is
unfortunately blocking, so we don't want to use it in all scenarios nor for too
long.
This patch makes GstSystemClock use clock_nanosleep (if available) as such:
* Any wait below 500us uses it
* Any wait below 2ms will first use the regular waiting system and then
clock_nanosleep
# modified: gst/gstsystemclock.c
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/688>
While the default implementation will spawn a thread per new
pushed task, this new implementation instead spawns a maximum
number of threads, then queues new tasks on existing threads.
The thread that the new task will be queued on is picked in
a pretty naive fashion, by simply popping the first thread
from a queue and pushing it back to the tail, but this is
an implementation detail and can always be sophisticated
in the future if the need arises.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/692>
While the default implementation passes NULL around as the
task handle, other implementations can only provide a safe
API by having that handle map to a refcounted opaque type.
While what's passed around is a gpointer, a valid transfer
type annotation has informative value.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/692>
