Changes since 0.10.13:
* State change handling improvements
* New improved GstController API
* Multi-queue fixes and enhancements
* Fewer dynamic library relocations
* Various other fixes and improvements
* Parallel installability with 0.8.x series
* Threadsafe design and API
Bugs fixed since 0.10.13:
* 457322 : registry is read twice on startup
* 406632 : pkgconfigify the check for Check in configure.ac
* 423700 : [multiqueue] leaks memory when flushing
* 437457 : saving relocations for GstElementDetails
* 444648 : [API] add GstParamSpecFraction
* 444820 : build error with grammar.tab.c
* 448794 : state changes are racy with decodebin after rewrite for 0...
* 449197 : [capsfilter] Should set src pad caps on buffer
* 449226 : Incomplete ghostpads in xml session save
* 449493 : minor quoting fix in gst-feature.m4 macro
* 450711 : [GstController] Improve extensibility by providing a GstC...
* 451939 : [API] add GST_TAG_LICENSE_URI and GST_TAG_COPYRIGHT_URI
* 459501 : gst_registry_lookup_feature_locked is slow
* 462558 : Regression in async state change handling
* 462752 : make gst_element_class_set_details_simple args const
* 430682 : multiqueue doesn't output data on unlinked pads properly
* 462085 : registry feature hash can access invalid memory
* 357959 : buffer_alloc is not proxied through tee
* 374639 : GstTee doesn't handle pad list resyncs properly
* 449414 : [capsfilter] Copy of passed GstCaps taken instead of refe...
* 453732 : [PATCH] multiqueue returns wrong internally linked pads
package name. Otherwise, an error like the following occurs:
line 28: warning: "/usr/sbin/pkg_info -K /var/db/pkg -Q PKG_OPTIONS \
gstreamer 2>/dev/null || { cd && ...
changes:
-Add a warning if GLib threading support is not properly initialised
-Add support for binary registry serialisation
-Preliminary support for pull-mode scheduling of entire chains
-Add support for HPPA
-Collectpads fixes
-Lots of bug fixes
changes:
* The testsuite can now generate coverage information
* Better support for non-linear editing
* Removed Glib 2.6 compatibility
* Performance improvements on cpus that support branch hints
* Bug fixes
and add a new helper target and script, "show-buildlink3", that outputs
a listing of the buildlink3.mk files included as well as the depth at
which they are included.
For example, "make show-buildlink3" in fonts/Xft2 displays:
zlib
fontconfig
iconv
zlib
freetype2
expat
freetype2
Xrender
renderproto
changes:
-many bugfixes
-API functions added
-gcc-4.1 fixes
-New version of data protocol now serializes events
-Registry cache updating is now done in a fork, so no plugins are left
opened
PKGLOCALEDIR and which install their locale files directly under
${PREFIX}/${PKGLOCALEDIR} and sort the PLIST file entries. From now
on, pkgsrc/mk/plist/plist-locale.awk will automatically handle
transforming the PLIST to refer to the correct locale directory.
RECOMMENDED is removed. It becomes ABI_DEPENDS.
BUILDLINK_RECOMMENDED.foo becomes BUILDLINK_ABI_DEPENDS.foo.
BUILDLINK_DEPENDS.foo becomes BUILDLINK_API_DEPENDS.foo.
BUILDLINK_DEPENDS does not change.
IGNORE_RECOMMENDED (which defaulted to "no") becomes USE_ABI_DEPENDS
which defaults to "yes".
Added to obsolete.mk checking for IGNORE_RECOMMENDED.
I did not manually go through and fix any aesthetic tab/spacing issues.
I have tested the above patch on DragonFly building and packaging
subversion and pkglint and their many dependencies.
I have also tested USE_ABI_DEPENDS=no on my NetBSD workstation (where I
have used IGNORE_RECOMMENDED for a long time). I have been an active user
of IGNORE_RECOMMENDED since it was available.
As suggested, I removed the documentation sentences suggesting bumping for
"security" issues.
As discussed on tech-pkg.
I will commit to revbump, pkglint, pkg_install, createbuildlink separately.
Note that if you use wip, it will fail! I will commit to pkgsrc-wip
later (within day).
Changes since 0.10.3:
* Parallel installability with 0.8.x series
* Threadsafe design and API
* 33% of library size removed by parse change
* (Experimental) QoS features (use max-lateness on BaseSink subclasses to enable)
Bugs fixed since 0.10.3:
* 323542 : GStreamer 0.10 hangs at EOS for all songs
* 320340 : triggering a gst_task_join from the streaming task can de...
* 322628 : document thread-unsafety of gst_value_foo() + gst_value_r...
* 324186 : Smarter (deterministic!) typefind decisions
* 324398 : varargs/valist functions aren't bindable
* 328873 : only one gst_bin_recalc_state allowed at a time
* 330899 : CVS directory in make dist tarball
* 330906 : API: add GST_QUERY_TYPE_NAME and GST_MESSAGE_TYPE_NAME
* 330996 : basesrc emits EOS unconditionally when going to READY
* 331255 : Shrinking gst_parse_launch bloat
* 331898 : Basetranform is not reusable
* 332045 : API: GstPipeline needs API to disable flushing of its bu...
* 332277 : [filesrc] seek does not correct read position in non-mmap...
* 332611 : [basesrc] needs way to prevent expensive start/stop in ch...
* 332723 : New API: gst_type_find_helper_for_buffer
* 333042 : New API: gst_type_find_helper_get_range
* 333266 : Registry file writing rewrite
* 333272 : tests should set G_SLICE=always-malloc now
* 333417 : GST_BOILERPLATE_WITH_INTERFACE can't be used in C++ code
* 333669 : Add pad accessor defines for GstBaseTransform
* 333779 : New API: add gst_uri_has_protocol
* 324818 : gst_registry_get_default doesn't sink newly created registry
* 330125 : _pad_send_event does not check flushing correctly
* 330684 : Add ability to emit sync-message without messing with the...
API added since 0.10.3:
* gst_pipeline_set_auto_flush_bus()
* gst_pipeline_get_auto_flush_bus()
* gst_uri_has_protocol()
* gst_pipeline_set_max_lateness()
* gst_pipeline_get_max_lateness()
* gst_pipeline_get_sync()
* gst_pipeline_do_sync()
* GST_BASE_TRANSFORM_SRC_PAD()
* GST_BASE_TRANSFORM_SINK_PAD()
* GstBaseSrc::check_get_range()
* GstBaseSink::max-lateness
* gst_type_find_helper_get_range()
* gst_type_find_helper_for_buffer()
* GST_MESSAGE_TYPE_NAME()
* GST_QUERY_TYPE_NAME()
* gst_bus_enable_sync_message_emission()
* gst_bus_disable_sync_message_emission()
This is GStreamer 0.10.3, "Like a virgin"
Changes since 0.10.2:
* Documentation updates
* Bug fixes
* Base class improvements
* Extra utility API
* More elements ported from 0.8
* Fixes for win32 build.
Changes since 0.10.1:
* documentation updates
* bug fixes
Bugs fixed since 0.10.1:
* 325429 : GST_BOILERPLATE_FULL can't be used when compiling with -W...
* 316086 : MinGW compilation: fileindexer depends on mmap
gstreamer0.8 to the 0.10 series. Changes:
Changes since 0.10.0:
* element factory leak fixes
* tests cleanup
* BaseSrc seeking in other format than BYTES
* collectpad fixes
Bugs fixed since 0.10.0:
* 322441 : getcontext() in common/m4/gst-mcsc.m4 uses incorrect stac...
* 323041 : [BASESRC] basesrc does not handle a seek to non-bytes for...
* 323870 : [tools] gst-feedback needs updating for 0.9/0.10
* 323882 : gst-launch manpage needs updates
* 323905 : use return val of g_list_append
API added since 0.10.0:
* GST_TYPE_STATIC_CAPS
* GST_TYPE_STATIC_PAD_TEMPLATE
* GstNetTimeProvider::active
* gst_base_src_set_format
Changes from the 0.8 series to 0.10 series:
* Refcounting and API revised for thread safety, see rules in
docs/design/part-MT-refcounting.txt
* Scheduling is now done by the elements, they start/stop
processing threads themselves. This removes the need for
a separate scheduler.
* State changes now happen from sink to source. This makes sure
that downstream elements are ready to process data when upstream
starts producing.
* EOS/ERROR/... does not change the state of elements anymore. This
means that when the application sets the state on an element, it
remains in this state.
* GMainLoop integration. Information on the state of the pipeline
is now received in the mainloop via the GstBus.
* Events move separate from the datastream, this allows for both
in and out of sync delivery of events.
* Generic _pull() replaced by _pull_region() so that random access
can be made faster.
* data passing happens from PAUSED, called the preroll phase. In this
state no data is rendered but it is queued in the sinks.
* negotiation simplified. Caps are now attached to buffers and travel
along with the datastream.
* GstBuffer is now subclassable.
* Ghostpads now look and feel like real pads.
* New Seeking/flushing policy.
* Addition of base classes and helper functions in the core to abstract
away the harder parts of plugin development.