Sebastian Dröge [Sun, 28 Mar 2010 16:05:36 +0000 (18:05 +0200)]
gst: Use GSlice instead of normal g_malloc in more places
Stefan Kost [Sun, 28 Mar 2010 10:14:06 +0000 (13:14 +0300)]
build: more some prototypes out if #ifndef GST_DISABLE_GST_DEBUG
The build was failing becasue of a new warning. There are still failures
(tracked via bug #614167).
Tim-Philipp Müller [Thu, 25 Mar 2010 20:04:37 +0000 (20:04 +0000)]
po: update translations for newly-added strings
Tim-Philipp Müller [Thu, 25 Mar 2010 19:56:00 +0000 (19:56 +0000)]
docs: fix intermittent make distcheck failures
Use .NOTPARALLEL when building docs. This avoids intermittent
make distcheck failures like 'cp: cannot create regular file
`build/image.entities': File exists' when using -jN.
Fixes #590718.
Sebastian Dröge [Thu, 25 Mar 2010 17:57:40 +0000 (18:57 +0100)]
elementfactory: Add FIXME 0.11 to remove GstElementDetails from the public API
It's not necessary anymore to expose this as public API and this allows
easier extension of the element details by new fields.
Wim Taymans [Thu, 25 Mar 2010 17:43:04 +0000 (18:43 +0100)]
queue2: handle write errors
Handle write errors to the temporary download file and post errors when
something went wrong.
Wim Taymans [Thu, 25 Mar 2010 17:13:02 +0000 (18:13 +0100)]
queue2: add element query function
Add an element query function that is a little more efficient than the generic
default query handler.
Wim Taymans [Thu, 25 Mar 2010 17:12:06 +0000 (18:12 +0100)]
bin: improve docs a little
Mention that a DURATION message does not mean that one can safely query the
duration on a bin, that only works when the bin is prerolled.
Wim Taymans [Thu, 25 Mar 2010 17:05:28 +0000 (18:05 +0100)]
queue2: remove fixed FIXME
Wim Taymans [Thu, 25 Mar 2010 16:36:45 +0000 (17:36 +0100)]
queue2: add the buffering percent in BUFFERING query
Wim Taymans [Thu, 25 Mar 2010 16:21:02 +0000 (17:21 +0100)]
queue2: improve buffer level measurement in download mode
Keep track of the current buffer level in the current range in download mode so
that we post the correct buffering messages.
Tim-Philipp Müller [Thu, 25 Mar 2010 15:54:19 +0000 (15:54 +0000)]
libs: don't use fancy shell features when invoking gobject-introspection scanner
It's POSIX, but tcsh doesn't seem to support it.
Tim-Philipp Müller [Thu, 25 Mar 2010 13:46:35 +0000 (13:46 +0000)]
libs: fix PKG_CONFIG_PATH used when calling gobject-introspection scanner
Our own pkgconfig directory should come first, so that pkg-config uses
the in-tree libgstreamer and not some external one when --pkg=gstreamer-0.10
is passed to g-ir-scanner.
See #603710.
Edward Hervey [Thu, 25 Mar 2010 09:27:00 +0000 (10:27 +0100)]
GstAdapter: add a unchecked variant of flush for internal usage
Trims off 10-20% cpu time when using gst_adapter_take[_buffer]
Thiago Santos [Fri, 19 Mar 2010 18:10:07 +0000 (15:10 -0300)]
tags: Add new _USER_RATING tag
Adds a new tag for user favorite media rating.
User rating informs how much (from 0 to 100) a user
'likes' a media.
Having an percent uint range for this is easy to map into other scales,
like some players that allow users to attribute 'stars' to its
media.
API: GST_TAG_USER_RATING
Fixes #520697
Wim Taymans [Wed, 24 Mar 2010 18:02:12 +0000 (19:02 +0100)]
queue2: add more info in the buffering query
Add the estimated download time and estimated time left to the buffering query
results along with the estimated download and playback speed.
Wim Taymans [Wed, 24 Mar 2010 17:18:13 +0000 (18:18 +0100)]
queue2: implement flushing in download buffering
Maintain a separate variable to control src and sink flowreturn values so that
we can unlock the src part without shutting down the sink part.
Add flushing for upstream pull based elements that unblocks our getrange
function. This implements seeking when blocking for more data.
Add some arbitrary threshold before attempting a seek. Add a FIXME for this
because we need to find a sensible threshold based on the input rate.
Edward Hervey [Wed, 24 Mar 2010 17:50:02 +0000 (18:50 +0100)]
Automatic update of common submodule
From 55cd514 to c1d07dd
Benjamin Otte [Wed, 24 Mar 2010 16:32:54 +0000 (17:32 +0100)]
Remove unused code
OPT_CFLAGS was never AC_SUBST()'ed so it wasn't used. And the last time
it was touched was in 2005.
Tim-Philipp Müller [Wed, 24 Mar 2010 15:47:22 +0000 (15:47 +0000)]
build: fix make distcheck
The change from GST_ALL_CFLAGS to GST_OPTION_CFLAGS dropped the includes,
putting them back fixes make distcheck.
Benjamin Otte [Wed, 24 Mar 2010 14:15:23 +0000 (15:15 +0100)]
Fix tests after set_element_details() deprecation
Robert Swain [Wed, 24 Mar 2010 12:33:58 +0000 (13:33 +0100)]
git-update: Fix and restructure logic
Benjamin Otte [Fri, 19 Mar 2010 21:36:07 +0000 (22:36 +0100)]
Deprecated gst_element_class_set_details()
Use gst_element_class_set_details_simple() instead. If you want to
convert automatically, here's a script:
for file in `git grep -l GstElementDetails`; do
sed -i -n -r '
1h
1!H
$ {
g
s/((\/\*[^\n]*\*\/)?\n)*[^\n]*GstElementDetails .* =\s*GST_ELEMENT_DETAILS\s*\((\"[^\"]*\",\s*\"[^\"]*\",\s*\"[^\"]*\",\s*(\"[^\"]*\"\s*)*)\);\n*(.*)gst_element_class_set_details \(([^,]*),\s*[^)]*\)/\n\n\5gst_element_class_set_details_simple (\6, \3)/
s/((\/\*[^\n]*\*\/)?\n)*[^\n]*GstElementDetails .* =\s*\{\s*(\"[^\"]*\",\s*\"[^\"]*\",\s*\"[^\"]*\",\s*(\"[^\"]*\"\s*)*)\};\n*(.*)gst_element_class_set_details \(([^,]*),\s*[^)]*\)/\n\n\5gst_element_class_set_details_simple (\6, \3)/
p
}' $file
~/gst/gstreamer/tools/gst-indent $file
done
Philippe Normand [Wed, 24 Mar 2010 09:57:08 +0000 (10:57 +0100)]
queue2: Fix uninitialized variable compiler warning
Tim-Philipp Müller [Tue, 23 Mar 2010 19:25:08 +0000 (19:25 +0000)]
tests: remove unused CHECK_CFLAGS and CHECK_LIBS from Makefile.am
Tim-Philipp Müller [Tue, 23 Mar 2010 19:23:22 +0000 (19:23 +0000)]
structure: add mapping for (uint) to allow deserialisation of unsigned integers
Unsigned ints are used in taglists, would be nice to be able to
deserialise them, esp. in connection with the taginject API.
Wim Taymans [Tue, 23 Mar 2010 18:25:29 +0000 (19:25 +0100)]
queue2: implement seeking in download mode
When in download mode and the requested offset is too far away, attempt to do a
seek request to fetch the data.
Keep track of all downloaded parts and merge ranges when needed.
Fixes #600877
André Dieb Martins [Mon, 22 Mar 2010 14:06:21 +0000 (11:06 -0300)]
git-update.sh: use autogen.sh instead of autoregen.sh on fresh repositories.
Fixes #613593.
Stefan Kost [Mon, 22 Mar 2010 15:10:06 +0000 (17:10 +0200)]
gettext: build fixes: #if -> #ifdef
Stefan Kost [Thu, 18 Mar 2010 08:22:09 +0000 (10:22 +0200)]
parse-launch: make delayed set recursive
Right now deleyed set would only try for first set of children. We need to keep
trying to support arbitrary deep hierarchies (like in playbin2 with auto*sinks).
Also GstBin would need to actualy emit the child-added/removed signal as it
implements the iface. Fixes #613215.
Sebastian Dröge [Mon, 22 Mar 2010 07:40:34 +0000 (08:40 +0100)]
pkgconfig: Use @LIBM@ instead of -lm
Tim-Philipp Müller [Fri, 19 Mar 2010 01:02:30 +0000 (01:02 +0000)]
pkgconfig: add back support for builddir != srcdir case in uninstalled setup
Attempt to add back support for builddir != srcdir. Use absolute paths
instead of relative paths based on pcfiledir this time to make things
clearer - there's not really any need for uninstalled trees to be
relocatable without re-running configure.
Robert Swain [Thu, 18 Mar 2010 11:18:39 +0000 (11:18 +0000)]
basetransform: Implement QoS message posting
And some more for bug #322947
Wim Taymans [Thu, 18 Mar 2010 12:41:35 +0000 (13:41 +0100)]
basesrc: catch, parse and store QoS event values
Catch, parse and store the QoS values from QoS events for later use.
Robert Swain [Wed, 17 Mar 2010 15:29:48 +0000 (15:29 +0000)]
basesink: Implement QoS message posting in basesink
Post QoS messages when frames are dropped.
This goes a little further towards resolving bug #322947
Wim Taymans [Wed, 17 Mar 2010 18:26:30 +0000 (19:26 +0100)]
message: improve docs a little
Robert Swain [Wed, 17 Mar 2010 18:16:42 +0000 (19:16 +0100)]
message: add QoS message to inform apps of lost data
This has been implemented as per part-qos.txt and partially addresses
bug #322947
Benjamin Otte [Wed, 17 Mar 2010 15:44:05 +0000 (16:44 +0100)]
test: Remove needless cast
GstValueArray.name is const now
Tim-Philipp Müller [Wed, 17 Mar 2010 14:33:35 +0000 (14:33 +0000)]
Revert "Add srcdir to includes for out-of-source builds"
I don't know how this ever worked, as it seems to put -I./..
and -I./../libs verbatim into the includes, at least with
current autotools versions.
This reverts commit
279a80ff27ba4c90f52981e89e710eb1181bc201.
Wim Taymans [Wed, 17 Mar 2010 14:46:46 +0000 (15:46 +0100)]
docs: avoid confusion between events and messages
Sebastian Dröge [Wed, 17 Mar 2010 11:18:18 +0000 (12:18 +0100)]
controller: Mark property_name in GstValueArray as const
This won't and should not be changed from any API
Mart Raudsepp [Wed, 17 Mar 2010 10:45:49 +0000 (12:45 +0200)]
docs: fix typo
Mart Raudsepp [Tue, 16 Mar 2010 07:56:16 +0000 (09:56 +0200)]
docs: add missing "Since: 0.10.26" marker for gst_data_queue_new_full()
Mart Raudsepp [Tue, 16 Mar 2010 07:42:21 +0000 (09:42 +0200)]
docs: fix since tag for gst_structure_id_has_field_typed()
gst_structure_id_has_field_typed was added in 0.10.26, not 0.10.16.
Apparent typo in commit
f9e3b72f when the API was added.
Tim-Philipp Müller [Wed, 17 Mar 2010 10:53:19 +0000 (10:53 +0000)]
docs: fix Since markers for gst_byte_writer_put_float*()
As the headers were broken in 0.10.26 the functions weren't really
usable back then, so we should advertise them as being there only
since 0.10.27.
Spotted by Mart Raudsepp.
Tim-Philipp Müller [Tue, 16 Mar 2010 17:56:13 +0000 (17:56 +0000)]
gst: use #if GLIB_CHECK_VERSION instead of #ifndef GLIBMACRO
for better greppability at the time we bump GLib version requirements.
Edward Hervey [Sun, 14 Mar 2010 10:47:23 +0000 (11:47 +0100)]
typefind: deactivate pad if we can't get length or it's a length of zero.
Fixes issues when re-using typefind after a file of length zero.
Thiago Santos [Tue, 16 Mar 2010 13:32:12 +0000 (10:32 -0300)]
basetransform: Accept non-fixed caps suggestions
When doing pad_allocs, use non-fixed caps suggestions and
try to fixate them before using. This makes possible to
have suggested buffer size with 0 in basetransform just
to signal upstream a renegotiation is needed
Fixes #576234
Fixes #609046
Wim Taymans [Tue, 16 Mar 2010 09:40:40 +0000 (10:40 +0100)]
docs: merge QoS message fields
There was already a section about QoS messages that is now merged with the new
information.
Stefan Kost [Tue, 16 Mar 2010 08:44:01 +0000 (10:44 +0200)]
pwg: mention how to build after using the project stamp
Stefan Kost [Tue, 16 Mar 2010 07:56:41 +0000 (09:56 +0200)]
task: snprintf needs to include "stdio.h"
Wim Taymans [Mon, 15 Mar 2010 17:48:14 +0000 (18:48 +0100)]
docs: update QOS docs to include QOS messages
Add some docs about the values needed for a QoS message and some use
cases.
See #322947
Wim Taymans [Mon, 15 Mar 2010 16:07:59 +0000 (17:07 +0100)]
task: use bionic/libc friendly arguments to prctl
prctl is supposed to take 5 arguments. It used to work with 2 arguments on some
versions of libc because it is defined as a varags function there.
See #611911
Wim Taymans [Mon, 15 Mar 2010 14:07:35 +0000 (15:07 +0100)]
task: update docs.
Sebastian Dröge [Mon, 15 Mar 2010 13:44:51 +0000 (14:44 +0100)]
taglist: Work around gtk-doc problem
Sebastian Dröge [Mon, 15 Mar 2010 13:44:16 +0000 (14:44 +0100)]
bytewriter: Use correct gtk-doc workaround
Sebastian Dröge [Mon, 15 Mar 2010 13:05:35 +0000 (14:05 +0100)]
utils: Handle iterator resyncs in gst_pad_proxy_setcaps()
Fixes bug #612881.
Wim Taymans [Mon, 15 Mar 2010 13:48:19 +0000 (14:48 +0100)]
task: configure the object name as thread name
When we have prctl available, use it to set the configured object name as the
thread name for better debugging.
Based on patch by Robert Swain.
Wim Taymans [Mon, 15 Mar 2010 12:48:30 +0000 (13:48 +0100)]
pad: set a good name on the task of the pad
Use the element:pad names to configure a good name for the pad task.
Wim Taymans [Mon, 15 Mar 2010 11:13:55 +0000 (12:13 +0100)]
task: retab
Stefan Kost [Mon, 15 Mar 2010 08:26:18 +0000 (10:26 +0200)]
logging: remove extra newline
Stefan Kost [Thu, 4 Mar 2010 21:36:50 +0000 (23:36 +0200)]
bus: turn g_return_if_fail into g_assert.
This either must never happen (which makes sense in this case) and thus should
use assert() or we should use a traditional if (poll_data->message) return;
to avoid differnet behaviour of intenal api when compiling with
G_DISABLE_CHECKS.
Stefan Kost [Wed, 3 Mar 2010 21:37:01 +0000 (23:37 +0200)]
caps: move the check to the public api.
This avoids creating empty caps and destroying them in the case of an error. We
also avoid double checking in other code path where we call the internal api.
Stefan Kost [Wed, 3 Mar 2010 21:34:58 +0000 (23:34 +0200)]
caps: this is internal API where we need to ensure !NULL higher up
Stefan Kost [Wed, 3 Mar 2010 19:50:21 +0000 (21:50 +0200)]
gst: this is an internal function where we already ensure !NULL when calling
Stefan Kost [Fri, 26 Feb 2010 07:29:23 +0000 (09:29 +0200)]
debugutils: fix case of pad flag
Due to a typo the code was always showing the flag as 's' (lower case).
Fixes #611075
Stefan Kost [Tue, 23 Feb 2010 21:50:36 +0000 (23:50 +0200)]
debug: add pretty printer for events
Adder is using GST_PTR_FORMAT for events already, so we might actualy
implement this and print out some useful info.
Benjamin Otte [Sat, 13 Mar 2010 10:03:59 +0000 (11:03 +0100)]
Add some 0.11 FIXMEs for GstPluginInitFunc
See
8fe63000de31bb2bcf346d59230dea06117997cd for why having a TRUE/FALSE
return value is a bad idea.
I've scanned a few plugins and they generally get it wrong and aren't
unloadable when they return FALSE.
Tim-Philipp Müller [Fri, 12 Mar 2010 19:07:22 +0000 (19:07 +0000)]
po: update for new strings
Leo Singer [Fri, 12 Mar 2010 19:05:16 +0000 (19:05 +0000)]
gstinfo: fix compilation error if HAVE_REGISTER_PRINTF_SPECIFIER is undefined
Use #if HAVE_FOO instead of #ifdef HAVE_FOO.
Fixes #612733.
Benjamin Otte [Fri, 12 Mar 2010 15:42:47 +0000 (16:42 +0100)]
plugins: Do not ever unload a plugin after calling into it
This is what can happen in a plugin_init function:
- An element based on GstBaseSink is registered
- Other elements fail to register
- The plugin_init function returns FALSE
Now if this the plugin is the first plugin to link against
libgstbase.so, it will have caused libgstbase.so to be loaded and static
strings from that library will have been added to gobject while
registering GstBaseSink.
So unloading the plugin will cause those strings to go stale and the
next plugin using GstBaseSink will crash. So we must not unload modules
after calling into them ever.
https://bugzilla.redhat.com/show_bug.cgi?id=572800
Sebastian Dröge [Fri, 12 Mar 2010 14:36:38 +0000 (15:36 +0100)]
interpolationcontrolsource: Don't pass NULL to the GSequence API
Robert Swain [Fri, 12 Mar 2010 13:33:00 +0000 (13:33 +0000)]
git-update: Fix error return value and make the script exit on errors
Newer versions of BASH (4.x?) seem to dislike using -1 for a return. Even
though it's documented as being signed, BASH complains about it, so use
255 instead.
Edward Hervey [Fri, 12 Mar 2010 12:54:29 +0000 (13:54 +0100)]
Automatic update of common submodule
From e272f71 to 55cd514
Edward Hervey [Wed, 17 Feb 2010 12:02:43 +0000 (13:02 +0100)]
gst: Use G_VALUE_COLLECT_INIT if available
This brings total call speedups between 5% and 25%.
gst_caps_set_simple_valist: +5%
gst_structure_set_valist: + 10%
gst_structure_id_set_valist: +25%
gst_tag_list_add_valist: +5%
Measured using valgrind when run over the discovery of 200 media files.
Fixes #610256
Benjamin Otte [Thu, 11 Mar 2010 19:29:29 +0000 (20:29 +0100)]
win32: Fix build failures of tests
Thiago Santos [Tue, 9 Mar 2010 23:38:47 +0000 (20:38 -0300)]
tags: Adds new geo location tags
Adds new tags GST_TAG_GEO_LOCATION_COUNTRY,
GST_TAG_GEO_LOCATION_CITY and GST_TAG_GEO_LOCATION_SUBLOCATION.
API: GST_TAG_GEO_LOCATION_COUNTRY
API: GST_TAG_GEO_LOCATION_CITY
API: GST_TAG_GEO_LOCATION_SUBLOCATION
Fixes #612410
Benjamin Otte [Thu, 11 Mar 2010 17:36:32 +0000 (18:36 +0100)]
win32: Add prototype for DllMain()
Edward Hervey [Thu, 11 Mar 2010 10:46:09 +0000 (11:46 +0100)]
Update .po files
Edward Hervey [Mon, 28 Dec 2009 16:25:20 +0000 (17:25 +0100)]
gstreamer: remove unneeded casts
G_PARAM_SPEC_VALUE_TYPE does an expensive type check, whereas the
value_type field is a public field, so we can just use it directly.
Benjamin Otte [Thu, 11 Mar 2010 10:39:40 +0000 (11:39 +0100)]
Remove -Winline flag again
It triggers for a lot of GStreamer API (even though those triggers are
wrong most of the time).
I missed it because it only triggers with -O2, and I was using -O0.
Benjamin Otte [Thu, 11 Mar 2010 10:20:35 +0000 (11:20 +0100)]
Automatic update of common submodule
From df8a7c8 to e272f71
Benjamin Otte [Thu, 11 Mar 2010 10:10:44 +0000 (11:10 +0100)]
Add a bunch more warning flags to configure
None of these flags cause warnings anymore, so no fixes necessary.
The flags are:
-Wformat-nonliteral
-Wformat-security
-Wold-style-definition
-Wcast-align
-Winline
-Winit-self
-Wmissing-include-dirs
-Waddress
-Waggregate-return
-Wno-multichar
-Wnested-externs
Benjamin Otte [Tue, 2 Mar 2010 21:58:06 +0000 (22:58 +0100)]
Fixes for -Wmissing-declarations -Wmissing-prototypes
Also adds those flags to the configure warning flags
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Tim-Philipp Müller [Thu, 11 Mar 2010 09:39:23 +0000 (09:39 +0000)]
buffer: fix printf format
Use %u to print unsigned integers.
Stefan Kost [Thu, 11 Mar 2010 08:29:23 +0000 (10:29 +0200)]
tests: cast the arg. to fix the build with new compiler opts.
This is ugly. I am not sure if we really want to have such casts all over the
place.
Stefan Kost [Thu, 11 Mar 2010 08:14:05 +0000 (10:14 +0200)]
i18n: define dummy ngettext if i18n is disabled.
We cannot blindly use gettext function and not define them when not using gettext.
Stefan Kost [Thu, 11 Mar 2010 08:00:45 +0000 (10:00 +0200)]
i18n: fix the build with i18n disabled.
Don't include gettext.h if !ENABLE_NLS.
Stefan Kost [Thu, 4 Mar 2010 08:44:52 +0000 (10:44 +0200)]
buffer: allow configurable memory alignment. Fixes #596832
The alignment guaranteed by malloc is not always sufficient. E.g. vector
instructions or hardware subsystems want specifically aligned buffers. The
attached patch will use posix_memalign if available to allocate buffers.
The desired alignment can be set when running configure using the new
--with-buffer-alignment option.
Benjamin Otte [Wed, 10 Mar 2010 20:51:50 +0000 (21:51 +0100)]
Automatic update of common submodule
From 9720a7d to df8a7c8
Benjamin Otte [Wed, 10 Mar 2010 19:52:06 +0000 (20:52 +0100)]
Merge branch 'work'
Benjamin Otte [Wed, 3 Mar 2010 10:45:38 +0000 (11:45 +0100)]
Fixes for -Wwrite-strings
This changes some APIs in compatible ways:
- Some functions now take "const char *" arguments, not "char *"
- Some structs now have "conts char *" members, not "char *"
The changes may cause warnings when compiling with the right warning
flags. You've been warned.
Also adds -Wwrite-strings as a warning flag in configure.ac.
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Wed, 3 Mar 2010 09:31:26 +0000 (10:31 +0100)]
Fixes -Wundef warnings
... and adds that flag to configure.ac
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Wed, 3 Mar 2010 09:31:05 +0000 (10:31 +0100)]
benchmarks: Remove unneeded g_thread_exit()
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Wed, 3 Mar 2010 09:26:14 +0000 (10:26 +0100)]
Fixes for -Wold-style-definition
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Tue, 2 Mar 2010 22:51:18 +0000 (23:51 +0100)]
Make code safe for -Wredundant-decls
Adds that warning to configure.ac
Includes a tiny change of the GST_BOILERPLATE_FULL() macro:
The get_type() function is no longer declared before being defined.
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Wed, 10 Mar 2010 19:43:35 +0000 (20:43 +0100)]
Automatic update of common submodule
From 0b6e072 to 9720a7d
Benjamin Otte [Wed, 3 Mar 2010 09:00:41 +0000 (10:00 +0100)]
Make sure generated code doesn't run with -Werror
https://bugzilla.gnome.org/show_bug.cgi?id=611692
Benjamin Otte [Wed, 10 Mar 2010 16:03:29 +0000 (17:03 +0100)]
Update to common/ changes to ERROR_CFLAGS
Benjamin Otte [Wed, 10 Mar 2010 18:17:42 +0000 (19:17 +0100)]
Revert "registry: remove unused function"
Turns out the function is not unused, but was in an #ifdef WIN32
section.
Whoops.
This reverts commit
57d5db424c68ab5a61f33ce36ce0179eb30251ac.