platform/upstream/gstreamer.git
23 years agoAdded ability to use .glade file in cwd
Erik Walthinsen [Sun, 31 Dec 2000 22:49:50 +0000 (22:49 +0000)]
Added ability to use .glade file in cwd

Original commit message from CVS:
Added ability to use .glade file in cwd

23 years agoStarted work on better ghostpad management, and started to build the low- level EOS...
Erik Walthinsen [Sun, 31 Dec 2000 22:18:05 +0000 (22:18 +0000)]
Started work on better ghostpad management, and started to build the low- level EOS mechanism.

Original commit message from CVS:
Started work on better ghostpad management, and started to build the low-
level EOS mechanism.

Also removed a couple of printouts that aren't needed any more.

23 years agoAdded stdlib.h to clean up exit() warnings
Erik Walthinsen [Sun, 31 Dec 2000 22:15:28 +0000 (22:15 +0000)]
Added stdlib.h to clean up exit() warnings

Original commit message from CVS:
Added stdlib.h to clean up exit() warnings

23 years agoNew test programs.
Erik Walthinsen [Sun, 31 Dec 2000 22:14:06 +0000 (22:14 +0000)]
New test programs.

Original commit message from CVS:
New test programs.

23 years agoI'm too lazy to comment this
Erik Walthinsen [Sun, 31 Dec 2000 22:13:31 +0000 (22:13 +0000)]
I'm too lazy to comment this

Original commit message from CVS:
*** empty log message ***

23 years agoConverted to a Makefile.am, fixed up the GstSrc thing, and removed a set to READY...
Erik Walthinsen [Sun, 31 Dec 2000 22:11:35 +0000 (22:11 +0000)]
Converted to a Makefile.am, fixed up the GstSrc thing, and removed a set to READY state before PLAYING that isn't nec...

Original commit message from CVS:
Converted to a Makefile.am, fixed up the GstSrc thing, and removed a set
to READY state before PLAYING that isn't necessary.

23 years agoVarious fixes to ensure that appropriate files are included in make dist.
Richard Boulton [Sun, 31 Dec 2000 17:32:46 +0000 (17:32 +0000)]
Various fixes to ensure that appropriate files are included in make dist.

Original commit message from CVS:
Various fixes to ensure that appropriate files are included in make dist.
Make dist now produces a nearly functioning distribution.

23 years agoUpdated the editor to use the capslist
Wim Taymans [Sun, 31 Dec 2000 17:32:16 +0000 (17:32 +0000)]
Updated the editor to use the capslist

Original commit message from CVS:
Updated the editor to use the capslist
Added change state in cdparanoia

23 years agoUpdated the docs
Wim Taymans [Sun, 31 Dec 2000 17:02:47 +0000 (17:02 +0000)]
Updated the docs

Original commit message from CVS:
Updated the docs
Updated gst/Makefile.am to add CFLAGS
Added test programs to tests/.cvsignore

23 years agoTest program for the pad factory
Wim Taymans [Sun, 31 Dec 2000 16:15:11 +0000 (16:15 +0000)]
Test program for the pad factory

Original commit message from CVS:
Test program for the pad factory

23 years agoAdded multiple caps to the pads
Wim Taymans [Sun, 31 Dec 2000 16:12:48 +0000 (16:12 +0000)]
Added multiple caps to the pads

Original commit message from CVS:
Added multiple caps to the pads
Extended the padfactory for the multiple caps
GstCaps now have a name so that future manipulations on them can be done by name
Updated the plugins for the new caps list

23 years agoAdd files in include to distributions.
Richard Boulton [Sun, 31 Dec 2000 16:07:04 +0000 (16:07 +0000)]
Add files in include to distributions.

Original commit message from CVS:
Add files in include to distributions.

23 years agoAdd gstpropsprivate.h to dist.
Richard Boulton [Sun, 31 Dec 2000 15:53:07 +0000 (15:53 +0000)]
Add gstpropsprivate.h to dist.

Original commit message from CVS:
Add gstpropsprivate.h to dist.

23 years agoAdd docs/fwg to DIST_SUBDIRS
Richard Boulton [Sun, 31 Dec 2000 15:19:12 +0000 (15:19 +0000)]
Add docs/fwg to DIST_SUBDIRS

Original commit message from CVS:
Add docs/fwg to DIST_SUBDIRS

23 years agoAdd docs/fwg/Makefile to things to be built.
Richard Boulton [Sun, 31 Dec 2000 15:18:08 +0000 (15:18 +0000)]
Add docs/fwg/Makefile to things to be built.

Original commit message from CVS:
Add docs/fwg/Makefile to things to be built.

23 years agoFix another bug in make install for fwg (complained becuase there were no pngs to...
Richard Boulton [Sun, 31 Dec 2000 14:22:59 +0000 (14:22 +0000)]
Fix another bug in make install for fwg (complained becuase there were no pngs to install).

Original commit message from CVS:
Fix another bug in make install for fwg (complained becuase there were no
pngs to install).
I should have got up more slowly today...

23 years agoFix .cvsignores
Richard Boulton [Sun, 31 Dec 2000 14:02:32 +0000 (14:02 +0000)]
Fix .cvsignores

Original commit message from CVS:
Fix .cvsignores

23 years agoFix bug in make install for manuals.
Richard Boulton [Sun, 31 Dec 2000 13:55:33 +0000 (13:55 +0000)]
Fix bug in make install for manuals.

Original commit message from CVS:
Fix bug in make install for manuals.
Added stuff to .cvsignore in fwg

23 years agoComplete rewrite of the launcher. Now supports pretty arbitrary pipelines.
Erik Walthinsen [Sun, 31 Dec 2000 10:46:16 +0000 (10:46 +0000)]
Complete rewrite of the launcher.  Now supports pretty arbitrary pipelines.

Original commit message from CVS:
Complete rewrite of the launcher.  Now supports pretty arbitrary pipelines.
Try '((fakesrc) ! identity ! (fakesink))' out.  Doesn't seem to recognize
pad names yet for some reason, so it goes with the first sink/src pad it
finds in a given element.  Other problems still remain, but the hard part
is done with.

23 years agoDisabled MMX optimized versions until they get fixed
Wim Taymans [Sat, 30 Dec 2000 19:37:48 +0000 (19:37 +0000)]
Disabled MMX optimized versions until they get fixed

Original commit message from CVS:
Disabled MMX optimized versions until they get fixed

23 years agoFixed autoplugging for avi
Wim Taymans [Sat, 30 Dec 2000 19:04:51 +0000 (19:04 +0000)]
Fixed autoplugging for avi

Original commit message from CVS:
Fixed autoplugging for avi
Fixed a bug in multiple typefind function loading: when a typefind is
requested for a mimetype, all plugins with a typefind function should be
loaded.

23 years agoFixed MPEG2 autoplugging
Wim Taymans [Sat, 30 Dec 2000 17:42:15 +0000 (17:42 +0000)]
Fixed MPEG2 autoplugging

Original commit message from CVS:
Fixed MPEG2 autoplugging
Added some more INFO to gstthread

23 years agoUpdated the makefile for new files
Wim Taymans [Sat, 30 Dec 2000 17:13:08 +0000 (17:13 +0000)]
Updated the makefile for new files

Original commit message from CVS:
Updated the makefile for new files

23 years agoDocs updates: remove sink, src, filter, connection added gstinfo to docs.
Wim Taymans [Sat, 30 Dec 2000 16:13:17 +0000 (16:13 +0000)]
Docs updates: remove sink, src, filter, connection added gstinfo to docs.

Original commit message from CVS:
Docs updates: remove sink, src, filter, connection
added gstinfo to docs.
Updated some core files for the docs
gsttypefind did not include gst_private
Added CFlags to Makefile.am in gst/elements

23 years agoAdded framework for writing filter writers guide.
Richard Boulton [Sat, 30 Dec 2000 15:36:24 +0000 (15:36 +0000)]
Added framework for writing filter writers guide.

Original commit message from CVS:
Added framework for writing filter writers guide.
Extracted much of the makefile from docs/manual into a shared file which
is included, so it can be re-used.

23 years agoFix a bug in manual/Makefile.am: bad shell script for make disting .ps files. Coupl...
Richard Boulton [Sat, 30 Dec 2000 14:46:09 +0000 (14:46 +0000)]
Fix a bug in manual/Makefile.am: bad shell script for make disting .ps files.   Couple of final fixes to libs/Makefil...

Original commit message from CVS:
Fix a bug in manual/Makefile.am: bad shell script for make disting .ps
files.   Couple of final fixes to libs/Makefile.am, too.
Add a few things to cvsignores.

23 years agoMakefile cleaned muchly - make should now cause gstreamer-libs docs to be created...
Richard Boulton [Sat, 30 Dec 2000 13:54:43 +0000 (13:54 +0000)]
Makefile cleaned muchly - make should now cause gstreamer-libs docs to be created.  Make dist may work too - untested...

Original commit message from CVS:
Makefile cleaned muchly - make should now cause gstreamer-libs docs to be
created.  Make dist may work too - untested.  No longer store tmpls in
CVS, either.
Only thing missing is make dependencies on the source files used to extract
the docs.

23 years agoRemove template files from CVS: they're autogenerated better.
Richard Boulton [Sat, 30 Dec 2000 13:14:09 +0000 (13:14 +0000)]
Remove template files from CVS: they're autogenerated better.

Original commit message from CVS:
Remove template files from CVS: they're autogenerated better.

23 years agoMakefile can now be used for a different manual by simply changing the initial variables
Richard Boulton [Sat, 30 Dec 2000 05:05:07 +0000 (05:05 +0000)]
Makefile can now be used for a different manual by simply changing the initial variables

Original commit message from CVS:
Makefile can now be used for a different manual by simply changing the
initial variables

23 years agoRemove old directories and files
Richard Boulton [Sat, 30 Dec 2000 04:40:07 +0000 (04:40 +0000)]
Remove old directories and files

Original commit message from CVS:
Remove old directories and files

23 years agoMisc tweaks to makefile for manual. Moved .figs into same directory so we can have...
Richard Boulton [Sat, 30 Dec 2000 04:36:40 +0000 (04:36 +0000)]
Misc tweaks to makefile for manual.  Moved .figs into same directory so we can have them built appropriately without ...

Original commit message from CVS:
Misc tweaks to makefile for manual.  Moved .figs into same directory so
we can have them built appropriately without having to hack makefiles
horribly.  Nearly have a core makefile which can be reused for other
manuals.

23 years agoScaled down the eps images to half their size
Wim Taymans [Sat, 30 Dec 2000 03:28:29 +0000 (03:28 +0000)]
Scaled down the eps images to half their size

Original commit message from CVS:
Scaled down the eps images to half their size

23 years agoAdded some INFO in gstelements
Wim Taymans [Sat, 30 Dec 2000 02:47:11 +0000 (02:47 +0000)]
Added some INFO in gstelements

Original commit message from CVS:
Added some INFO in gstelements
Reverted pad names in gstfakesrc back to what they were

23 years agoUpdated more files to use INFO.
Erik Walthinsen [Sat, 30 Dec 2000 02:41:15 +0000 (02:41 +0000)]
Updated more files to use INFO.

Original commit message from CVS:
Updated more files to use INFO.

23 years agoMinor neaten to makefile.
Richard Boulton [Sat, 30 Dec 2000 02:25:05 +0000 (02:25 +0000)]
Minor neaten to makefile.

Original commit message from CVS:
Minor neaten to makefile.

23 years agoConvert GstSrc's to GstElements.
Richard Boulton [Sat, 30 Dec 2000 01:53:05 +0000 (01:53 +0000)]
Convert GstSrc's to GstElements.

Original commit message from CVS:
Convert GstSrc's to GstElements.

23 years agoAdded configure test for liblame.
Richard Boulton [Sat, 30 Dec 2000 01:06:19 +0000 (01:06 +0000)]
Added configure test for liblame.

Original commit message from CVS:
Added configure test for liblame.

23 years agoAdd configure tests for cdparanoia
Richard Boulton [Sat, 30 Dec 2000 00:46:17 +0000 (00:46 +0000)]
Add configure tests for cdparanoia

Original commit message from CVS:
Add configure tests for cdparanoia

23 years agoFix a couple of MMX issues.
Richard Boulton [Fri, 29 Dec 2000 21:56:20 +0000 (21:56 +0000)]
Fix a couple of MMX issues.

Original commit message from CVS:
Fix a couple of MMX issues.

23 years agoInclude libgst.la when linking gbtest so it compiles.
Richard Boulton [Fri, 29 Dec 2000 21:50:18 +0000 (21:50 +0000)]
Include libgst.la when linking gbtest so it compiles.

Original commit message from CVS:
Include libgst.la when linking gbtest so it compiles.

23 years agoFixed buffer flag handling gstplay used old flag handling updated some plugins for...
Wim Taymans [Fri, 29 Dec 2000 19:45:45 +0000 (19:45 +0000)]
Fixed buffer flag handling gstplay used old flag handling updated some plugins for the new objects/error handling

Original commit message from CVS:
Fixed buffer flag handling
gstplay used old flag handling
updated some plugins for the new objects/error handling
Fixed a serious buffer error in gst_buffer_append

23 years agoUpdated some test programs
Wim Taymans [Fri, 29 Dec 2000 18:31:06 +0000 (18:31 +0000)]
Updated some test programs

Original commit message from CVS:
Updated some test programs

23 years agoFixed a serious bug in gst_props_new: properties with a 0 value causes a segfault
Wim Taymans [Fri, 29 Dec 2000 18:23:41 +0000 (18:23 +0000)]
Fixed a serious bug in gst_props_new: properties with a 0 value causes a segfault

Original commit message from CVS:
Fixed a serious bug in gst_props_new: properties with a 0 value causes
a segfault
Fixed a possible bug in gstautoplug when one of the constructed paths
was empty
Commented out a too liberal assert in gstscheduler.c
Added GST_BUFFER_DISCONTINUOUS
tried to fix asyndisksrc

23 years agoAdded DEBUG, INFO, and ERROR systems. Very little code is converted yet.
Erik Walthinsen [Fri, 29 Dec 2000 10:02:17 +0000 (10:02 +0000)]
Added DEBUG, INFO, and ERROR systems.  Very little code is converted yet.

Original commit message from CVS:
Added DEBUG, INFO, and ERROR systems.  Very little code is converted yet.
Policy decisions need to be made as to what kinds of cases get what kind
of response, and what the default ERROR response should be.  Right now it
will print out all the information, then have gdb spew a stack trace.

23 years agoUpdated copyright notices.
Erik Walthinsen [Fri, 29 Dec 2000 05:38:06 +0000 (05:38 +0000)]
Updated copyright notices.

Original commit message from CVS:
Updated copyright notices.

23 years agoAdd the pipeline to the thread after autoplugging.
Wim Taymans [Fri, 29 Dec 2000 04:36:54 +0000 (04:36 +0000)]
Add the pipeline to the thread after autoplugging.

Original commit message from CVS:
Add the pipeline to the thread after autoplugging.

23 years agoAdded a missing file...
Wim Taymans [Fri, 29 Dec 2000 03:05:47 +0000 (03:05 +0000)]
Added a missing file...

Original commit message from CVS:
Added a missing file...

23 years agoReparented everything to GstElement, removing GstSrc, GstSink, GstFilter, and GstConn...
Erik Walthinsen [Fri, 29 Dec 2000 02:28:04 +0000 (02:28 +0000)]
Reparented everything to GstElement, removing GstSrc, GstSink, GstFilter, and GstConnection.  Also fixed a bug in gst...

Original commit message from CVS:
Reparented everything to GstElement, removing GstSrc, GstSink, GstFilter,
and GstConnection.  Also fixed a bug in gst_bin_iterate_func.

23 years agomissed one file for copyright
Erik Walthinsen [Thu, 28 Dec 2000 22:13:35 +0000 (22:13 +0000)]
missed one file for copyright

Original commit message from CVS:
missed one file for copyright

23 years agoUpdated copyright in all the libgst files.
Erik Walthinsen [Thu, 28 Dec 2000 22:12:02 +0000 (22:12 +0000)]
Updated copyright in all the libgst files.

Original commit message from CVS:
Updated copyright in all the libgst files.
Created gst_private.h, set up all files to use it.

23 years agoUpdated the API docs.
Wim Taymans [Thu, 28 Dec 2000 21:42:23 +0000 (21:42 +0000)]
Updated the API docs.

Original commit message from CVS:
Updated the API docs.

23 years agoremoved some dead pad_pull/push code
Erik Walthinsen [Thu, 28 Dec 2000 21:12:40 +0000 (21:12 +0000)]
removed some dead pad_pull/push code

Original commit message from CVS:
removed some dead pad_pull/push code

23 years agoPending editor changes. fixes for autoplug of elementary MPEG1 video streams make...
Wim Taymans [Thu, 28 Dec 2000 02:51:07 +0000 (02:51 +0000)]
Pending editor changes. fixes for autoplug of elementary MPEG1 video streams make sure mp3 types get compiled videosi...

Original commit message from CVS:
Pending editor changes.
fixes for autoplug of elementary MPEG1 video streams
make sure mp3 types get compiled
videosink changes for xml loading

23 years agoFixes for the new scheduler.
Wim Taymans [Thu, 28 Dec 2000 00:18:26 +0000 (00:18 +0000)]
Fixes for the new scheduler.

Original commit message from CVS:
Fixes for the new scheduler.
Autoplug works again with mp3/mpeg1/ogg

23 years agoAdded some cleanup, so create_plan can be called multiple times.
Erik Walthinsen [Wed, 27 Dec 2000 23:42:15 +0000 (23:42 +0000)]
Added some cleanup, so create_plan can be called multiple times.

Original commit message from CVS:
Added some cleanup, so create_plan can be called multiple times.

23 years agoUpdates for the gst_pad_push changes
Wim Taymans [Wed, 27 Dec 2000 22:39:45 +0000 (22:39 +0000)]
Updates for the gst_pad_push changes

Original commit message from CVS:
Updates for the gst_pad_push changes

23 years agoAdded loop_based option
Wim Taymans [Wed, 27 Dec 2000 04:16:01 +0000 (04:16 +0000)]
Added loop_based option

Original commit message from CVS:
Added loop_based option
Added multiple pad output

23 years agoMostly minor little changes, but two interesting things: 1) removed a pthread_join...
Erik Walthinsen [Wed, 27 Dec 2000 03:13:20 +0000 (03:13 +0000)]
Mostly minor little changes, but two interesting things: 1) removed a pthread_join from the gst_thread_main_loop.  co...

Original commit message from CVS:
Mostly minor little changes, but two interesting things:

1) removed a pthread_join from the gst_thread_main_loop.  commented out
because the thread isn't supposed to run pthread_join, the main process is.

2) Fixed a major bug with cothreads in threads.  Had to add MAP_FIXED to
the mmap() of the cothread stack.  Presumably the gilbc that ships with
redhat 7.0 now places these mmap requests somewhat randomly.  Since they
*must* be exactly where we expect them, it was failing.  MAP_FIXED forces
it to put it where we say.

23 years agoAdded code to attempt the scheduling of chained cases. Very simply right now. Also...
Erik Walthinsen [Wed, 27 Dec 2000 00:46:26 +0000 (00:46 +0000)]
Added code to attempt the scheduling of chained cases.  Very simply right now.  Also added code to allow the use_coth...

Original commit message from CVS:
Added code to attempt the scheduling of chained cases.  Very simply right
now.  Also added code to allow the use_cothread flag of a bin to force all
the chains to use cothreads as well.

23 years agoadded case4
Erik Walthinsen [Wed, 27 Dec 2000 00:01:27 +0000 (00:01 +0000)]
added case4

Original commit message from CVS:
added case4

23 years agoMore massive changes to the scheduling system. Moved the scheduling code to gstsched...
Erik Walthinsen [Tue, 26 Dec 2000 23:51:04 +0000 (23:51 +0000)]
More massive changes to the scheduling system.  Moved the scheduling code to gstscheduler.[ch], so an child bin can r...

Original commit message from CVS:
More massive changes to the scheduling system.  Moved the scheduling code
to gstscheduler.[ch], so an child bin can replace the scheduler.
Introduced the concept of chains, which are subsets of the list of managed
elements for a given manager bin, which get scheduled as separate entities.
gst_bin_iterate_func should be pretty much fixed now, the scheduling code
gets to do all the hard work.

Cothreaded case work in the couple tests I've tried, chained is next.

23 years agoChanged "bin" to "thread" in get_thread_new.
Erik Walthinsen [Tue, 26 Dec 2000 20:49:46 +0000 (20:49 +0000)]
Changed "bin" to "thread" in get_thread_new.

Original commit message from CVS:
Changed "bin" to "thread" in get_thread_new.

23 years agoThis file will contain several use cases to reason through
Wim Taymans [Tue, 26 Dec 2000 20:24:47 +0000 (20:24 +0000)]
This file will contain several use cases to reason through

Original commit message from CVS:
This file will contain several use cases to reason through

23 years agoLots of editor changes:
Wim Taymans [Mon, 25 Dec 2000 01:33:06 +0000 (01:33 +0000)]
Lots of editor changes:

Original commit message from CVS:
Lots of editor changes:
- restucturing of object creation
- loading of xml files
- visual eye candy: color changes on mouseover
- active object is indicated
- property box changes for boolean and enum types
- property box: create properties per element
- show pads/caps in propertybox
added gst_util_get_bool_arg in gstutils.c
added default properties for audiosink

23 years agoAdded handoff signals to fakesrc and fakesink
Wim Taymans [Sat, 23 Dec 2000 03:17:52 +0000 (03:17 +0000)]
Added handoff signals to fakesrc and fakesink

Original commit message from CVS:
Added handoff signals to fakesrc and fakesink
Added scheduling property to identity
Added GST_STATE_TRANSITION macro to check for state changes.
Modified gstbin and gstthread to the new state change macros

23 years agoFixed a bug in gstbin.c the manager was not detected.
Wim Taymans [Fri, 22 Dec 2000 23:23:10 +0000 (23:23 +0000)]
Fixed a bug in gstbin.c the manager was not detected.

Original commit message from CVS:
Fixed a bug in gstbin.c the manager was not detected.
Some other testcases
bin, thread, pipeline use the elementfactory on gst_*_new

23 years agoImplemented some test cases... they all fail :-(
Wim Taymans [Fri, 22 Dec 2000 16:14:33 +0000 (16:14 +0000)]
Implemented some test cases... they all fail :-(

Original commit message from CVS:
Implemented some test cases... they all fail :-(
Added a flag to identity to for loop_based behaviour.
Added a signal to fakesink when a buffer is consumed.

23 years agoMore testcases
Wim Taymans [Thu, 21 Dec 2000 20:21:19 +0000 (20:21 +0000)]
More testcases

Original commit message from CVS:
More testcases

23 years ago70+ testcases to start with
Wim Taymans [Thu, 21 Dec 2000 19:53:56 +0000 (19:53 +0000)]
70+ testcases to start with

Original commit message from CVS:
70+ testcases to start with

23 years agoSome fixes to the non-chained case, it works in a very basic way. More use cases...
Erik Walthinsen [Thu, 21 Dec 2000 07:10:28 +0000 (07:10 +0000)]
Some fixes to the non-chained case, it works in a very basic way.  More use cases will hopefully provide some sanity.

Original commit message from CVS:
Some fixes to the non-chained case, it works in a very basic way.  More use
cases will hopefully provide some sanity.

23 years agoFixing two bugs: 1) I had commented out an unlock. duh.
Erik Walthinsen [Thu, 21 Dec 2000 01:27:27 +0000 (01:27 +0000)]
Fixing two bugs: 1) I had commented out an unlock.  duh.

Original commit message from CVS:
Fixing two bugs:

1) I had commented out an unlock.  duh.
2) changed the _get function to return the buffer rather than call _push

Also uncommented some debugging I'd turned off.  Need to solve the verbosity
problem somehow, I think by way of debug levels as well as info levels...

23 years agoMassive scheduling changes (again). Not entirely complete, but getting closer. ...
Erik Walthinsen [Wed, 20 Dec 2000 09:39:43 +0000 (09:39 +0000)]
Massive scheduling changes (again).  Not entirely complete, but getting closer.  Need to think about various scheduli...

Original commit message from CVS:
Massive scheduling changes (again).  Not entirely complete, but getting
closer.  Need to think about various scheduling plans that we might want
to produce, and figure out the rules for what is legal, and how to get
the results we need as far as the plan.

23 years agoAdded explanation for current autoplugging
Wim Taymans [Wed, 20 Dec 2000 00:55:45 +0000 (00:55 +0000)]
Added explanation for current autoplugging

Original commit message from CVS:
Added explanation for current autoplugging

23 years agoSmall test program for the property system
Wim Taymans [Tue, 19 Dec 2000 16:38:16 +0000 (16:38 +0000)]
Small test program for the property system

Original commit message from CVS:
Small test program for the property system

23 years agoAdded API to dynamically create GstCaps and GstProps
Wim Taymans [Tue, 19 Dec 2000 16:36:10 +0000 (16:36 +0000)]
Added API to dynamically create GstCaps and GstProps

Original commit message from CVS:
Added API to dynamically create GstCaps and GstProps
Changed typefind of mpeg1 to set the GstProps.
autoplugging now works again for mpeg1, whoohoo!

23 years agoAdded the autoplugger. some .dia drawings of the objects the autoplug test program.
Wim Taymans [Tue, 19 Dec 2000 13:44:23 +0000 (13:44 +0000)]
Added the autoplugger. some .dia drawings of the objects the autoplug test program.

Original commit message from CVS:
Added the autoplugger.
some .dia drawings of the objects
the autoplug test program.

23 years agoCleanup in gsttypes.c: removed the crazy GList of GHashTables, since the autoplugger...
Wim Taymans [Tue, 19 Dec 2000 13:41:55 +0000 (13:41 +0000)]
Cleanup in gsttypes.c: removed the crazy GList of GHashTables, since the autoplugger will use the GstCaps and element...

Original commit message from CVS:
Cleanup in gsttypes.c:
removed the crazy GList of GHashTables, since the autoplugger will
use the GstCaps and elementfactories instead of the type system.
We don't maintain a list anymore of the elements for the specific
removed unused code in XML loading/saving.

Cleanup in gstelementfactory:
removed the register/unregister methods, register is now implicit when
gst_elementfactory_new is called. _unregister is now _destroy.
Removed logic to register/unregister the types in gsttypes.
added methods to query if the factory can src/sink a GstCaps
Make sure the elementfactory is set in the element_class when a new
element is registered with gst_elementfactory_new.

gst.c: properly register the basic bins
gst_pipeline: use new gstautoplug (next checkin)
gstprops: fixed an error in compatibility check
registry test program changes
plugins: misc changes for the new caps system.

23 years agoThe typefind function now returns a GstCaps structure instead of a gboolean. modified...
Wim Taymans [Sun, 17 Dec 2000 16:24:14 +0000 (16:24 +0000)]
The typefind function now returns a GstCaps structure instead of a gboolean. modified some plugins to this new behaviour

Original commit message from CVS:
The typefind function now returns a GstCaps structure instead of a gboolean.
modified some plugins to this new behaviour
Fixed autoplugging for the mpg123 case. When an element is selected in
autoplugging and the caps do not match, another element is selected until the
caps match. only examples/autoplug works because gstmediaplay uses a threaded
setup that does not seem to work with the current scheduling.

23 years agoREADME and the beginning of the first code-review. Here's the README:
Erik Walthinsen [Sun, 17 Dec 2000 06:26:30 +0000 (06:26 +0000)]
README and the beginning of the first code-review.  Here's the README:

Original commit message from CVS:
README and the beginning of the first code-review.  Here's the README:

Code reviews:
=============

Files are to be named by file or subsystem, and CVS revision number or date:

gstbin.c-1.41
editor-20001216

A file should look something like the following:

----------------------------------------------------------------------
Code Review
===========
File: gst/gstbin.c
Revision: 1.41
Date: Dec 16, 2000
Reviewer: Erik Walthinsen <omega@cse.ogi.edu>

-----
Line 20:
#define GST_DEBUG_ENABLED

Shouldn't be here, DEBUG should be enabled globally.  May leave until
scheduling changes are done.

-----
Line 117: (gst_bin_class_init)
gstelement_class->elementfactory =    gst_elementfactory_find("bin");

Not sure this is such a great idea.  I thought the GstElement code did this
kind of stuff?

----------------------------------------------------------------------

The format will evolve as we do more stuff, such as putting in fields for
recommended actions, comments regarding any later changes made and when, etc.

23 years agoSome minor compilation fixes...
Erik Walthinsen [Sun, 17 Dec 2000 04:54:57 +0000 (04:54 +0000)]
Some minor compilation fixes...

Original commit message from CVS:
Some minor compilation fixes...

23 years agosome editor changes= loading of save pipelines (not working) added include in gstdebu...
Wim Taymans [Sat, 16 Dec 2000 17:12:28 +0000 (17:12 +0000)]
some editor changes= loading of save pipelines (not working) added include in gstdebug.h for getpid property loading ...

Original commit message from CVS:
some editor changes= loading of save pipelines (not working)
added include in gstdebug.h for getpid
property loading implemented
GstXML can now be queried for the toplevel elements.

23 years agoWARNING: Don't grab this updated unless you're really, REALLY sure.
Erik Walthinsen [Sat, 16 Dec 2000 10:18:09 +0000 (10:18 +0000)]
WARNING: Don't grab this updated unless you're really, REALLY sure.

Original commit message from CVS:
WARNING: Don't grab this updated unless you're really, REALLY sure.
WARNING: Wait for the next one.

Whole lotta changes here, including a few random bits:

examples/*/Makefile: updated to use `libtool gcc`, not just `gcc`
gst/
gstbuffer.h: updated to new flag style
gst.c, gstdebug.h: added new debugging for function ptrs
gstpipeline.c: set type of parent_class to the class, not the object
gstthread.c: ditto
plugins/
cdparanoia/cdparanoia.c: added an argument type, updated some defaults
cobin/spindentity.c: updated to new do/while loopfunction style
mp3encode/lame/gstlame.c: argument types, whole lotta lame options
tests/: various changes

Now, for the big changes:  Once again, the scheduling system has changed.
And once again, it broke a whole bunch of things.  The gist of the change
is that there is now a function pointer for gst_pad_push and gst_pad_pull,
instead of a hard-wired function.  Well, currently they are functions, but
that's for debugging purposes only, they just call the function pointer
after spewing lots of DEBUG().

This changed the GstPad structure a bit, and the GstPad API as well.
Where elements used to provide chain() and pull() functions, they provide
chain() and get() functions. gst_pad_set_pull[region]_function has been
changed to get_pad_set_get[region]_function.  This means all the elements
out there that used to have pull functions need to be updated.  The calls
to that function have been changed in the normal elements, but the names
of the functions passed is still _pull[region](), which is an aesthetic
issue more than anything.

As for what doesn't work yet, just about anything dealing with Connections
is hosed, meaning threaded stuff won't work.  This will be fixed about 12
hours from now, after I've slept, etc.  The simplefake.c test works in
both cothreaded and chained cases, but not much else will work due to the
Connection problem.  Needless to say, don't grab this unless you *need*
these features *now*, else wait to update this stuff until tomorrow.

I'm going to sleep now.

23 years agoMove a step closer to autoplugging. The element selection in the pipeline works.
Wim Taymans [Fri, 15 Dec 2000 16:43:26 +0000 (16:43 +0000)]
Move a step closer to autoplugging. The element selection in the pipeline works.

Original commit message from CVS:
Move a step closer to autoplugging. The element selection in the
pipeline works.
Slight modification to the typefind function handling.
removed the stupid mime types in the mpeg decoders.
Adjust videosink for the new caps handling.
Modified the queue example to show an error in plan creation; Erik can you
solve this? This is the type of pipelines autoplugging uses.

23 years agoHeader cleanup: try to include as little as possible; this will probably speed up...
Wim Taymans [Fri, 15 Dec 2000 01:57:34 +0000 (01:57 +0000)]
Header cleanup: try to include as little as possible; this will probably speed up compilation a bit.

Original commit message from CVS:
Header cleanup: try to include as little as possible; this will probably
speed up compilation a bit.
changed the .c files to use #include "..."
Fix for the 'plugins are loaded twice' bug.
Fix 22186: GstObject flags are now used everywhere. Added *_FLAG_LAST so
elements do not use the same flags. Added some padding in the flag enum
for future expansion.

23 years agoSome plugins are adjusted to the new pad template mechanisms.
Wim Taymans [Thu, 14 Dec 2000 17:21:29 +0000 (17:21 +0000)]
Some plugins are adjusted to the new pad template mechanisms.

Original commit message from CVS:
Some plugins are adjusted to the new pad template mechanisms.

23 years agoFirst few TODO items to be done after 0.1.0 (i.e. in 2001)
Erik Walthinsen [Wed, 13 Dec 2000 21:42:19 +0000 (21:42 +0000)]
First few TODO items to be done after 0.1.0 (i.e. in 2001)

Original commit message from CVS:
First few TODO items to be done after 0.1.0 (i.e. in 2001)

23 years agoChanged the GstPadFactory and added the GstPadTemplate.
Wim Taymans [Wed, 13 Dec 2000 19:29:35 +0000 (19:29 +0000)]
Changed the GstPadFactory and added the GstPadTemplate.

Original commit message from CVS:
Changed the GstPadFactory and added the GstPadTemplate.

The GstPadFactory is an easy way to define a pad with capabilities.
The GstPadFactory is converted into a GstPadTemplate. The template is
used to create new pads and to expose the possible pads used in an
element to the plugin system.

updated mp3parse, mpg123, cdparanoia to the new API.
Rerun gstreamer-register because the XML definition has changed.

23 years agoAdded a document about how we handle plugins and dynamic loading.
Wim Taymans [Tue, 12 Dec 2000 19:31:00 +0000 (19:31 +0000)]
Added a document about how we handle plugins and dynamic loading.

Original commit message from CVS:
Added a document about how we handle plugins and dynamic loading.

23 years agoMore Caps changes, this change introduces the padfactory. the GstPadFactory will...
Wim Taymans [Tue, 12 Dec 2000 19:29:43 +0000 (19:29 +0000)]
More Caps changes, this change introduces the padfactory. the GstPadFactory will be used to describe the pads that an...

Original commit message from CVS:
More Caps changes, this change introduces the padfactory. the GstPadFactory
will be used to describe the pads that an element will contain during
its lifetime.
mpg123 and mp3parse serve as an example.
More XML changes to incorporate the padfactories.

23 years agoAdded LAME encoder. Wow.
Erik Walthinsen [Tue, 12 Dec 2000 09:40:25 +0000 (09:40 +0000)]
Added LAME encoder.  Wow.

Original commit message from CVS:
Added LAME encoder.  Wow.

23 years agoBrand new source, the cdparanoia source. Reads audio data from a CD, writes out...
Erik Walthinsen [Tue, 12 Dec 2000 06:49:26 +0000 (06:49 +0000)]
Brand new source, the cdparanoia source.  Reads audio data from a CD, writes out raw audio.  The tests/paranoia.c pro...

Original commit message from CVS:
Brand new source, the cdparanoia source.  Reads audio data from a CD,
writes out raw audio.  The tests/paranoia.c program will simply hook this
up to a sound card.  It works perfectly.

Next step is to flesh out the rest of the element, including pad caps,
better seek and playout control, signals, and whatever else comes up.

A minor patch to the editor is included here, the GstElementFactory details
struct has a name change from 'class' to 'klass' that wasn't reflected
in the elementselect widget.  Fixd.

23 years agoUpdated to latest code for testing scheduling.
Erik Walthinsen [Mon, 11 Dec 2000 00:25:06 +0000 (00:25 +0000)]
Updated to latest code for testing scheduling.

Original commit message from CVS:
Updated to latest code for testing scheduling.

23 years agoImplemented the hybrid scheduling system for sources and connections outside the...
Erik Walthinsen [Mon, 11 Dec 2000 00:24:32 +0000 (00:24 +0000)]
Implemented the hybrid scheduling system for sources and connections outside the current Bin.  Is a bit hackish in on...

Original commit message from CVS:
Implemented the hybrid scheduling system for sources and connections
outside the current Bin.  Is a bit hackish in one place, but I'll work out
a way to make that cleaner soon.  queue.c in tests now works nicely in all
cases.  More to come later.

23 years agoTest program to check the registry handling
Wim Taymans [Mon, 11 Dec 2000 00:06:16 +0000 (00:06 +0000)]
Test program to check the registry handling

Original commit message from CVS:
Test program to check the registry handling

23 years agoFirst attempt at rebuilding the type/plugin system
Wim Taymans [Mon, 11 Dec 2000 00:04:25 +0000 (00:04 +0000)]
First attempt at rebuilding the type/plugin system

Original commit message from CVS:
First attempt at rebuilding the type/plugin system
- make sure caps/props are saved in the registry
autoplugging is completely broken.
removed the typefactories and typeids from all the pads in the plugins
XML load/save is ok (be sure to rerun gstreamer-register)

23 years agoSnapshot of work-in-progress do deal with out-of-bin elements. Current system is...
Erik Walthinsen [Sat, 9 Dec 2000 09:13:09 +0000 (09:13 +0000)]
Snapshot of work-in-progress do deal with out-of-bin elements.  Current system is very likely going to be ditched due...

Original commit message from CVS:
Snapshot of work-in-progress do deal with out-of-bin elements.  Current
system is very likely going to be ditched due to its complexity and the
fact that it doesn't work right now.  More than that, it's generating some
really odd results in my test program, which aren't readily explainable.

If you want to keep a working copy of your working copy (heh), don't update
until this is replaced with a hybrid approach.  This approach will be the
beginning of a plan generator that can construct full hybrid schedules
given hints from various places.

Hybrid means that cothreads are used, but there are chunks of the pipeline
that are actually dealt with by chaining. This can improve speed by
reducing cothread switches (which are much cheaper than thread switches,
but still far from free, about 570 cycles on my PIII), but does carry a
complexity burden.  Luckily, the structure of GStreamer allows that burden
to live entirely in create_plan.  Luck?  I think not ;-)

23 years agoMostly debugging changes. What's in CVS seems to work rather well, actually.
Erik Walthinsen [Fri, 8 Dec 2000 23:38:12 +0000 (23:38 +0000)]
Mostly debugging changes.  What's in CVS seems to work rather well, actually.

Original commit message from CVS:
Mostly debugging changes.  What's in CVS seems to work rather well, actually.

Next is the work on task 23098...

23 years agoTest program for simple queue operations. Good for testing thread/cothread interations.
Erik Walthinsen [Fri, 8 Dec 2000 20:31:40 +0000 (20:31 +0000)]
Test program for simple queue operations.  Good for testing thread/cothread interations.

Original commit message from CVS:
Test program for simple queue operations.  Good for testing thread/cothread
interations.

23 years agoAdded pullregion in again, it got removed with the previous commit.
Wim Taymans [Fri, 8 Dec 2000 18:24:16 +0000 (18:24 +0000)]
Added pullregion in again, it got removed with the previous commit.

Original commit message from CVS:
Added pullregion in again, it got removed with the previous commit.

23 years agoAnother big set of changes. Connections are now also pullfunc based. gstqueue has...
Erik Walthinsen [Fri, 8 Dec 2000 10:33:01 +0000 (10:33 +0000)]
Another big set of changes.  Connections are now also pullfunc based. gstqueue has been updated, I don't know of any ...

Original commit message from CVS:
Another big set of changes.  Connections are now also pullfunc based.
gstqueue has been updated, I don't know of any other connections offhand.

There are still a few things that need doing, specifically the concept
of a source or connection with connections to multiple thread contexts is
not dealt with.  This may force us to move the threadstate from the
element to the pad, maybe keeping the element's copy for simple cases.
Then the Bin would create a structure to pass to the cothreaded _wrappers
of any such elements, which would detail the pads that are to be dealt with
by this particular cothread context.

That will speed things up to, since we don't have to look through the list
of all pads for every Src or Connection element for every iteration, we can
simply step through the list provided by the plan.  Special case might even
have a single pad pointer sitting there to trump the list, if there's only
one (the common case anyway).

Task 23098 is tracking these changes.  The main task 22588 depends on that
subtask, as well as 22240, which is a consistency check on PAD_DISABLED.