Original commit message from CVS:
2004-11-07 Martin Soto <martinsoto@users.sourceforge.net>
* gst/schedulers/fairscheduler.c:
* gst/schedulers/faircothreads.c:
* gst/schedulers/faircothreads.h:
New cothread based scheduler: Fair scheduler.
* gst/schedulers/gthread-cothreads.h:
Add the standard #if around the whole file.
Defining symbol GTHREAD_COTHREADS_NO_DEFINITIONS will now prevent
compilation of the functions defined in this file. This is
necessary to be able to use this file as a normal header.
* gst/schedulers/Makefile.am: Add compiling support for fair
scheduler.
* docs/gst/Makefile.am (IGNORE_HFILES): Exclude internal fair
scheduler cothreads layer from documentation generation.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (add_to_group),
(remove_from_group), (normalize_group), (group_migrate_connected),
(gst_opt_scheduler_iterate):
* testsuite/schedulers/.cvsignore:
* testsuite/schedulers/Makefile.am:
* testsuite/schedulers/queue_link.c: (main):
Added testcase for schduler segfault.
Fix scheduler segfault when removing a decoupled
entry point as the last element from a group.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (add_to_group),
(gst_opt_scheduler_iterate):
Aplied patch #154061. Running a pipeline in which an element
calls GST_ELEMENT_ERROR in the chain function, the opt
scheduler doesn't unref the chain so it never gets freed.
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* gst/gstscheduler.c: (gst_scheduler_register):
* gst/gstscheduler.h:
add gst_scheduler_register shortcut similar to gst_element_register
* gst/schedulers/entryscheduler.c: (plugin_init):
* gst/schedulers/gstbasicscheduler.c: (plugin_init):
* gst/schedulers/gstoptimalscheduler.c: (plugin_init):
use it
Original commit message from CVS:
* gst/schedulers/entryscheduler.c:
(gst_entry_scheduler_remove_element):
fix for GST_DISABLE_DEBUG
* tools/gst-launch.c: (print_tag):
fixes for G_DISABLE_ASSERT
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (remove_from_group),
(gst_opt_scheduler_pad_link):
fix 147894-2 and the group_link problem.
Original commit message from CVS:
* gst/schedulers/entryscheduler.c:
(gst_entry_scheduler_loop_wrapper),
(gst_entry_scheduler_chain_wrapper),
(gst_entry_scheduler_get_wrapper):
reset the state when the cothread starts, so we don't get assertion
failures on restarting of cothreads
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (gst_opt_scheduler_init),
(create_chain), (destroy_chain), (create_group), (destroy_group),
(add_to_group), (merge_groups), (group_elements), (group_inc_link),
(group_dec_link), (gst_opt_scheduler_pad_link),
(group_inc_links_for_element), (group_migrate_connected):
Call group_inc_link with the proper src->sink ordering --
break this, and we break sort_chain. patch from wingo for bug
147713.
Partially revert patch 1.89. When adding a loop based element to
the scheduler, the links to other groups are automatically followed
and incremented. This should not happen because the bin will call
pad_link explicitly for those connection, resulting in them counted
twice. Results in assertion failure on pipeline cleanup.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (group_dec_link),
(gst_opt_scheduler_pad_link), (group_migrate_connected):
Fix stupid warning when an element is to be migrated but
is already migrated.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (group_dec_link),
(gst_opt_scheduler_pad_link), (group_migrate_connected):
Make sure that a single non-loop-based element does not
end up in a group. This fixes the testsuite again.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (create_group),
(add_to_group), (merge_groups), (schedule_group),
(gst_opt_scheduler_get_wrapper), (group_elements),
(group_dec_link), (gst_opt_scheduler_pad_link),
(group_migrate_connected), (gst_opt_scheduler_pad_unlink),
(gst_opt_scheduler_iterate):
move isolated groups to a new chain.
Emit a warning instead of segfaulting in some error cases.
Fix a bug where the link count between groups was not calculated
correctly. Fixes#144510.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c:
(gst_opt_scheduler_pad_link), (group_elements_set_visited),
(element_get_reachables_func), (element_get_reachables),
(debug_element), (rechain_group), (group_migrate_connected),
(gst_opt_scheduler_pad_unlink):
Do not try to migrate decoupled elements to a new group since
they are not added to groups.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c:
(chain_recursively_migrate_group), (create_group),
(schedule_group), (gst_opt_scheduler_pad_link),
(group_elements_set_visited), (element_get_reachables_func),
(element_get_reachables), (group_can_reach_group), (debug_element),
(rechain_group), (group_migrate_connected),
(gst_opt_scheduler_pad_unlink):
* testsuite/schedulers/Makefile.am:
Implemented group splitting and rechaining.
Fixes 143777 and 143777-2 in the testsuite.
Original commit message from CVS:
* gst/autoplug/gstspideridentity.c: (gst_spider_identity_chain):
extra debugging
* gst/gstevent.h:
* gst/gstinfo.c: (gst_debug_log_default):
print time nicely. add thread pointer until someone figures out
a completely portable way of getting at thread id's.
* gst/gstpad.c: (gst_pad_set_active), (gst_pad_link_try),
(_invent_event), (gst_pad_pull), (gst_pad_event_default_dispatch),
(gst_pad_call_chain_function):
extra debugging
* gst/schedulers/gstoptimalscheduler.c:
(get_group_schedule_function), (loop_group_schedule_function),
(gst_opt_scheduler_loop_wrapper), (gst_opt_scheduler_get_wrapper),
(pad_clear_queued), (gst_opt_scheduler_iterate):
rename BUFPEN and friends to DATAPEN since that's what they are.
Original commit message from CVS:
- removing some trailing commas in enums to please C90 only compilers (AIX, Forte, ...)
('should' fix#143290)
should/could we compile gst with -ansi -pedantic to catch those in the build-bots ?
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (remove_from_chain),
(chain_group_set_enabled), (create_group), (add_to_group),
(merge_groups), (setup_group_scheduler), (group_elements),
(gst_opt_scheduler_iterate), (gst_opt_scheduler_show):
Don't try to follow the pad connections with other groups
when a loop based element is added to the scheduler because
the bin will inform the scheduler about the pad links a little
later.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (add_to_chain),
(remove_from_chain), (chain_group_set_enabled),
(setup_group_scheduler), (group_element_set_enabled),
(gst_opt_scheduler_state_transition), (gst_opt_scheduler_iterate),
(gst_opt_scheduler_show):
Elements without a group can do a state change as well, just wait
with the setup of the scheduling function when it is added to a
chain.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (add_to_chain),
(remove_from_chain), (chain_group_set_enabled), (add_to_group),
(merge_groups), (setup_group_scheduler),
(group_inc_links_for_element), (gst_opt_scheduler_iterate),
(gst_opt_scheduler_show):
Fixes to maintain internal consistency of the scheduler data
structures.
- adding an enabled group to a chain should increment the
number of enabled elements in that chain.
- removing an enabled group from a chain could disable the
chain.
- removing a disabled group from a chain could enable the
chain.
- add g_assert when internal inconsistency is detected.
- adding an element to a group could increase the number of
links this group has with other groups.
- merging two groups also merges the chains.
- also show group links in the _show method.
Please report regressions ASAP, so I can fix them.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_iterate_func), (gst_bin_iterate):
don't
* gst/schedulers/entryscheduler.c: (safe_cothread_switch),
(safe_cothread_destroy),
(gst_entry_scheduler_remove_all_cothreads),
(gst_entry_scheduler_reset), (_remove_cothread),
(gst_entry_scheduler_state_transition):
hold off cothread destruction if we're not in main cothread
* configure.ac:
* testsuite/Makefile.am:
add new test dir
* testsuite/schedulers/.cvsignore:
* testsuite/schedulers/Makefile.am:
add tests
* testsuite/schedulers/relink.c: (cb_handoff), (main):
check relinking and adding/removing elements from a running pipeline
* testsuite/schedulers/unlink.c: (cb_handoff), (main):
check unlinking in a running pipeline
* testsuite/schedulers/unref.c: (cb_handoff), (main):
check unreffing a running pipeline
* testsuite/schedulers/useless_iteration.c: (main):
check iterating a pipeline that contains running threads works
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (remove_from_group),
(setup_group_scheduler), (gst_opt_scheduler_pad_link):
Fixed an error introduced with patch for 1.63. When setting
a get based element as the entry point in a group, make sure
to mark the group as GET based.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (create_group),
(setup_group_scheduler), (loop_group_schedule_function),
(gst_opt_scheduler_pad_link):
Added some more debug info and fixed a bug where the group
type was set to LOOP but it was in fact unknown.
Original commit message from CVS:
* gst/gstelement.c: (gst_element_class_init):
whoops, store the signal id correctly
* gst/schedulers/gstbasicscheduler.c:
(gst_basic_scheduler_chain_wrapper):
detect infinfinfinfinfinf^Cinite loops in chain wrapper when the
chain function isn't linked
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c: (remove_from_group),
(group_dec_link), (gst_opt_scheduler_pad_unlink):
when removing an element from a group, we always need to
decrement the link count that this group had with other
groups through the element.
added an extra assert to catch inconsistencies when decrementing
the link count.
Original commit message from CVS:
* gst/schedulers/gstoptimalscheduler.c
(gst_opt_scheduler_pad_unlink): Check if element is non-NULL and
really is a GstElement. Avoids critical when running gst-launch -v
and a oggdemux/decoding pipeline.
Original commit message from CVS:
reviewed by Benjamin Otte <otte@gnome.org>
* gst/schedulers/gthread-cothreads.h:
free allocated data for main cothread, too when destroying context
(fixes#141417)
Original commit message from CVS:
* gst/autoplug/Makefile.am: Fix some little buglets in last checkin.
* gst/indexers/Makefile.am:
* gst/schedulers/Makefile.am:
* libs/gst/bytestream/Makefile.am:
* libs/gst/control/Makefile.am:
* libs/gst/getbits/Makefile.am:
Original commit message from CVS:
* common/as-libtool.mak: Fine-tune DLL building.
* configure.ac: Link plugins against libgstreamer. Define plugindir
(like gst-plugins)
* examples/plugins/Makefile.am: remove plugindir
* gst/autoplug/Makefile.am: DLL building fixes
* gst/elements/Makefile.am: DLL building fixes. Disable pipefilter on
Windows.
* gst/elements/gstelements.c: Conditionally disable pipefilter.
* gst/indexers/Makefile.am: DLL building fixes
* gst/schedulers/Makefile.am: DLL building fixes.
* libs/gst/bytestream/Makefile.am: DLL building fixes.
* libs/gst/control/Makefile.am: same
* libs/gst/getbits/Makefile.am: same
* testsuite/Makefile.am: New dlopen directory
* testsuite/dlopen/Makefile.am: Tests to check if libgstreamer works
when dlopened.
* testsuite/dlopen/dlopen_gst.c: (main): same
* testsuite/dlopen/loadgst.c: (do_test): same
Original commit message from CVS:
* gst/schedulers/entryscheduler.c:
(gst_entry_scheduler_yield):
refuse to yield when decoupled elements insist on doing that.
At least it's better than crashing
Original commit message from CVS:
* gst/gstpad.c: (gst_pad_set_active), (gst_pad_link_new),
(gst_pad_link_free), (gst_pad_link_try), (_invent_event),
(gst_pad_pull):
implement enforcing discont events before buffers are passed. This
allows state changes of only some elements and later correctly going
on where they left off (or in short: you can now set audio sinks to
NULL to release the device when the pipeline is paused)
* gst/gstpad.c: (gst_pad_call_chain_function),
(gst_pad_call_get_function):
* gst/gstpad.h:
add gst_pad_call_chain_function and gst_pad_call_get_function for
scheduler interaction. They are required because of the changes
above.
* gst/schedulers/entryscheduler.c: (get_buffer),
(gst_entry_scheduler_chain_wrapper),
(gst_entry_scheduler_get_wrapper),
(gst_entry_scheduler_state_transition),
(gst_entry_scheduler_pad_link):
* gst/schedulers/gstbasicscheduler.c:
(gst_basic_scheduler_chain_wrapper),
(gst_basic_scheduler_src_wrapper),
(gst_basic_scheduler_chainhandler_proxy),
(gst_basic_scheduler_gethandler_proxy),
(gst_basic_scheduler_cothreaded_chain),
(gst_basic_scheduler_chain_elements):
* gst/schedulers/gstoptimalscheduler.c:
(get_group_schedule_function), (pad_clear_queued),
(gst_opt_scheduler_pad_link):
use the new functions instead of calling get/chain-functions
directly.
Original commit message from CVS:
* configure.ac: Changes to remove POSIXisms (mmap in this case)
and to build DLLs on Windows.
* gst/Makefile.am:
* gst/elements/gstfilesrc.c: (gst_filesrc_get),
(gst_filesrc_open_file):
* gst/schedulers/Makefile.am:
Original commit message from CVS:
* gst/schedulers/gstbasicscheduler.c: (plugin_init):
rename categories to basic_*
* gst/schedulers/gstbasicscheduler.c:
(gst_basic_scheduler_chain_wrapper),
(gst_basic_scheduler_chainhandler_proxy),
(gst_basic_scheduler_gethandler_proxy),
(gst_basic_scheduler_eventhandler_proxy):
debugging category fixes - put common stuff in log category
* gst/schedulers/gstbasicscheduler.c:
(gst_basic_scheduler_chain_elements):
dirty fix: call gst_basic_scheduler_cothreaded_chain when already
active and linking two active chains
Original commit message from CVS:
* gst/schedulers/gstbasicscheduler.c:
make sure we can switch back to the main function if we're still in
the main function (supposed to fix#139617)
* gst/schedulers/gthread-cothreads.h:
don't throw an error when switching to the same cothread
Original commit message from CVS:
* gst/gstbuffer.h:
remove macro that wouldn't work anymore because struct member has
been removed.
* gst/schedulers/entryscheduler.c: (schedule_forward):
fix segfault for unconnected pads
Original commit message from CVS:
* gst/schedulers/entryscheduler.c: (some functions, dunno which,
prepareChangeLog doesn't work when cvs indents):
don't throw an error when no element can be scheduled, there's too
many weird reasons why it doesn't work. Return STOPPED instead.
decoupled elemts' schedulability doesn't depend on bufpens.
