When writing a cache chunk fails, we were freeing the node and jump to a final
cleanup which dereferenced a null pointer. Leve freeing the node to the cleanup
code in fail_free_list. (sorry for committing wrong fix before).
When writing a cache chunk fails, we were freeing the node and jump to a final
cleanup which dereferenced a null pointer. Leve freeing the node to the cleanup
code in fail_free_list.
The previous commit was bogus, as was the check before. We just point m to the file data,
so neither it nor its members will be NULL. Better check if we have enough data.
Allocate every structure that is directly written to the binary
registry with g_malloc0(). Otherwise some parts of it will be
uninitialized (struct padding because of alignment, etc) and
valgrind will complain about it.
Original commit message from CVS:
* gst/gstregistrybinary.c: (unpack_element), (unpack_const_string),
(unpack_string)::
Wrap multi-line macros in G_STMT_{START|END}.
Original commit message from CVS:
* gst/gstregistrybinary.c:
In win32 codepath, if we fail to write the registry, create the
directory for it and try again, matching the behaviour in non-win32
codepaths.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_write_chunk),
(gst_registry_binary_initialize_magic),
(gst_registry_binary_write_cache),
(gst_registry_binary_check_magic):
* gst/gstregistrybinary.h:
Don't write and check a CRC for the binary registry file. It's
guaranteed that the registry is completely written (it's first written
to a temporary file and then moved) and if the registry was corrupted
by some hardware failure we would have bigger problems.
Bump binary registry version to 0.10.21.1 for this as it's an
incompatible change and to ensure that the registry gets rebuild
after the update.
This saves some milliseconds for reading/writing the registry.
Fixes bug #560399.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_write_cache):
If we can't get a cache file don't try to save something to it.
Dereferencing NULL pointers usually isn't a good idea.
Original commit message from CVS:
* gst/gstregistrybinary.c:
Don't use g_mkstmp() on win32, it's unsafe if glib is using a different
libc.
Fixes#544776.
Original commit message from CVS:
* configure.ac:
* docs/gst/gstreamer-sections.txt:
* docs/gst/gstreamer.types:
* docs/gst/gstreamer.types.in:
* gst/Makefile.am:
* gst/gst.c:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_set_index_func):
* gst/gstconfig.h.in:
* gst/gstelement.c: (gst_element_get_index):
* gst/gstregistrybinary.c: (gst_registry_binary_save_feature),
(gst_registry_binary_load_feature),
(gst_registry_binary_read_cache):
* gst/gstregistryxml.c: (load_feature),
(gst_registry_xml_read_cache), (gst_registry_xml_save_feature):
* plugins/Makefile.am:
* tools/gst-indent:
* tools/gst-inspect.c: (print_index_info), (print_element_list),
(print_plugin_features), (print_element_features):
* tools/gst-xmlinspect.c: (print_event_masks),
(print_element_info):
* win32/common/gstconfig.h:
Remove GST_DISABLE_(ENUMTYPES|INDEX|URI) everywhere.
Disabling the indexers and URI handler code will only reduce the
required amount of memory by a very small amount but on the other hand
requires much more maintaince work. Apart from that many places of
code are broken when disabling them.
Disabling the enum types doesn't reduce the required amount of memory
by more than a few bytes and makes it hard to fix bugs like #539772,
i.e. use the enums as GObject properties.
Original commit message from CVS:
* gst/gstplugin.c: (CHECK_PLUGIN_DESC_FIELD), (gst_plugin_load_file):
Print error debug message if plugin description fields that should
be set are NULL.
* gst/gstregistrybinary.c: (gst_registry_binary_save_const_string):
Don't crash if the string to serialise is NULL (it really should
not be, but apparently this used to work with the xml registry ...).
Original commit message from CVS:
* gst/gstclock.c: (gst_clock_set_calibration):
Make some checks actually useful.
* gst/gstregistrybinary.c: (gst_registry_binary_load_plugin):
Remove some unused code. Unsigned integers tend to be >= 0.
Original commit message from CVS:
* gst/gstregistrybinary.c:
Add a warning of we hit unhandled factories when saving.
More debug logging detail, but move to LOG category.
Original commit message from CVS:
* gst/gstregistrybinary.c: (_gst_crc32):
Return the old CRC instead of 0 if we give a NULL buffer
or a buffer with a length of 0.
Original commit message from CVS:
* gst/gstregistrybinary.c: (_gst_crc32),
(gst_registry_binary_write),
(gst_registry_binary_initialize_magic),
(gst_registry_binary_write_cache),
(gst_registry_binary_check_magic),
(gst_registry_binary_read_cache):
* gst/gstregistrybinary.h:
Add crc32 checksum to the binary registry file and check this before
accepting a registry file.
Also free the data list when writing to the registry file fails.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_save_feature),
(gst_registry_binary_load_feature),
(gst_registry_binary_load_plugin):
If an element supports the Uri interface, returns a valid pointer
to the supported URI protocols but this pointer contains nothing
don't try to save that as it will corrupt the registry.
Don't unref the plugin if we added it to the registry already but
fail to load a feature as gst_registry_add_plugin() takes ownership
of the plugin.
Improve debugging a bit.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_load_feature),
(gst_registry_binary_load_plugin):
Fix a typo in a debug message and revert change from yesterday as
gst_registry_add_plugin() will only fail if something is really wrong
already and we can't survive it anyway.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_load_plugin):
If gst_registry_add_plugin() fails our reference to the plugin is
invalid so don't try to use it anymore and instead error out.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_write):
Align memory to the pointer size even if the architecture allows
unaligned memory access. Unaligned memory access usually comes with
performance penality.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_write),
(gst_registry_binary_check_magic),
(gst_registry_binary_load_pad_template),
(gst_registry_binary_load_feature),
(gst_registry_binary_load_plugin):
Align memory to the pointer size instead of always 32 bit. Fixes
unaligned memory accesses on ia64 and friends.
* gst/gstregistrybinary.h:
Bump binary registry format version for this as it changes the
format on those architectures that don't have unaligned access
and 64 bit pointers.
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* gst/gstregistrybinary.c: (gst_registry_binary_initialize_magic),
(gst_registry_binary_check_magic):
* gst/gstregistrybinary.h:
Call the version GST_MAGIC_BINARY_VERSION_STR to be more consistent
and add it to the (private part) of the docs to fix the build.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_initialize_magic),
(gst_registry_binary_check_magic),
(gst_registry_binary_read_cache):
* gst/gstregistrybinary.h:
Don't use GST_MAJORMINOR for the binary registry version. Instead
hardcode a value that must be changed whenever the format changes
in an incompatible way.
Also don't GST_ERROR when there is a version mismatch, just
regenerate the registry silently.
Original commit message from CVS:
Patch by: Ole André Vadla Ravnås
<ole dot andre dot ravnas at tandberg dot com>
* gst/gstregistrybinary.c:
Include io.h for write() and close() when building with MSVC. Fixes
bug #520877.
Original commit message from CVS:
* configure.ac:
* gst/gst_private.h:
* gst/gstconfig.h.in:
* gst/gstregistry.h:
* gst/gstregistrybinary.c:
* win32/common/gstconfig.h:
Move registry backend API to private headers where we can. Add
fixme-0.11 comments for the others. Add stubs for the xml backend when
using the binary to ensure they functions exists (they should not be
used though). Fixes#520756.
Original commit message from CVS:
* gst/gstregistrybinary.c: (gst_registry_binary_write_cache):
Use g_remove() and g_rename(). Check result of g_rename(), and
don't leak the open file descriptor if we error out when writing.
* gst/gstregistryxml.c: (load_plugin), (gst_registry_xml_write_cache):
Must check the return value of close() after writing out the new
registry file. Sometimes write problems such as out-of-diskspace
are only reported when the file is closed and not already during
the write. This may have caused partial/broken registry files in
some rare circumstances. Should fix#503675.
Original commit message from CVS:
* gst/gstplugin.c:
* gst/gstplugin.h:
* gst/gstregistrybinary.c:
* gst/gstregistryxml.c:
Put more strings into the GLib quark table. No need to keep
a hundred-something copies of identical version strings,
license strings, package name strings and package origin
strings around.
Original commit message from CVS:
* gst/glib-compat-private.h:
Add compatibility macro for g_intern_string() for
GLib-2.8 (any reason we can't just bump the
requirement to at least 2.10?)
* gst/gstpadtemplate.h:
* gst/gstelementfactory.c:
* gst/gstregistryxml.c:
* gst/gstregistrybinary.c:
Make GstStaticPadTemplate's templ_name field a const gchar * and fix
up the internal code accordingly. This shouldn't be a problem, since
there is no reason external code could ever assume the string in such
a structure is dynamically allocated unless it did that itself; the
use of g_strdup() is private to element factories. The new code also
saves some memory by putting pad template name strings into the GLib
quark table instead of allocating them dynamically.
Declaring this field constant fixes warnings with g++-4.2 when using
the GST_STATIC_PAD_TEMPLATE macro in c++ code (#478092).
