| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Holding handles is not needed with Connection Managers having
immortal handles (any Connection Manager using telepathy-glib >= 0.13.8).
Other Connection Managers are considered deprecated, clients wanting to
still support them should continue using this deprecated function.
|
|
|
|
| |
Everywhere we need an handle, we can give an ID nowadays.
|
|
|
|
|
| |
Contact attributes is an internal concept that should not appear
in our API.
|
|
|
|
|
|
| |
tp_connection_unref_handles()
They are no-op now, so we can safely remove them all
|
|
|
|
|
|
| |
Based on a patch from Jonny Lamb, updated for current master.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=46470
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Replace g_(ptr_)array_free (foo, TRUE) and g_hash_table_destroy
with respectively g_(ptr_)array_unref (foo) and g_hash_table_unref.
I used this command to generate this patch:
for f in `find -name "*.c"`; do sed -i $f -re 's/g_ptr_array_free \(([^ ,]+), TRUE\)/g_ptr_array_unref \(\1\)/'; done
See Danielle's blog for explanation of possible bug _free can do:
http://blogs.gnome.org/danni/2011/11/16/mistakes-with-g_value_set_boxed/
|
| |
|
|
|
|
| |
There's no need, now that we never release them.
|
|
|
|
|
| |
This gets us a little closer to the "immortal handles" behaviour even for
CMs that have refcounting.
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
docs/reference/telepathy-glib-sections.txt
examples/cm/contactlist/contact-list-manager.c
spec/Connection_Interface_Contact_Groups.xml
spec/Connection_Interface_Contact_List.xml
spec/errors.xml
telepathy-glib/errors.c
|
| |\ |
|
| | | |
|
| | |
| | |
| | |
| | |
| | | |
This renames GetContactListAttributes to RequestContactList; all the other
spec changes were already implemented in this branch.
|
| | |
| | |
| | |
| | | |
As per Guillaume's review.
|
| | |
| | |
| | |
| | |
| | |
| | | |
This needs a specific binding for the same reasons as
GetContactAttributes. I've called it _request_ rather than _get_ since
I plan to rename the underlying D-Bus method at some point.
|
| |/
|/|
| |
| | |
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|/ |
|
| |
|
|
|
|
|
|
| |
Skipped tp_g_value_slice_* functions in dbus.c
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|\
| |
| |
| | |
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
Previously, we mistakenly passed error (which is NULL on this code path)
to client code; this should have been e, the new GError we just allocated.
Also, don't leak that GError.
|
|/
|
|
| |
TpConnectionManager
|
| |
|
|
|
|
|
|
|
| |
occured
If something went wrong attributes is NULL so we can't call g_hash_table_size
on it (fdo #27537).
|
|
|
|
|
|
| |
Fixes fd.o #19741
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|
|
|
|
|
|
| |
Calling _tp_connection_ref_handles on an invalidated connection (which
self will be here if error is non-NULL) asserts.
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
| |
|
| |
|
|
|
|
|
| |
This is a wrapper for GetContactAttributes that correctly updates the
handle reference tracking infrastructure if @hold is TRUE.
|
| |
|
|
|
|
|
|
| |
obscurely indicates user error, indicate what they did wrong
From the "abuse of the comma operator" department.
|
|
|
|
| |
DEBUG()
|
|
|
|
|
|
|
|
| |
of data structures
_tp_connection_init_handle_refs now preallocates some of the data
structures, so we can just assert that they're good rather than allocating
them lazily.
|
|
|
|
| |
handles
|
| |
|
|
|
|
| |
on invalidated connections
|
|
|
|
|
| |
Passing an out-of-range handle type is library user error, so we should
g_return_if_fail on it.
|
|
|
|
|
|
|
| |
bucket have been invalidated
TpConnection emits invalidated when it runs out of refs, so this is
necessary for correct sharing of buckets.
|
|
|
|
|
|
|
| |
invalidated
This avoids repeated alloc/free cycles if a connection oscillates
between holding 0 and 1 handles.
|
| |
|
|
|
|
|
| |
It was previously unclear whether it was the same pointer, or a pointer
to a copy.
|
|
This centralizes handle ref tracking, because otherwise calling
ReleaseHandles is inherently unsafe - you never know whether
another part of the process is still using the handles.
In this implementation the handles are stored in a data-slot on the
DBusConnection.
See mailing list thread "Uniqueness of objects", starting
with Message-ID: <20080930200125.GA15304@carbon.pseudorandom.co.uk>
|