| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When passwords are stored in the Parameters by MC and no
SASLAuthentication handler is used, e.g. under Sugar, we currently
save passwords into gnome-keyring, if enabled at compile time; so, we
want to read them back out of gnome-keyring.
This conflicts slightly with wanting to ignore passwords in gnome-keyring
that are left over from an Empathy 3.0 schema migration that failed
due to fd.o #42088. To avoid failure to authenticate/re-prompt if the
password has been changed on the service side, ignore passwords that
are going to be deleted; but to avoid breaking Sugar, don't ignore
any other passwords we might find in gnome-keyring.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=59468
|
| |
|
| |
|
|
|
|
|
| |
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Guillaume Desmottes <guillaume.desmottes@collabora.co.uk>
|
|
|
|
|
|
|
| |
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Bug-Debian: http://bugs.debian.org/687933
Reviewed-by: Guillaume Desmottes <guillaume.desmottes@collabora.co.uk>
|
|
|
|
|
|
|
| |
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Cherry-picked-from: c21e2b09bf4cdda6fb05d64f9198f6db7b776c76
|
|
|
|
|
|
|
| |
"make distcheck" in an already-out-of-tree build turns out not to
work because of this.
This corresponds to commit beb2d971 in telepathy-glib.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Otherwise the upgrade from Debian 6 to 7 forgets MSN passwords, which
is not the end of the world, but seems less nice than it could be.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Xavier Claessens <xavier.claessens@collabora.co.uk>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Mission Control has a bug/design flaw/whatever in which it doesn't
know the types of parameters without looking at the corresponding
connection manager, which doesn't work if that connection manager has
been uninstalled. This turns out to be bad for the upgrade from
Debian 6 to Debian 7, which upgrades MC to a version that will
perform the migration from Butterfly to Haze, but also uninstalls
Butterfly...
Since we know what the type of 'account' is, with a bit of refactoring
we can tell McdAccount that we know it's a string, get its value
anyway, and get on with our lives.
Bug-Debian: http://bugs.debian.org/686835
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Xavier Claessens <xavier.claessens@collabora.co.uk>
|
|
|
|
|
|
|
|
|
|
|
|
| |
will look for
Deleting secrets with param="param-password" isn't a whole lot of use
when we save, and look up, param="password".
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Cherry-picked-from: 1d9e8f5cfb9e7b426a99ae6e16c35c1101d55a91
|
|
|
|
|
|
|
|
|
|
| |
Otherwise we'd just delete it, then (because it's still in secrets)
re-commit it!
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Cherry-picked-from: 16e4507c5001f6f45158ef3d0b46998fa8c1ca2a
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'removed' is essentially a set of (account, key) tuples that should
be deleted. What we were doing was:
foreach account in removed
foreach key in secrets[account]
delete (account, key)
which makes little sense - if we have param-password and
param-proxy-password and we want to unset Parameters['password'],
the current implementation would delete both. This commit changes it to:
foreach account in removed
foreach key in removed[account]
delete (account, key)
which has the advantage of actually making sense.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Cherry-picked-from: d4ca35cffea9d0093e127e0be633501d22ded35f
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Our tracking of whether something is "secret" (in the gnome-keyring) is
pretty shaky, and in particular, we can forget that things are meant
to be "secret" sometimes (we lose that information when deleting
parameters).
Happily, when we're deleting things, it doesn't actually matter: the
right thing to do is clearly to delete from both locations, regardless
of where we think it ought to be.
Similarly, when we're setting a property to a new value, it's appropriate
to delete it from both locations, then put it in the right location
(only).
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42088
Cherry-picked-from: 5b4954d2c959c4ba7f6c67b7d8e16eaa5ab272dd
|
|
|
|
|
|
|
|
| |
To be considered as online, we have to be connected to the network *and* not
in the proccess of sleeping.
Fix a race where an account we just disconnected because we are going to sleep
is put back online because the monitor was considering us online again.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This wasn't really an appropriate change for a stable branch: it
introduced code churn without actually fixing a bug. We should
ignore deprecation warnings instead.
This reverts commit 88642511bb9c6f66d2fe4aae1550893eb0d781b2.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Acked-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|
|
|
|
|
|
|
| |
There's no immediately obvious way to install GSettings schemas there.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=53497
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If you do:
dbus-send --dest=org.freedesktop.Telepathy.MissionControl5 \
--type=method_call --print-reply / org.freedesktop.DBus.Peer.Ping & \
dbus-send --dest=org.freedesktop.Telepathy.AccountManager \
--type=method_call --print-reply / org.freedesktop.DBus.Peer.Ping
then one or both of the calls will fail because the executable claims
one of its bus names, but can't get the other, and exits 1.
By making one wait for the other, we get the desired behaviour.
I've made activating MissionControl5 wait for AccountManager because
in MC's current implementation AccountManager is taken last.
Reviewed-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=53220
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=52231
|
| |
|
| |
|
|
|
|
| |
It make sure "make -j3 maintainer-foo" works
|
|
|
|
|
|
|
|
| |
If the account is already in the hashtable, g_hash_table_insert()
will set @removed as value, but free it since the key already in
the table is kept.
https://bugs.freedesktop.org/show_bug.cgi?id=51842
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
_mcd_client_proxy_handle_channels sets the McdChannel state to
HANDLER_INVOKED, but when we were calling this in both delegate and
present operations, the state was not being set back to the DISPATCHED
state afterwards.
The state being wrong meant that dispatching a channel after if had
been delegated or presented would not happen, as MC thought we were
still waiting for a handler response.
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
| |
|
|
|
|
|
|
|
| |
tp_*_call_when_ready were replaced by tp_proxy_prepare_async a while
ago.
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
|
|
|
| |
This is needed for newer libnm-glib. We should start using
nm_client_new_async().
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=47054
|
|
|
|
|
|
| |
GLib 2.30 is necessary for g_unix_open_pipe().
https://bugs.freedesktop.org/show_bug.cgi?id=47054
|
|
|
|
|
|
|
| |
This makes it output the complete dependency chain for each
library, e.g. libffi for GObject. This is necessary because Android
toolchains don't follow shared library interdependencies like
normal GNU toolchains.
|
| |
|
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the connection manager is not installed, or is installed but doesn't
have the protocol for the account, then _mcd_account_dup_parameters()
returns %NULL. (As the comment I've added says, I think this is
ridiculous—it's fall-out from MC not storing the types of parameters.)
As a result, Get('Parameters') or GetAll() would crash, because
dbus-glib would try to serialize a NULL GHashTable and blow up. (See
<https://bugs.freedesktop.org/show_bug.cgi?id=44939>.) So with this
patch, we handle NULL as "no parameters" for the sake of D-Bus
properties. (Because this situation should only arise if the account is
invalid, we also warn if _dup_parameters() returns NULL for valid
accounts.)
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
| | |
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | | |
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42814
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
| | |
| | |
| | |
| | |
| | | |
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=42814
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|