| Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
I always get confused as to how to run the tests as I look for a
tests/README and it's not there.
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
So so annoying.
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
When this test creates a new account, it expects four events to occur:
• A.AccountPropertyChanged, to say the new account is valid;
• AM.AccountValidityChanged, to say the new account is valid;
• CreateAccount returning successfully;
• RequestConnection on the fake CM, because ConnectAutomatically was
set to True and RequestedPresence was set to non-offline in
CreateAccount.
However, if the account's not yet been announced by
AccountValidityChanged, MC has no reason to emit AccountPropertyChanged
on it. It happens that, in some cases, it does; but this depends on
whether or not MC has already introspected the fake CM by the time the
account is created. So the test is wrong to expect
AccountPropertyChanged. Instead, we check that Account.Valid is in sync
with what AccountValidityChanged said.
This fixes an issue where this test would only pass if it was run after
account-manager/create-twice, but not if it was run individually.
|
|
This test runs right before a test which passed or failed depending
on whether this test was run. The logs were made much harder to read by
them both creating accounts with very similar names.
|
|
|
|
|
|
|
|
|
|
|
|
Fixes: <https://bugs.freedesktop.org/show_bug.cgi?id=29812>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Fixes: <https://bugs.freedesktop.org/show_bug.cgi?id=32414>
|
|
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>
|
|
returned
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
Signed-off-by: Jonny Lamb <jonny.lamb@collabora.co.uk>
|
|
|
|
Now that preparing the connection takes one more step, it can take
longer before ObserveChannels is called with the channel that just
appeared on connecting, so let's allow that to appear earlier.
This only relaxes the test assumptions a little. Although it currently
fails on 0.13.9 and this fixes it, it'll still work on tp-glib <
0.13.9.
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>
|
|
I feel kind of bad doing this, but it turned out not to be needed. I
think it's better to take dead code out, rather than leaving it kicking
around to potentially complicate matters later.
|
|
The only property that was implemented in an async fashion was
Account.Parameters. Since that's now synchronous, we can bin this code
for getting properties asynchronously.
|
|
This is a really massive patch, and I'm very sorry.
Previously _mcd_account_dup_parameters() took a callback parameter. This
was because mcd_account_get_parameter() took a callback parameter. The
latter used to actually be asychronous in the gnome-keyring case, but it
is not any more, thanks to the harmonious account backend API.
So we can remove a *lot* of unnecessary complexity if we make the latter
return the parameter's value synchronously, and hence make the former
return a GHashTable synchronously. In particular, this means we can make
retrieving the Parameters D-Bus property synchronous, which means we can
port to TpDBusPropertiesMixin.
I did try to split this patch up, but as you, dear reader, will be able
to see from what I ended up with, this code was all really entangled
with itself. This is the best I could do.
|
|
No known subclass of McdAccount overrides get_parameter, set_parameter,
delete or load these days. People should be, and are, using the new
pluggable storage backends API. So it's safe to remove these methods
from the vtable. This will simplify refactoring their (default)
implementations.
|
|
dbus-glib treats a GValue containing a NULL string as equivalent to a
GValue containing "", at least in this context. Many other properties
rely on this behaviour; so let's just rely on it here too.
|
|
The property's actually called Condition, not Conditions. But because
McdDBusProp returns the same error for “unknown property” and “wrong
type”, the test passed.
|
|
These have been in tp-glib approximately forever.
|
|
|
|
|
|
|
|
|
|
git+ssh://git.collabora.co.uk/git/telepathy-mission-control
|
|
|
|
|
|
telepathy-mission-control-5.6
Fixes: <https://bugs.freedesktop.org/show_bug.cgi?id=24779>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
|
|
|
|
git+ssh://git.collabora.co.uk/git/telepathy-mission-control
|
|
|
|
Also only ship the changelog since 5.4, which should be plenty.
Reviewed-by: Will Thompson <will.thompson@collabora.co.uk>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|