First, it was not easily possible to set a private key without
also providing a password. This used to be OK, but now with
secret flags it may be the case that when the connection is read,
there's no private key password. So functions that set the
private key must account for NULL passwords.
Unfortunately, the crytpo code did not handle this case well.
We need to be able to independently (a) verify that a file looks
like a certificate or private key and (b) that a given password
decrypts a private key. Previously the crypto code would fail
to verify the file when the password was NULL.
So this change fixes up the crytpo code for a more distinct
split between these two operations, such that if no password is
given, the file is still checked to ensure that it's a private
key or a certificate. If a password is given, the password is
checked against the private key file.
This commit also changes how private keys and certificates were
handled with the BLOB scheme. Previously only the first certificate
or first private key was included in the property data, while now
the entire file is encoded in the data. This is intended to fix
cases where multiple private keys or certificates are present in
a PEM file. It also allows clients to push certificate data to
NetworkManager for storage in system settings locations, which was
not as flexible before when only part of the certificate or key
was sent as the data.
Clients need to do their own logging using glib or whatever; these
macros while somewhat helpful were not flexible and are not a
substitute for actual logging in the client. g_warning, g_message,
and g_error are more suitable anyway.
It's always used with a GByteArray anyway, as are most
functions in nm-utils.h. Even better, we can skip the
memcpy since it turns out to be pointless.
Because most of the time they will. They need special handling all
around anyway because only the VPN plugin itself knows whether the
connection needs secrets.
The old function took a string value, which wasn't really correct as
the property type is a GHashTable of string:string. For whatever
reason this is how nm-applet passed VPN secrets back to NM in the return
from the GetSecrets() D-Bus call. This was probably easier or
something but it was a special case that's magic and quite unclear.
Since we use nm_connection_update_secrets() more these days, and we
depend on the GValue types we pass into it matching the property
types of the setting property the secret is for, we need to fix that
up for VPN connections. But keep the old code for backwards
compatibility.
In the future secret agents should pass back VPN secrets in the same
form as the VPN setting specifies them for the "secrets" property:
a GHashTable of string:string. But the old mechanism of just dumping
the key/value pairs into the returned VPN hash as string:string will
still work.
Not all connections will require every secret, and sometimes we
can't automatically figure out whether we need the secret. For
vpnc sometimes the group password isn't used, and sometimes PPP
providers require a username but don't want a password, etc.
Becuase there's only one 'flags' property for WEP keys (because it's pretty
dumb to have different flags for all 4 WEP keys) we need to do some tap dancing
with the secret name, so that requests for "wep-keyX" look up the "wep-key-flags"
property.
This allows the necessary flexibility when handling secrets; otherwise
it wouldn't be known when NM should save secrets returned from agents
to backing storage, or when the agents should store the secrets. We
can't simply use lack of a secret in persistent storage as the indicator
of this, as (for example) when creating a new connection without
secrets the storage method would be abmiguous.
At the same time, fold in "always ask" functionality for OTP tokens
so user agents don't have to store that attribute themselves out-of-band.
There was never a property for it anyway, so it never got serialized
across D-Bus, because it was folded into the "password" property in
wpa_supplicant between 0.5 and 0.6.
The first-level hash table key should be the setting name itself,
not the GType name of the setting's GObject. There's probably a
better way to do this to reduce that confusion.
Simplifies code internally, and makes it easier for clients as well in
some cases where they want to control what ends up in the resulting
hash and what does not.
Add the necessary annotations (the mininum required, that is those
on return values. NULL parameters or container types may require
more), and the Autotools stuff to get a NetworkManager GIR for
libnm-util and a NMClient for libnm-glib.
If a pointer to a valid GError was not passed the function could
crash. Make it simpler and fix the possible crash by just converting
to hash table iters instead.
