Track both device and active connections at the same time and decide
whether activation finished (either successfully or not) in a single
place, check_activated().
This makes the already too complex code path a bit more straightforward
and also makes it possible to be more reasonable about the diagnostics.
Now that the active connection signals the reason, we include it; but if
the failure is due to the device disconnection while we're activating,
include a device reason instead, since it's often more useful. Like:
Before:
Error: Connection activation failed.
Without considering the device:
Error: Connection activation failed: the base network connection was interrupted
After:
Error: Connection activation failed: The Wi-Fi network could not be found
In practice, this should only matter when there are multiple
header files with the same name. That is something we try
to avoid already, by giving headers a distinct name.
When building NetworkManager itself, we clearly want to use
double-quotes for including our own headers.
But we also want to do that in our public headers. For example:
./a.c
#include <stdio.h>
#include <nm-1.h>
void main() {
printf ("INCLUDED %s/nm-2.h\n", SYMB);
}
./1/nm-1.h
#include <nm-2.h>
./1/nm-2.h
#define SYMB "1"
./2/nm-2.h
#define SYMB "2"
$ cc -I./2 -I./1 ./a.c
$ ./a.out
INCLUDED 2/nm-2.h
Exceptions to this are
- headers in "shared/nm-utils" that include <NetworkManager.h>. These
headers are copied into projects and hence used like headers owned by
those projects.
- examples/C
This is sort of ugly, because it includes the domain and log levels
verbatim. They're just plain strings on the API, there's no way the
client would know which ones are valid.
On the other hand this kills one of two uses of nmc_parse_args(), which
probably means it's not a very good abstraction and maybe we should get
rid of it altogether. It is in particular unfriendly to argument
completion.
We can't pass the password obtained from
nm_setting_802_1x_get_*private_key_password() to
nm_setting_802_1x_set_*private_key() as the latter also frees the old
password.
Fixes: afd2811028
If the connection spontaneously disappears (perhaps along with the whole
daemon on crash) while we're deleting it, then the removal callback
would free up the context structure the delete operation is using.
Let's cancel the in-flight delete operations so that they won't touch
the structure after it's gone.
* all getter/setter/describe, ... functions grouped according to properties
* sort the settings groups alphabetically
(cherry picked from commit 418733f2c1)
It results in a rather confusing behavior:
# nmcli c modify wifi \
802-1x.private-key /etc/pki/themostsecret.key \
802-1x.private-key-password verystrong
Error: failed to modify 802-1x.private-key: private key
password not provided.
Moreover, the user might have a good reason for not specifying it on a command
liue and it's not strictly required anyway -- we'll do fine if we don't verify
a private key at addition time.
It helps listing or searching for slave connections. For example, one can do
$ nmcli -f name,uuid,type,slave connection show | awk '$NF == "bridge"'
to show all bridge slave connections.
We now require settings-docs.c to be present. Either, configure
with --enable-gtk-doc to have it build, or use the pre-generated file
from the source tarball.
Also, be a bit more careful about the layers of errors. Just don't do this:
(process:236): nmcli-CRITICAL **: Error: Could not create NMClient object:
Permissions request failed: Authorization check failed:
The name org.freedesktop.PolicyKit1 was not provided by any .service files.
