ApplicationCommandLine

class ApplicationCommandLine(**properties: Any)

Superclasses: Object

GApplicationCommandLine represents a command-line invocation of an application.

It is created by Application and emitted in the command_line signal and virtual function.

The class contains the list of arguments that the program was invoked with. It is also possible to query if the commandline invocation was local (ie: the current process is running in direct response to the invocation) or remote (ie: some other process forwarded the commandline to this process).

The GApplicationCommandLine object can provide the argc and argv parameters for use with the OptionContext command-line parsing API, with the get_arguments function. See [gapplication-example-cmdline3.c][gapplication-example-cmdline3] for an example.

The exit status of the originally-invoked process may be set and messages can be printed to stdout or stderr of that process.

For remote invocation, the originally-invoked process exits when done method is called. This method is also automatically called when the object is disposed.

The main use for GApplicationCommandLine (and the command_line signal) is ‘Emacs server’ like use cases: You can set the EDITOR environment variable to have e.g. git use your favourite editor to edit commit messages, and if you already have an instance of the editor running, the editing will happen in the running instance, instead of opening a new one. An important aspect of this use case is that the process that gets started by git does not return until the editing is done.

Normally, the commandline is completely handled in the command_line handler. The launching instance exits once the signal handler in the primary instance has returned, and the return value of the signal handler becomes the exit status of the launching instance.

static int
command_line (GApplication            *application,
              GApplicationCommandLine *cmdline)
{
  gchar **argv;
  gint argc;
  gint i;

  argv = g_application_command_line_get_arguments (cmdline, &argc);

  g_application_command_line_print (cmdline,
                                    "This text is written back\n"
                                    "to stdout of the caller\n");

  for (i = 0; i < argc; i++)
    g_print ("argument ``%d``: ``%s``\n", i, argv[i]);

  g_strfreev (argv);

  return 0;
}

The complete example can be found here: gapplication-example-cmdline.c

In more complicated cases, the handling of the commandline can be split between the launcher and the primary instance.

static gboolean
 test_local_cmdline (GApplication   *application,
                     gchar        ***arguments,
                     gint           *exit_status)
{
  gint i, j;
  gchar **argv;

  argv = *arguments;

  if (argv[0] == NULL)
    {
      *exit_status = 0;
      return FALSE;
    }

  i = 1;
  while (argv[i])
    {
      if (g_str_has_prefix (argv[i], "--local-"))
        {
          g_print ("handling argument ``%s`` locally\n", argv[i]);
          g_free (argv[i]);
          for (j = i; argv[j]; j++)
            argv[j] = argv[j + 1];
        }
      else
        {
          g_print ("not handling argument ``%s`` locally\n", argv[i]);
          i++;
        }
    }

  *exit_status = 0;

  return FALSE;
}

static void
test_application_class_init (TestApplicationClass *class)
{
  G_APPLICATION_CLASS (class)->local_command_line = test_local_cmdline;

  ...
}

In this example of split commandline handling, options that start with --local- are handled locally, all other options are passed to the command_line handler which runs in the primary instance.

The complete example can be found here: gapplication-example-cmdline2.c

If handling the commandline requires a lot of work, it may be better to defer it.

static gboolean
my_cmdline_handler (gpointer data)
{
  GApplicationCommandLine *cmdline = data;

  // do the heavy lifting in an idle

  g_application_command_line_set_exit_status (cmdline, 0);
  g_object_unref (cmdline); // this releases the application

  return G_SOURCE_REMOVE;
}

static int
command_line (GApplication            *application,
              GApplicationCommandLine *cmdline)
{
  // keep the application running until we are done with this commandline
  g_application_hold (application);

  g_object_set_data_full (G_OBJECT (cmdline),
                          "application", application,
                          (GDestroyNotify)g_application_release);

  g_object_ref (cmdline);
  g_idle_add (my_cmdline_handler, cmdline);

  return 0;
}

In this example the commandline is not completely handled before the command_line handler returns. Instead, we keep a reference to the GApplicationCommandLine object and handle it later (in this example, in an idle). Note that it is necessary to hold the application until you are done with the commandline.

The complete example can be found here: gapplication-example-cmdline3.c

Methods

class ApplicationCommandLine

create_file_for_arg(arg: str) → File

Creates a File corresponding to a filename that was given as part of the invocation of cmdline.

This differs from new_for_commandline_arg() in that it resolves relative pathnames using the current working directory of the invoking process rather than the local process.

Added in version 2.36.

Parameters:: arg – an argument from cmdline

done() → None

Signals that command line processing is completed.

For remote invocation, it causes the invoking process to terminate.

For local invocation, it does nothing.

This method should be called in the command_line handler, after the exit status is set and all messages are printed.

After this call, set_exit_status() has no effect. Subsequent calls to this method are no-ops.

This method is automatically called when the ApplicationCommandLine object is disposed — so you can omit the call in non-garbage collected languages.

Added in version 2.80.

get_arguments() → list[str]

Gets the list of arguments that was passed on the command line.

The strings in the array may contain non-UTF-8 data on UNIX (such as filenames or arguments given in the system locale) but are always in UTF-8 on Windows.

If you wish to use the return value with OptionContext, you must use parse_strv().

The return value is None-terminated and should be freed using strfreev().

Added in version 2.28.

get_cwd() → str | None

Gets the working directory of the command line invocation. The string may contain non-utf8 data.

It is possible that the remote application did not send a working directory, so this may be None.

The return value should not be modified or freed and is valid for as long as cmdline exists.

Added in version 2.28.

get_environ() → list[str]

Gets the contents of the ‘environ’ variable of the command line invocation, as would be returned by get_environ(), ie as a None-terminated list of strings in the form ‘NAME=VALUE’. The strings may contain non-utf8 data.

The remote application usually does not send an environment. Use SEND_ENVIRONMENT to affect that. Even with this flag set it is possible that the environment is still not available (due to invocation messages from other applications).

The return value should not be modified or freed and is valid for as long as cmdline exists.

See getenv() if you are only interested in the value of a single environment variable.

Added in version 2.28.

get_exit_status() → int: Gets the exit status of cmdline. See set_exit_status() for more information.

Added in version 2.28.

get_is_remote() → bool: Determines if cmdline represents a remote invocation.

Added in version 2.28.

get_options_dict() → VariantDict

Gets the options that were passed to g_application_command_line().

If you did not override local_command_line() then these are the same options that were parsed according to the OptionEntry added to the application with add_main_option_entries() and possibly modified from your GApplication::handle-local-options handler.

If no options were sent then an empty dictionary is returned so that you don’t need to check for None.

The data has been passed via an untrusted external process, so the types of all values must be checked before being used.

Added in version 2.40.

get_platform_data() → Variant | None

Gets the platform data associated with the invocation of cmdline.

This is a Variant dictionary containing information about the context in which the invocation occurred. It typically contains information like the current working directory and the startup notification ID.

It comes from an untrusted external process and hence the types of all values must be validated before being used.

For local invocation, it will be None.

Added in version 2.28.

get_stdin() → InputStream | None

Gets the stdin of the invoking process.

The InputStream can be used to read data passed to the standard input of the invoking process. This doesn’t work on all platforms. Presently, it is only available on UNIX when using a D-Bus daemon capable of passing file descriptors. If stdin is not available then None will be returned. In the future, support may be expanded to other platforms.

You must only call this function once per commandline invocation.

Added in version 2.34.

getenv(name: str) → str | None

Gets the value of a particular environment variable of the command line invocation, as would be returned by getenv(). The strings may contain non-utf8 data.

The return value should not be modified or freed and is valid for as long as cmdline exists.

Added in version 2.28.

Parameters:: name – the environment variable to get

print_literal(message: str) → None

Prints a message using the stdout print handler in the invoking process.

Unlike print(), message is not a printf()-style format string. Use this function if message contains text you don’t have control over, that could include printf() escape sequences.

Added in version 2.80.

Parameters:: message – the message

printerr_literal(message: str) → None

Prints a message using the stderr print handler in the invoking process.

Unlike printerr(), message is not a printf()-style format string. Use this function if message contains text you don’t have control over, that could include printf() escape sequences.

Added in version 2.80.

Parameters:: message – the message

set_exit_status(exit_status: int) → None

Sets the exit status that will be used when the invoking process exits.

The return value of the Application::command-line signal is passed to this function when the handler returns. This is the usual way of setting the exit status.

In the event that you want the remote invocation to continue running and want to decide on the exit status in the future, you can use this call. For the case of a remote invocation, the remote process will typically exit when the last reference is dropped on cmdline. The exit status of the remote process will be equal to the last value that was set with this function.

In the case that the commandline invocation is local, the situation is slightly more complicated. If the commandline invocation results in the mainloop running (ie: because the use-count of the application increased to a non-zero value) then the application is considered to have been ‘successful’ in a certain sense, and the exit status is always zero. If the application use count is zero, though, the exit status of the local ApplicationCommandLine is used.

This method is a no-op if done() has been called.

Added in version 2.28.

Parameters:: exit_status – the exit status

Properties

class ApplicationCommandLine

props.arguments: Variant: The commandline that caused this command_line signal emission.

Added in version 2.28.

props.is_remote: bool: Whether this is a remote commandline.

Added in version 2.28.

props.options: Variant: The options sent along with the commandline.

Added in version 2.28.

props.platform_data: Variant: Platform-specific data for the commandline.

Added in version 2.28.

Virtual Methods

class ApplicationCommandLine

do_done() → None

Signals that command line processing is completed.

For remote invocation, it causes the invoking process to terminate.

For local invocation, it does nothing.

This method should be called in the command_line handler, after the exit status is set and all messages are printed.

After this call, set_exit_status() has no effect. Subsequent calls to this method are no-ops.

This method is automatically called when the ApplicationCommandLine object is disposed — so you can omit the call in non-garbage collected languages.

Added in version 2.80.

do_get_stdin() → InputStream | None

Gets the stdin of the invoking process.

You must only call this function once per commandline invocation.

Added in version 2.34.

do_print_literal(message: str) → None

Prints a message using the stdout print handler in the invoking process.

Unlike print(), message is not a printf()-style format string. Use this function if message contains text you don’t have control over, that could include printf() escape sequences.

Added in version 2.80.

Parameters:: message – the message

do_printerr_literal(message: str) → None

Prints a message using the stderr print handler in the invoking process.

Unlike printerr(), message is not a printf()-style format string. Use this function if message contains text you don’t have control over, that could include printf() escape sequences.

Added in version 2.80.

Parameters:: message – the message

Fields

class ApplicationCommandLine

parent_instance

priv