The `new` and `delete` operators that don't need `libtsdc++` are also provided by module `sys/cpp_new_delete` for platforms that don't have the `libstdc++`. Since these operators in `sys/cpp11-compat` are just wrappers for standard `malloc`/`free` functions like in `sys/cpp_new_delete`, the `new`/`delete` operators that don't need `libstdc++` are used from `sys/cpp_new_delete`. Only the `new`/`delete` operators that require the `libstd++` are left in `sys/cpp11-compat`.
So it is sufficient for the use of the (re)defined `new`/`delete` operators that an application uses the module `sys/cpp_new_delete` instead of `sys/cpp11-compat` and thus also works on platforms without `libstdc++`.
This PR makes `event_timeout` and `event_timeout_ztimer` two distinct
pseudomodules, where the only api difference is in the init function.
If only `event_timeout_ztimer` is selected then no default ZTIMER
backend is selected and the old init function is not implemented.
If only `event_timeout` is selected then `xtimer` is used unless
`ztimer_usec` is included. In which case the `xtimer` wrapper on top
of `ztimer` is used and `xtimer` is not directly selected. This
allows for the legacy api to be supported with `ztimer_usec` as
a drop-in replacement.
If `event_timeout` and `event_timeut_ztimer` are selected then
`event_timeout` SRC file is excluded from compilation.
The `udp` command is a valuable debugging tool that is also useful
outside of the gnrc_networking example.
To enable easy sending of udp messages in other applications during
development, move the `udp` command to the shell module and introduce
the `gnrc_udp_cmd` pseudo-module to enable it.
This commit fixes the dependency resolution of `gnrc_netif_cmd_lora`.
As it was, this module was pulled by the driver if `gnrc` was used.
Besides pulling an extra dependency in applications that don't use
`shell_commands` or `gnrc_lorawan`, this hardcodes the module
resolution in the drivers.
This commit pulls `gnrc_netif_cmd_lora` if `shell_commands` and
`gnrc_lorawan` are present.
There is no clear reason why this should be configurable and causes an abstraction of USEMODULE.
The true reason is to make Kconfig easier but it seems having a variable in USEMODULE is a bit strange.
