The examples/webserver app can be built in two modes:
(1) in standalone mode, or
(2) as a NSH built-in app.
When run in standalone mode, the webserver program is responsible for
bringing up the network (including DHCP if configured). Also, the
webserver program must never exit, so if httpd fails (i.e., if
httpd_listen() returns), webserver_main() goes into an endless loop.
When run as a NSH built-in app, network bring-up is the responsibility
of other processes and the webserver program assumes the network is
already properly configured when it starts. Also, if httpd_listen()
returns, the webserver program should terminate.
Prior to this change, the webserver program would *not* terminate,
even when running as a NSH built-in app. For example:
nsh> webserver &
webserver [6:100]
nsh> Starting webserver
nsh> kill -9 6
nsh> webserver_main: Still running
nsh> webserver_main: Still running
nsh> webserver_main: Still running
nsh> webserver_main: Still running
The line "webserver_main: Still running" would be forever printed
every 3 seconds, however httpd_listen() is no longer running and the
webserver is not functional.
This change makes the webserver play nicely when running as a NSH
built-in app. With this change applied:
nsh> webserver &
webserver [6:100]
nsh> Starting webserver
nsh> kill -9 6
nsh> webserver_main: Exiting
apps/examples/webserver/webserver_main.c:
* main(): Infer from CONFIG_NSH_BUILTIN_APPS if this is a
standalone program or a NSH built-in app. (See [1], where
similar logic was added to decide whether to do network bring-up
or not.) If standalone, run forever as before. If built-in
app, exit when httpd terminates.
References:
[1] Commit 3a21b0b22263c06b5ba54ea2525521c8ca84b349
819454ee6e