mb2hal: Fix module info placement and cppcheck problem

[BsAtHome]

The new cppcheck in Ubuntu 26.04 (for future CI, see #4477) would trip on a define test of MODULE_VERBOSE. The placement was wrong in the header and the usage should be replaced with MODULE_INFO as used everywhere else.

This PR addresses the issue. It also removes the RTAPI dependent inclusion check of rtapi_app.h (probably copy-pasted from a compiled component). This is a user-space program that does not need rtapi_app_main() and uses plain main().

[grandixximo]

Good catch...

[BsAtHome]

Well, there is still a problem. Just installed a Ubuntu 26.04 in a VM and it still complains. Need to look deeper what is happening.

[BsAtHome]

The real issue here is that the MODULE_INFO stuff is only available when RTAPI is defined. However, that is never the case with user-space components and therefore, the whole MODULE_xxx exercise is a hopeless endeavor. It simply needs to be removed.

[hdiethelm]

The issue looks to be how cppcheck handles #ifdef. cppcheck has to figure out which cases it should check.

You see that here, it sees #ifdef MODULE_VERBOSE and does a check with MODULE_VERBOSE=1 due to --force, which of course fails:

Checking src/hal/user_comps/mb2hal/mb2hal.c ...
Checking src/hal/user_comps/mb2hal/mb2hal.c: MODULE_VERBOSE...
src/hal/user_comps/mb2hal/mb2hal.h:36:1: error: syntax error [syntaxError]
MODULE_VERBOSE(emc2, "component:mb2hal:Userspace HAL component to communicate with one or more Modbus devices");
^
Checking src/hal/user_comps/mb2hal/mb2hal.c: RTAPI..

If there are missing includes, this generates more errors, due to cppckeck does not find KERNEL_VERSION and just assumes KERNEL_VERSION=1 for example which of course also fails. As soon as rtapi.h is found, problem solved.

See:
https://cppcheck.sourceforge.io/manual.html Automatic configuration of preprocessor defines