Add max-msg run-time option

[RosscoG]

Intro

Thank you @mity for a great tool that I use for every bit of C I work with.
I wanted to get into contributing to open source projects for practice and internet points. Working on something I have a passion for and am familiar with seemed like a good place to start.

I have implemented the cmd line option for changing message output limit at run-time.
Resolves #33

New features

Added run-time option to change the max output message length without having to recompile. Features:

• cmd line option --max-msg=NUMBER
• TEST_MSG_MAXSIZE is kept and used as default. It can still be defined before acutest include.
• Added --help line for --max-msg usage
• Changes match formatting, style and layout of existing code.

Tested with:
x64 Windows: gcc -Wall -Wextra -pedantic.

Recreate issue and Verify solution

A test rig was setup to recreate a character limited test message.
Length of test message was larger than TEST_MSG_MAXSIZE (1024), increased to (1024 + 200 - 24) = 1200 characters (for a nice round number).
On a failing test, as shown below, the # symbols represent long TEST_MSG output. As expected, the character count is 1024 - 1 (null terminated), defined by TEST_MSG_MAXSIZE. It does not show the final sentence, confirming the long message output is limited by the compile-time defined value.

Normal default length limited output:

Using the new run-time option --max-msg=1200, the full message can be seen without having to recompile. It includes the closing words right up to the final period, verifying intended behaviour.

Run-time option extended output:

Re-compiling with pre-defined TEST_MSG_MAXSIZE 1200 is confirmed to still operate as expected, needing no option to show full message.

Compile-time define extended output:

Bad input argument behaviour

There are several possible arg errors, nonsense input or genuine user misunderstanding. eg. user could misinterpret "--max-msg" is to set max possible etc. To handle these type of cases, the acutest_cmdline_options_ argument field was set to ACUTEST_CMDLINE_OPTFLAG_REQUIREDARG_. Other bad input is handled by setting exit value and usage hint output, matching other options.

Bad arg outputs:

1. Missing arg
2. Empty NUMBER arg
3. Non-numeric NUMBER arg
4. Negative NUMBER arg

--help options output includes --max-msg line:

Conclusion

This implementation meets requirements and is sufficiently verified for expected behaviour and error handling (IMO). I have not tested on other machines although I do not see these changes causing any problem. I will provide an update if anything arises from that testing.

Issue #34 is an equivalent feature for TEST_DUMP_MAXSIZE. I think it would be best to have that implemented for consistency. I'm happy to do that one next.