Fix ISO c++ does not allow assigning string literals to char*

[peter-urban]

First partial fix for #77

This fixes the c++ warnings related to assigning constant strings to char*
The fix is simple and works in c and c++ :

• replaced several char* with const char*

Most of the changes are unproblematic, since they only happen in gsw_check_functions.c

However, there is one change in gsw_saar_data.h (line 6,7 and 9). I wouldn't mind someone with better C command to double check this:

• const char** is a mutable pointer to a const char* and not a const pointer to mutable char* right?
• Also: While this is a minor change, it modifies the interface of gsw_get_version (const char** instead of char**). So it potentially require the libraries building on top of GSW-C to be updated? (e.g. GSW-Python)

[efiring]

@peter-urban thank you for jumping on this. But why are about 500,000 changes showing up in gsw_saar_data.h?

[efiring]

gsw_get_version is not used in GSW-Python, and I doubt it is used anywhere else.
Nevertheless, we might as well get it right. @mauzey1, I suspect you are the expert here,
so if you have time, please advise on this PR and #77.

[peter-urban]

@peter-urban thank you for jumping on this. But why are about 500,000 changes showing up in gsw_saar_data.h?

Oops, that was accidental auto-formatting. Should be fixed now.

[efiring]

Googling turned up a nice utility program:

$ cdecl explain "const char **x"
declare x as pointer to pointer to const char

[peter-urban]

Reverted the changes I mentioned in the last comment.
It caused problems that I don't intend to fix in this PR.

(multiple_declaration failures when including gsw_saar_data.h because gsw_get_version is neither declared static or inline.)

[peter-urban]

However, I believe the function does exactly the same as it did before this PR, so it should be fine

[efiring]

@peter-urban what does your latest force-push do? And do we need to keep all 3 commits, given that numbers 2 and 3 just cancel out? I could use the "squash and merge" option to simplify the history.

[peter-urban]

Sorry for the force push. I just rebased the branch to the latest main not realizing that this may cause confusion after approval.

[peter-urban]

I squashed the commits; (I actually didn't know that this was possible until you just mentioned it)