Answer by Lundin for Calculate the free electron concentration and drift speed in a segment of wire

In addition to the other reviews, some misc comments:

• The switch RUN_REAL_PROGRAM is weird, doesn't seem like something that should be part of the final program - remove it.

• In general, avoid non-standard C if there is no good reason for it. That is, replace for example non-standard #pragma once with standard C #ifndef SOMETHING_H #define SOMETHING_H ... #endif.

• Don't create some generic header called "defines.h". Instead place those defines in the module where they actually belong. AVOGADROS_NUMBER should probably be in material.h (?), INPUT_BUFFER_SIZE should be in prompt.h and so on.

• I prefer to place all includes in the .h file rather than the .c file since the former is the interface to the user. The user needs to know all file dependencies of a .h + .c pair and shouldn't need to go digging through the .c file (which might not even be available) to find them.

• For very long function declarations/definitions, consider using new line style:

  char prompt_quit(const char *  msg,                  char *        buffer,                  size_t        num_chars,                  const char *  quit_str);

• Avoid "for ever" loops when they aren't called for. Also, the presence of continue in a C program is almost always an indication of a poorly designed loop. So instead of the somewhat obscure: while(1) { ... if(conversion_error) { ... continue; } ... return result; }, write something like this instead:

    char conversion_error = /* non-zero value */;  while(conversion_error)  {    ...    long result = cstr_to_long(buf, &conversion_error);    if(conversion_error) {      fputs("Invalid input\n\n", stdout);    }  }  *request_quit = 0;  return result;} // end of function

• If conversion_error is only either 1 or 0 then it should be bool not char. Overall, it is common convention to reserve the return type of functions for the error handling, especially when supported more detailed errors (through an enum type etc).

• Your list of different materials should be implemented as an array of struct instead.

• All constants like 36.0 and 39 have a type in C, in this case double and int respectively. Always avoid mixing floating point and fixed point in the same expression, since that's an easy way to get subtle bugs. For example 123 / 456 + 789.0 first gets the 123 / 456 calculated on int then implicitly promoted to double afterwards.

  So for example change return 0.127 * pow(92, (36.0 - awg) / 39); to return 0.127 * pow(92.0, (36.0 - awg) / 39.0);

• Is AWG 0 really a thing?