Cleaned code, removed smc.h, fixed typos

[perfaram]

List :

• Cleaned code (by removing smc.h and transferring missing typedefs into SMCWrapper.h, by wiping useless macros, by reordering typedefs, by having all includes in the same place - the header, by normalising carriage returns and indentation)
• Corrected some silly mistakes (SMCWrapper.m includes its header, fixed a segfaulting typo that took me an hour to find)
• stringRepresentationForBytes:...:toNSString now actually returns a value
• Added methods to request raw hex values (e.g. 68 65 6c 6c 6f) from SMCVals
• Added FAILURE_REPRESENTING_STRUCT in SMCState enum, to be used when the SMCVal couldn't be represented (through stringRepresentation... methods). (NOT YET IN USE)

[perfaram]

Bump ?

[perfaram]

Take a look at the 'SMCWrapper' branch in http://github.com/perfaram/xline, this is where most improvements I made here come from.

[perfaram]

Just asking, why are we passing NSStrings by reference anyways ? Couldn't we just return the NSString (which is usually easier for the programmer who uses SMCWrapper) and have the error passed by ref ? Or have a @property for errors, making the whole thing very much easier, especially knowing that we are going to have more detailed errors ?

Cleaner list of options, in which Error stands either for NSError** or SMCState_t * :

1. -(BOOL) readKey:(NSString *)key intoString:(NSString **)str; (current)
2. -(NSString*) readKey:(NSString *)key withError:(Error **)err; (passing byref)
3. -(NSString*) readKey:(NSString *)key; (setting @property in method)

I vote for the last one.

[perfaram]

After working a bit on other projects, I changed my mind. The current version seems good. BUT, to have more precision, we could also have a @property [...] error, so that it doesn't break with current version, but we can have additional data for debugging.