Utilize blockwise I/O in ModelicaInternal_copyFile

[beutlich]

Blockwise I/O usually is faster then byte-wise I/O.

[HansOlsson]

Really good general idea, but:

• Always use "rb" and "wb". They are C-standard and supposed to work the same on Linux.
• It is unclear where BUFSIZ comes from.
• Is it intentional to ignore read-errors? Since the previous version ignored all errors that may be ok

[beutlich]

Always use "rb" and "wb". They are C-standard and supposed to work the same on Linux.

Yes, that is actually how I edited the code.

It is unclear where BUFSIZ comes from.

It is defined in the C standard library, in header stdio.h.

Is it intentional to ignore read-errors? Since the previous version ignored all errors that may be ok

How can read errors be detected? len==0 or len<BUFSIZ are completely valid read results.

[HansOlsson]

How can read errors be detected? len==0 or len<BUFSIZ are completely valid read results.

Well, one can use ferror() - but it is a bit messy and adds extra code.

[HansOlsson]

Looks good.
The extra check for read-errors can be added later.