struct bitfields - Re: C trivia - was Re: Dennis Ritchie has died

On 16 Oct 2011 at 11:03, Toby Thain wrote: Not the same at all. I don't believe that in C I can declare: bit anything[60000]; K&R have bitfields as comprising parts of larger datatypes, say int, but not as a datatype in their own right. Can you imagine how inconvenient it would be if the type "char" was available only as a member of type "int"? Bit arrays are enormously useful to facilitate vector math, either to accompany sparse arrays or as a control vector for storing results. And then there's the obvious application as a simple bitmap. As I said, K&R C is extremely myopic and to my eye always appeared as PDP-11 shorthand assembler. It's not as if bit vectors are anything new; APL has had them for a very long time. Another thing that's dogged me over the years is the extremely weak preprocessor in C. To be sure, the template feature in C++ has remedied a lot of it, but that's not C--it's C++. C's preprocessor never came up to the level of the macro facility in a good macro assembler or, for that matter, the preprocessor in PL/I. Yes, I know the recommendation is to use a generic external macro facility, such as M4 or a host of other tools, but none of those are part of the C language. What can be done with a really good macro-assembler is very eye- opening. Sadly, not many newer assemblers, particularly those for microcontrollers, have assemblers worth spitting at. --Chuck