Byte Ordering. Answer: The choice is made by the computer designers, and varies among architectures. Two standard choices:

Transcription

1 Byte Ordering Memory is byte-addressible, but the CPU manipulates multi-byte values, such as 4-byte integer or 8-byte double-precision floating point values. Question: When a 4-byte integer is stored in memory, which of the four bytes is stored in the lowest-numbered address? the highest-numbered address? Answer: The choice is made by the computer designers, and varies among architectures. Two standard choices: Little-endian: The least-significant byte goes at the lowest address. Big-endian: address. The most-significant byte goes at the lowest (Endianness)

2 Big-Endian and Little-Endian Byte Orderings

3 Endianness: Why Do You Care? If you write a multi-byte quantity into memory and then read it back one byte at a time, you will get different results on different architectures. Example: You write the integer value 260 (a 4-byte quantity) at address You read back bytes from locations, 100, 101, 102, 103: little-endian: You get 4, 1, 0, 0. big-endian: You get 0, 0, 1, 4. The time this makes a difference is when you are reading or writing multi-byte quantities from memory to external storage or a network.

4 Alignment Another issue related to multi-byte quantities is that the CPU will only read or write them to memory in an aligned fashion. 1-byte values can be read or written to any address. 2-byte values can only be read or written to even-numbered addresses. 4-byte values can only be read or written to addresses that are multiples of 4. etc. If you try to read or write a multi-byte quantity to a non-aligned address, an exception will occur.

5 Encoding Characters The ASCII character set or various Extended ASCII character sets use one byte per character, for a maximum of 256 distinct characters. Many legacy files are encoded in this fashion. The Unicode standard has over 1 million code points (currently U+0000 to U+10FFFF), of which 109,449 are currently used (Unicode 6.0).

6 ASCII

7 Extended ASCII

8 Unicode

9 Encoding Text Strings Text is typically encoded in memory by placing character codes in successive locations. The Java char data type consists of 16-bit unsigned 16-bit integers which can directly represent the first 65,536 Unicode code points. Strings in Java use UTF-16 encoding, which uses 16-bit codes placed consecutively in memory. Each Unicode code point is represented by either one or two 16-bit codes. Byte order marks are used at the beginning of UTF-16 text files, to detect endianness. UTF-8 is a scheme for encoding Unicode as a a sequence of bytes, which is backward compatible with ASCII. It is now commonly used in communication.

10 Introduction to C: History In 1970, Ken Thompson of Bell Labs designed the B language for the first Unix system on the DEC PDP-7. Soon after, the C language was designed by Dennis Ritchie at Bell Labs for implementing Unix on the PDP-11. In 1978, Brian Kernighan and Dennis Ritchie published the first edition of The C Programming Language (K&R). About 1981, Bjarne Stroustrup of Bell Labs defined an extended C language that eventually evolved into C++. In 1983 ANSI established a committee to define a standard C. In late 1988 the ANSI C standard was submitted for approval. The language includes some of Stroustrup s extensions. Today C is no longer changing, but C++ continues to evolve separately from ANSI C.

11 Ken Thompson 1970, Bell Labs Created B language Used in 1 st UNIX system (DEC PDP 7)

12 Dennis Ritchie

13

14 The C Programming Language

15 Bjarne Stroustrup 1981 Bell Labs Created C++

16 ANSI C versus K&R C There are two main dialects of C that you might encounter: K&R and ANSI. K&R (Kernighan and Ritchie) C is the old version of the C language, defined about ANSI C is the new standard, defined in ANSI C includes significant extensions, primarily w.r.t type checking, that make it an improvement over K&R C. Most new C programming is now done in ANSI C. We will focus exclusively on ANSI C. K&R C is only important if you work on legacy code.

17 C and Java C and Java look superficially alike, because Java adopted C s syntax. But... C and Java are very different! C and Java have much less in common than you might expect from a cursory glance.

18 Java-to-C: Forget Everything! To move from Java to C, you first have to forget everything you learned from Java really! Forget objects, forget classes, forget methods. Forget packages, forget interfaces, forget exceptions. Forget new, forget garbage collection. Forget array bounds-checking. Forget the JVM and on-demand class loading. Forget platform independence.

19 Java-to-C: Add Low-level Features C adds low-level, potentially dangerous features that Java does not have: Add macros. Add pointers. Add structures. Add malloc and free. Add manual linking. C requires a disciplined programming style to avoid pitfalls and simulate facilities that Java provides for free.

20 C is High-level Machine Language C began as high-level machine language for the DEC PDP-11 computer. Many C constructs reflect closely the PDP-11 instruction set. Even today, the underlying machine shows through to a great extent in C. The programmer must understand the underlying machine and pay close attention to low-level considerations. C philosophy: Programmers know what they are doing. They can do whatever they want, as long as they say it explicitly. This can lead to trouble...

21 A Simple C Program Here is the classic Hello world! program in C: #include <stdio.h> int main(void) { printf("hello world!\n"); }

22 A Simple C Program: Notes C programs have functions, not methods. Every C program must have a main() function. printf() is a C library function. (it is not a C language construct) #include is a preprocessor directive. (handled by the C preprocessor)

23 Compiling a C Program The C compiler is traditionally invoked as a command-line tool: $ gcc hello_world.c (the $ is the command shell prompt). On Unix-like systems, this produces the executable file a.out. On Windows, it produces a.exe. The executable is launched by giving its name as a command: $ a.out Hello world!