update README

This commit is contained in:
Andrew Kelley 2016-01-26 01:39:45 -07:00
parent 4e43973413
commit bc89614996

View File

@ -12,25 +12,39 @@ Zig is not afraid to roll the major version number of the language if it
improves simplicity, fixes poor design decisions, or adds a new feature which
compromises backward compatibility.
## Goals
## Existing Features
* Compile units do not depend on libc unless explicitly linked.
* Provides standard library which competes with the C standard library and is
always compiled against statically in source form.
* Pointer types do not allow the null value. Instead you can use a maybe type
which has several syntactic constructs to ensure that the null pointer is
not missed.
* Provides an error type with several syntatic constructs which makes writing
robust code convenient and straightforward. Writing correct code is easier
than writing buggy code.
* No header files required. Top level declarations are entirely
order-independent.
* Powerful constant expression evaluator. Generally, anything that *can* be
figured out at compile time *is* figured out at compile time.
* Tagged union enum type. No more accidentally reading the wrong union field.
* Easy to parse language so that humans and machines have no trouble with the
syntax.
* The binaries produced by Zig have complete debugging information so you can,
for example, use GDB to debug your software.
* Release mode produces heavily optimized code. What other projects call
"Link Time Optimization" Zig does automatically.
* Supported architectures: `x86_64`
* Supported operating systems: Linux
## Planned Features
* Completely compatible with C libraries with no wrapper necessary.
* In addition to creating executables, creating a C library is a primary use
case. You can export an auto-generated .h file.
* Do not depend on libc unless explicitly linked.
* Provide standard library which competes with the C standard library and is
always compiled against statically in source form.
* Generics so that one can write efficient data structures that work for any
data type.
* No null pointer. Convenient syntax for dealing with a maybe type so that
null pointer is not missed.
* An error type, combined with some syntatical constructs which makes writing
robust code convenient and straightforward.
* Eliminate the need for configure, make, cmake, etc.
* Eliminate the need for header files (when using zig internally).
* Tagged union enum type. No more accidentally reading the wrong union field.
* Easy to parse language so that humans and machines have no trouble with the
syntax.
* Eliminate the preprocessor, but (most) everything you can accomplish with
the preprocessor, you can accomplish directly in the language.
* Ability to mark functions as test and automatically run them in test mode.
@ -42,22 +56,9 @@ compromises backward compatibility.
* Shebang line OK so language can be used for "scripting" as well.
* Debug mode optimizes for fast compilation time and crashing when undefined
behavior *would* happen.
* Release mode produces heavily optimized code. What other projects call
"Link Time Optimization" Zig does automatically.
### Current Status
* Have a look in the example/ folder to see some code examples.
* Most language features are available, but many edge cases and errors are
not yet implemented.
* Linux x86_64 is supported.
* Building for the native target is supported.
* Optimized machine code that Zig produces is indistinguishable from
optimized machine code produced from equivalent C program.
* Zig can generate dynamic libraries, executables, object files, and C
header files.
* The binaries produced by Zig have complete debugging information so you can,
for example, use GDB to debug your software.
* Compiler exposes itself as a library.
* Support for all popular architectures and operating systems.
* Easy cross-compiling.
## Building