An argument parser that is truly zero-cost, similar to Unix's getopts.
getargs is a low-level, efficient and versatile argument parser that
works similarly to getopts. It works by producing a stream of options,
and after each option, your code decides whether to require and retrieve
the value for the option or not.
You don't have to declare a list of valid options up-front, so getargs
does not have to allocate space for them or spend runtime searching for
them. This also means that you have to write your own help message, but
since --help is just another flag, there are no restrictions on what
you do with it.
- Short
-fand long--flagflags - Required implicit values
-i VALUEand--implicit VALUE - Required or optional explicit values
-eVALUEand--explicit=VALUE - Positional arguments and
-- - Parse options at the beginning of the argument list, or anywhere
- Zero cost
- Zero copy
- Zero unsafe code
- Zero dependencies
- Zero allocation
- Simple to use yet versatile
#![no_std]-compatible- Compatible with
&strand&[u8]
getargs has had a lot of attention put into profiling and
optimization, and on a modern machine it takes under 0.2μs to parse a
short array of 12 arguments.
In our testing, getargs is faster than every other argument parsing
library on crates.io. Its closest competitor is gumdrop, which is only
~30% slower in the worst case, and its second-closest competitor is
getopt, which takes three times as long. Other libraries degrade
quickly; clap takes 45x longer. (This is not an entirely fair
comparison though, as clap is not just an argument-parsing library,
but an entire command-line application framework. It is perhaps
overqualified for simple tasks.)
For examples, see [the examples directory][./examples/] for small
programs that you can compile and run yourself to see how getargs
works.