 |
README - libzahl - big integer library |
 |
git clone git://git.suckless.org/libzahl |
 |
Log |
|
Files |
|
Refs |
|
README |
|
LICENSE |
|
--- |
|
README (1220B) |
|
--- |
|
1 NAME |
|
2 libzahl - Big integer library |
|
3 |
|
4 ETYMOLOGY |
|
5 The bold uppercase 'Z' which represents the set of |
|
6 all integers is derived from the german word 'Zahlen', |
|
7 whose singular is 'Zahl'. |
|
8 |
|
9 DESCRIPTION |
|
10 libzahl is a C library for arbitrary size integers, |
|
11 that aims to be usable for robust programs, and be |
|
12 fast. |
|
13 |
|
14 libzahl will accomplish this by using long jumps |
|
15 when an error is detected, rather than letting the |
|
16 caller also perform a check. This shall make the |
|
17 code in the user program cleaner too. libzahl will |
|
18 use dedicated temporary bignum integers whether |
|
19 possible, and necessary, for its internal calculations. |
|
20 libzahl will not deallocate allocations, but rather |
|
21 cache them for reuse. |
|
22 |
|
23 With the exception of functions working with strings, |
|
24 all output parameters are before the input parameters. |
|
25 |
|
26 RATIONALE |
|
27 GMP MP cannot be used for robust programs. LibTomMath |
|
28 is too slow, probably because of all memory allocations, |
|
29 and has a nonintuitive API. TomsFastMath has an a |
|
30 nonintuitive API, has limited precision (selected at |
|
31 compile-time), and has limited functionality. All the |
|
32 above are also bloated. Hebimath is promising, but I |
|
33 think it can be done better. |
|
34 |
|
35 NOTES |
|
36 libzahl is currently not thread-safe. |
