bodrato at mail.dm.unipi.it
Fri Dec 21 22:36:22 UTC 2018
Il Ven, 21 Dicembre 2018 7:13 pm, Niels Möller ha scritto:
> tg at gmplib.org (Torbjörn Granlund) writes:
>> mini-gmp's main goal of being small en simple.
> Agree, we should not compromise simplicity. To me it might make sense to
I agree too.
Before my changes, lcov (
https://gmplib.org/devel/lcov/shell/gmp/mini-gmp/ ) reported for the file
Date: 2018-12-15 08:18:37
Line Coverage: 98.5 % 2034 / 2066
Functions: 99.0 % 197 / 199
Branches: 87.6 % 1117 / 1275
Now it says:
Date: 2018-12-21 08:03:29
Line Coverage: 98.4 % 2029 / 2061
Functions: 99.0 % 196 / 198
Branches: 87.4 % 1107 / 1267
I mean, the lines of code, and branches actually used if mp_limb_t is
unsigned long, did not increase.
> 2. Reduce the number of places that depend on mp_limb_t == unsigned
> long, where that also makes code simpler. E.g, using init_set_ui or
> roinit_ui in more of the _ui functions.
I agree, let's do that!
> 3. Make it easy to omit remaining the functions/macros depending on
> mp_limb_t == unsigned long, to enable some separate wrapping which
> defines substitutes and then includes mini-gmp.c.
That's an interesting approach. After the step 2 you proposed, we will
decide the best way: code protected with compile time optimised if (cond),
or completely removed by #ifdef with replacements in another file.
More information about the gmp-devel