precompiles: Optionally use silkpre as implementation #660
Conversation
|
A Clang build of silkpre does not work. I'm getting |
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #660 +/- ##
==========================================
+ Coverage 97.41% 97.43% +0.02%
==========================================
Files 81 83 +2
Lines 8013 8042 +29
==========================================
+ Hits 7806 7836 +30
+ Misses 207 206 -1
Flags with carried forward coverage won't be shown. Click here to find out more.
|
Not sure if it's relevant, but perhaps erigontech/silkpre#8 can help. |
Add CMake option EVMONE_PRECOMPILES_SILKPRE (default OFF). If enabled CMake will fetch the silkpre precompiles library. https://github.com/torquem-ch/silkpre
chfast
force-pushed
the
state_precompiles_silkpre
branch
from
72c533c to
12955a1
Compare
This seems coming from libff. |
chfast
force-pushed
the
state_precompiles_silkpre
branch
2 times, most recently
from
f067d1b to
1717844
Compare
| @@ -172,6 +176,16 @@ inline constexpr auto traits = []() noexcept { | |||
| {ecpairing_analyze, dummy_execute<PrecompileId::ecpairing>}, | |||
| {blake2bf_analyze, dummy_execute<PrecompileId::blake2bf>}, | |||
| }}; | |||
| #ifdef EVMONE_PRECOMPILES_SILKPRE | |||
There was a problem hiding this comment.
why not #ifdef with two branches directly in the initialization above ?
There was a problem hiding this comment.
This is structured as optional override. It also allows us quickly change what we override with silkpre.
This is from the stTimeConsuming.CALLBlake2f_MaxRounds state test.
chfast
force-pushed
the
state_precompiles_silkpre
branch
from
97f3493 to
3f06da2
Compare
Add CMake option (off by default)
EVMONE_PRECOMPILES_SILKPREto enable building of silkpre as the precompiles implementation.