In case ARGTYPE=3; the register functions do not work

[danzadok]

Prerequisites

• Checked the developer manual
• Checked that your issue isn't already filed: https://github.com/issues?utf8=✓&q=repo%3Alibtom%2Flibtomcrypt
• Checked that your issue isn't related to TomsFastMath's limitation that PK operations can by default only be done with max. 2048bit keys

Description

[Description of the issue]

Steps to Reproduce

Version

You can get this information from the define SCRYPT in src/include/tomcrypt.h or your local git repository by running git describe --always --tags --dirty.
Also, please include the compiler, the compiler version, the architecture and (if applicable) the MPI provider, the OS and what version of the OS you're experiencing the issue.

Additional Information

Any additional information, configuration or data that might be necessary to reproduce the issue.

[danzadok]

This is a macro definition error

[sjaeckel]

1. can you please write proper bug reports and not just an issue line? those templates exist for a reason.
2. ARGTYPE=3 is a very bad idea
3. I've created a fix/474 branch which should fix the issue when ARGTYPE=3
4. there seems to be a regression somewhere in either ECC (ping @karel-m) or the updated prime checking in libtommath (ping @czurnieden)! currently the latest libtommath can't be used to run the ltc ECC tests!

[karel-m]

I confirm that with current libtommath/develop the ecc_tests and dh_test are failing

[karel-m]

libtommath's mp_prime_is_prime considers the following prime as a non-prime

15525180923007089351309181312584817556313340494345143132023511949029662399491021
07258669453876591642442910007680288864229150803718918046342632727613031282983744
380820890196288509170691316593175367469551763119843371637221007210577919

FFFFFFFFFFFFFFFFC90FDAA22168C234C4C6628B80DC1CD129024E088A67CC74020BBEA63B139B22
514A08798E3404DDEF9519B3CD3A431B302B0A6DF25F14374FE1356D6D51C245E485B576625E7EC6
F44C42E9A63A3620FFFFFFFFFFFFFFFF

[czurnieden]

@karel-m libtommath has been compiled how exactly for that test?

There are a bunch of defines to include/exclude algorithms. the default is M-R with bases 2 and 3, mp_prime_strong_lucas_selfridge and at least one M-R test with a random base.

The API changed for mp_prime_is_prime, too, although it should be down-compatible, so how did you call it?

I have no problem with the develop branch in my fork (v1.0.1-181-g48c95f2), where that prime (it is a certified prime, just if somebody asks) gets detected as a prime when checked with testme.sh (actually the whole bunch of tests from .travis.yml ). mp_prime_is_prime has been called for that prime with t = 8, that means M-R with bases 2,3, strong-lucas, and 8 M-R tests with random bases.

Great.
And I was already thinking: "Oh, that get's a bit too smoothly, doesn't it?"
*grrr*
*sigh*

[sjaeckel]

#476 should've fixed this... and some more :)