to_bytes/from_bytes() requires a newer version of the Math::BigInt::GMP library

[eserte]

The test suite fails on a few of my smokers:

...
from_bytes() requires a newer version of the Math::BigInt::GMP library. at t/04-Helpers.t line 20.
# Tests were run but no plan was declared and done_testing() was not seen.
# Looks like your test exited with 255 just after 1.
t/04-Helpers.t ......... 
Dubious, test returned 255 (wstat 65280, 0xff00)
All 1 subtests passed 
...
to_bytes() requires a newer version of the Math::BigInt::GMP library. at /home/cpansand/.cpan/build/2021092212/Bitcoin-Crypto-1.002-1/blib/lib/Bitcoin/Crypto/Helpers.pm line 144.
# Tests were run but no plan was declared and done_testing() was not seen.
# Looks like your test exited with 255 just after 1.
t/14-ExtPublicKey.t .... 
Dubious, test returned 255 (wstat 65280, 0xff00)
All 1 subtests passed 
...

[bbrtj]

Thanks for the report, but I'm not sure if I can do anything about it

• Bitcoin::Crypto does not depend on GMP, it is optional
• it tries to load Math::BigInt GMP backend before LTM to avoid a nasty bug with mixing Math::BigInt backends (plus GMP seems to be a tiny bit faster)
• users may or may not have GMP, but since it is not in dependencies, I don't think I can have any control over their version
• I already load Math::BigInt with a explicitly stated version, but that does not affect the backend

I'm open to suggestions on how to fix this

[eserte]

You may specify the dependency if the user have already an older version of Math::BigInt::GMP installed, for example by using code like this in Makefile.PL (untested):

my $minimum_Math_BigInt_GMP_VERSION = ...;
if (eval { require Math::BigInt::GMP; 1 } && $Math::BigInt::GMP::VERSION < $minimum_Math_BigInt_GMP_VERSION ) {
    ... extend PREREQ_PM ...
}

Or use skip_all on the two failing tests if you detect that Math::BigInt::GMP is too old here.

[bbrtj]

Hello @eserte,
I'm having trouble figuring out which Math::BigInt::GMP version I should require. Math::BigInt::GMP never got the required methods (_to_bytes, _from_bytes), instead it relies on inheritance from Math::BigInt::Lib to get it. I already require Math::BigInt (which contains Math::BigInt::Lib) in the sufficient version, which has the required methods.

Is it possible that these smokers have different version of Math::BigInt::Lib than Math::BigInt? Could you check the version of Math::BigInt, Math::BigInt::Lib and Math::BigInt::GMP on one of the smokers?

[bbrtj]

Seems like Math::BigInt::GMP 1.6003 was the first one to get an author test for existence of _to_bytes method, so it is the most probable version to require. I'm still interested why is it failing in the first place though

[bbrtj]

None the tests of my development release reported the problem with Math::BigInt::GMP version, so I'll upload a new version and close this ticket. Thanks!