Skip Menu |

This queue is for tickets about the Math-Prime-Util-GMP CPAN distribution.

Report information
The Basics
Id: 92508
Status: resolved
Worked: 1.5 hours (90 min)
Priority: 0/
Queue: Math-Prime-Util-GMP

People
Owner: DANAJ [...] cpan.org
Requestors: CHORNY [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Important
Broken in: 0.17
Fixed in: 0.18



Subject: 50-factoring.t hangs
Looking at cpantesters results, I think that it hangs on perls with 64-bit int. 1..145 ok 1 - factor(0) ok 2 - factor(1) ok 3 - factor(2) ok 4 - factor(3) ok 5 - factor(4) ok 6 - factor(5) ok 7 - factor(6) ok 8 - factor(7) ok 9 - factor(8) ok 10 - factor(16) ok 11 - factor(30) ok 12 - factor(57) ok 13 - factor(64) ok 14 - factor(210) ok 15 - factor(377) ok 16 - factor(403) ok 17 - factor(629) ok 18 - factor(779) ok 19 - factor(808) ok 20 - factor(989) ok 21 - factor(1363) ok 22 - factor(2310) ok 23 - factor(2727) ok 24 - factor(9592) ok 25 - factor(12625) ok 26 - factor(30030) ok 27 - factor(30107) ok 28 - factor(34643) ok 29 - factor(78498) ok 30 - factor(134431) ok 31 - factor(221897) ok 32 - factor(496213) ok 33 - factor(510510) ok 34 - factor(664579) ok 35 - factor(692759) ok 36 - factor(1228867) ok 37 - factor(2214143) ok 38 - factor(2463289) ok 39 - factor(3008891) ok 40 - factor(5115953) ok 41 - factor(5761455) ok 42 - factor(6961021) ok 43 - factor(8030207) ok 44 - factor(9699690) ok 45 - factor(10486123) ok 46 - factor(10893343) ok 47 - factor(12327779) ok 48 - factor(50847534) ok 49 - factor(114256942) ok 50 - factor(223092870) ok 51 - factor(455052511) ok 52 - factor(547308031) ok 53 - factor(701737021) ok 54 - factor(999999929) ok 55 - factor(2147483647) ok 56 - factor(4118054813) (hangs) -- Alexandr Ciornii, http://chorny.net
On Mon Jan 27 12:15:14 2014, CHORNY wrote: Show quoted text
> Looking at cpantesters results, I think that it hangs on perls with > 64-bit int.
First, huge thanks and I've reproduced on Cygwin. Can you point me to one of the results with 0.17? I don't see anything on the matrix or cpantesters.
Fixed in 0.18, just released. On systems where sizeof(UV) > sizeof(unsigned long), the integer sqrt overflow check wasn't properly working because it was set for UVs and GMP uses unsigned long. v0.17 added the test that exercised this path, because it was found on another platform.
Before resolution, check: https://buildd.debian.org/status/package.php?p=libmath-prime-util-gmp-perl&suite=sid to make sure all builds are successful. It looks like some are stopping in gcd, which is a different place. I have not been able to reproduce the gcd failure.