[libmath-prime-util-gmp-perl] annotated tag upstream/0.23 created (now 6bd178e)
gregor herrmann
gregoa at debian.org
Mon Sep 15 18:57:50 UTC 2014
This is an automated email from the git hooks/post-receive script.
gregoa pushed a change to annotated tag upstream/0.23
in repository libmath-prime-util-gmp-perl.
at 6bd178e (tag)
tagging f7af018ea0fcac412684ac972275d77b492cb414 (commit)
replaces upstream/0.21
tagged by gregor herrmann
on Mon Sep 15 20:54:23 2014 +0200
- Log -----------------------------------------------------------------
Upstream version 0.23
Dana Jacobsen (380):
Initial commit
Initial checkin with miller_rabin code
Strong Lucas pseudoprime, is_prime, is_prob_prime
Fix memory leaks with mpz objects
Return 2 for def prime
Move code to .c file, add next_prime, prev_prime, rudimentary factoring
Add HOLF and P-1, fix for general factor
small GMP performance enhancements
More factoring code
Speedup for p-1
Fix prototypes
Factoring speedups
Very small factor changes
Stage 2 for p-1
Factor breaks composite factors now, p-1 second stage improved a lot
add primes method, write most of the documentation
Add tests, update for impending release
Factoring update
Add Devel::CheckLib, so we'll only try to install if they have GMP
Update date
Rewrite factoring tests
Update for v0.03
Fix issue with 5.6.2 64-bit
Add lcm to XS
Changes for pedantic ansi compiling
Turn on debug for the p-1 factoring, in vain attempt to see why CPAN testers are seeing failures that I cannot duplicate
Uninit variable
Update Changes for 0.03
Add prime_count
Don't include stdint on MS compilers
Remove commented out dependencies on Math::Big*
Add primorial and consecutive_integer_lcm functions
Make primorial and pn_primorial match MPU.
Fix typo in internal next_small_prime function. Broke primorial.
Fix HOLF issue with perfect sqrts, add simple perfect power factoring test
Make factor() test for perfect powers, add ability to call trial division
Make ..._factor functions test for 2/3/5/7 divisibility first.
Add quick Pocklington primality test and make is_prime() call it
Add basic ECM factoring
Run a bit more initial p-1
Add BLS primality test, is_provable_prime, some factor adjustments, enhancements to SQUFOF
Complete rewrite of p-1 factoring
Fix compiler warnings
Another p-1 optimization, doubles speed of stage 1
Efficiently precalc the first 20 differences, simplify lazy calculations
Small p-1 change
Try out Travis CI
Remove Perl 5.8 from Travis
Rewrite consecutive_integer_lcm to not use sqrt or log
Cleanup for impending CPAN release
Add a couple more files to .gitignore
Add AKS test
7000 tests -> 2 is_deeply tests. Big speedup on slow systems, same testing.
Speedup affine ECM
Small prime iterator
Update for release
Fix 1-byte memory overflow
Add small factoring, setup for next release
Handle Math::BigInt inputs a little better
Speedups for factoring
Update documentation
M-R with large bases
Travis-CI seems to have changed
Final changes for release
Split out ECM, add QS (SIMPQS from old FLINT)
Some QS tuning
Factoring speedups (QS and mix)
Restructuring QS
Update KS, round 1
Redo K-S algorithm, part 2
Add simple is_prime functionality
New ECM code
Compilation cleanup
More code cleanup
Only use primes for K-S multipliers
Fix compile warnings
Adjust firstprimes; compilation cleanup; a couple memory issues
Update docs
Change factoring mix
Convert SIMPQS to New/Safefree
Small updates
Switch to non-normalized for stage 1. A little slower.
Adjust mix -- remove pbrent, add tiny ECM
QS returns multiple factors, adjust factor and BLS75 factoring mix
Use native int factoring for small values in BLS75 test
Speed up primality proofs
More primality proving rework
Code cleanup
Turn on random seeding for ECM
A little code cleanup
Update for release
Code cleanup
Add PRAC as alternative EC multiplication
Minor cleanup
Update for release
Start moving some functions to utility.c
Verbose > 2 printing for p-1, ecm, and qs. QS does trialdiv to 1000.
Add Cornacchia and modified Cornacchia
Code cleanup
Add return of n-1 primality proof to is_provable_prime
Update version
Adjust primality proof return
Updates to primality proving
Fix for 32-bit machines
Huge AKS speedup: binary segmentation for polynomial multiplication
AKS: More accurate log2n => smaller r => fewer loops => slightly faster
Revert a change in mpz_order
Constify some tables
Add support (turned off) for BLS75 theorem 7
Start of function to return Hilbert polynomials
Fix preliminary hilbert polys, minor speedup for really big AKS values
Lots of HCPs, needs more storage efficiency work
Much better HCP storage method, verify and add a value from SAGE
Export affine multiply
Add commented out AKS helper function
Big updates to utility,c to add many new functions (internal C+GMP)
Change BLS proof to take effort int instead of boolean.
ECPP implementation
Minor ecpp update
Primality proof returns for ECPP
Callee responsible for proof memory allocation
Put BLS75 into its own file
Redo is_prime / is_provable_prime with ECPP and random MRs
Add spell checker to release testing
Add is_nminus1_prime(). Spelling fixes
ECPP: make stage 1 and 2 factoring lighter (faster but less factoring). Big speedup for small sizes.
Add a simple test for AKS, N-1, and ECPP
Comment changes
Perl 5.8.8 and earlier croak on newSVpv with null argument
typo in manifest
Updates for release
Turn on BLS75 theorem 7
Fix a T7 hang, use QS earlier on small inputs
Fix polyz_print. Don't croak if no roots found
Small p-1 stage 2 speedup (only mod b every 4 primes)
Add big poly data to xt directory
Change ECPP to factor all with backtracking
ECPP FAS fixes for factoring harder at step 0
Lower is_prime opportunistic proof effort
Actually commit the pseudoprime and is_prime effort code
Faster is_prob_prime (faster composite check, rewrite of Lucas)
Small changes to fix issues found on Win32
Update class poly file
Updates for release
Update Changes for release
Minor updates ahead of release
Some tweaks to is_prime number of tests, and streamline strong Lucas a little
Braces. Important.
Update Lucas test docs and tests
Add standard and extra strong lucas prp tests
Enable making a standalone ECPP
Plug memory leak for new standard Lucas test
Tweaks for ECPP
Some ECPP speedups and options
Add script to create standalone ECPP directory
Small speedups for class poly functions
ECPP tweaks
Merge pull request #2 from dsteinbrunner/patch-1
ECPP with n-1 shortcut needs to print something when verbose
Small changes for p-1 performance
Updates for release
Add FU pseudoprime
Trial division for ecpp via giant gcds
Add general Lucas sequence code, rearrange Lucas tests to combine and use it.
Add BLS75 theorem 3 and 15 tests for n-1 and n+1 support
Add integrated n-1 and n+1 tests. Still needs a lot of work.
Document the Frobenius-Underwood pseudoprime function
Speedup for large composites in is_prob_prime
For n+1 proof, send back p and q, not just d
Bump version number
Add lucas_sequence
Fix documentation typo
Add very simple p+1 factoring, is_prime does fewer extra M-R tests
Use a little p+1 and pbrent
Wrap gcd init inside init-status test
Add some degree 18 and 20 polys, and a few 21 and 23 also
Start changing certificates to new text format
Update Lucas tests
Add 'almost extra-strong' Lucas test
Bump up prime iterator sizes slightly
Sort factors largest to smallest. Reduces number and provides consistency.
Switch to extra strong BPSW. Clean up primality proofs.
Add Lucas D search limit, move perfect square test into D search
Small ECPP change
Much better p+1 factoring (10x faster), use in ECPP.
Optimize extra strong Lucas test
Add tests for various Lucas prp tests
Proof updates
Don't limit degrees as early with big numbers
Add todo info about missing ECPP discriminants
Add more polys. Complete through degree 17.
ECPP updates including text certificate verifier
Certificate verifier, primo converter.
Remove debugging GMP disable, doh
More primality proof adjustments
More efficient poly root finding
Make sqrtmod do verification of result
Certificate verifier now does Primo and MPU formats
Updates for cert verifier
Add tentative Weber root conversion for D%3=0,D%8=7 case
Base 10 is the default, don't spit it out in every cert
ECPP updates
Update todo file
Add C+GMP verifier for Primo and MPU
Weber roots with D%3=0 and D%8 = {1,2,5,6,7}
Speedup for pn_primorial
Add GMP-ECPP cert convert script
Merge
Updates for proofs
Add projective code (defined out)
Grammarian Pod Checker fixes
Fix UV / GMP UI mismatch issue with primorial optization
ECPP updates for release
Update Changes for release
Fix small certs
Fix broken tests to match fixed code
Updates for release
Merge pull request #3 from neilbowers/master
factoring commands should return F,R instead of R,F
Trim ecpp.c and modify poly degree heuristics
Extra compositeness tests, and Q composite doesn't imply N composite
Expand constants to allow huge proofs, add comments about projective
Use degree 5 minimum degree instead of 4 when trimming
Some more thoughts on projective EC (not compiled on by default)
Comment about crypto use of random primes
Add documentation for each proof type
Some more comments added to ECPP documentation
Allow different ECM B2, might speed up ECPP
cert v0.91
Add bigger gcd for 200+ bit prob primes, speed up pn_primorial
Slightly deeper ECPP poly search
Test string for divisibility by 2,3,5 before converting to GMP
Fix string test and disable for pseudoprimes
Initial treesieve + redo is_prob_prime trial
Next round of tree sieve changes
Speedup large primorials and cilcms.
Double up multiplies in pn_primorial
Clean up start of trial_factor
trial_factor no longer does a probable prime test
Merge branch 'master' of github.com:danaj/Math-Prime-Util-GMP
Next set of changes for tree sieve
Revert "Next set of changes for tree sieve"
Next set of changes for tree sieve
Swapped limit and primary limit meant we made the aux sieve every segment
Update version; adjust standalone ECPP files including a set of simple tests
Add 2013 to date. About time :)
Use dynamic size for treesieve
Add miller_rabin_random
Use floor(2*sqrt(n)) instead of 2*floor(sqrt(n)) in verify
Sort m's by size for performance
Use mpz_swap
More robust parsing and output tweaks
Mention skipped bad lines in verbose mode
Add return codes to standalone ECPP
Add bpsw option to usage text
Repeated factors could cause problems in BLS75 minimization
Merge branch 'master' of https://github.com/danaj/Math-Prime-Util-GMP
Update Devel::CheckLib to 1.01
Merge branch 'master' of github.com:danaj/Math-Prime-Util-GMP
Updates for release
Merge branch 'master' of https://github.com/danaj/Math-Prime-Util-GMP
Update documentation
factor in scalar context; miller_rabin_random should return only 0 or 1
Update Changes date for release
Update meta spec
Only use TEST_REQUIRES with new enough MM
Delete TEST_REQUIRES after merging
Remove testing code
Update resources
Fix memory leak
Have factor do QS earlier for 100-159 bit numbers
Add partitions function
Export partitions, bump version to 0.16
Merge branch 'master' of github.com:danaj/Math-Prime-Util-GMP
Document partitions
Add tests for partitions()
Fixes for is_aks_prime
Fix verbose output for n+1
Merge branch 'master' of https://github.com/danaj/Math-Prime-Util-GMP
Minor updates for release
Add verify_primegap script
fix decl after statement
Fix spelling error (RT 91561)
Primo now allows S=1 in ECPP proofs
Fix for word boundary problem in trial factor
Shortcut for iterator isprime
Add gcd, lcm, kronecker
Add a few more gcd/lcm/kronecker tests
Documentation for gcd lcm kronecker
Update date
Small changes
Add is_bpsw_prime documentation
5.6.2 has XSRETURN_IV but no XSRETURN_UV. What a POS.
Fix unsigned long isqrt for places where unsigned long != UV
Add another factoring test, get ready for release
Add improved AKS parameters
Fix AKS overflow in log
Adjust AKS r/s tradeoff for larger values
Add is_power
Add exp_mangoldt
Small performance changes for poly roots
Documentation and add stage 1 ECM for degree <= 2
More small ecpp performance tweaks
ECPP initial root speedup
ECPP: Better initial trial division
Add another big pseudoprime
ECPP: Bring back stage 0
AKS, remove unused variables, use prime r for v6 test (perf and simplicity)
Comment ECPP changes in Changes file
ECPP: make polyz_div perform mod n
ECPP: performance updates
Remove unused variable
ECPP: check all factors for a discriminant and select smallest. Smaller proofs, maybe faster, maybe slower.
Use Weber polys for more values, reduces overall size
Add a few small degree 22/24 polys to big set
Fix large poly coefficient truncation
Don't do both m/q pairs at once for larger numbers
Update class polys with latest tiny file
Update big ECPP class poly table
Update README for standalone ECPP
Merge branch 'master' of github.com:danaj/Math-Prime-Util-GMP
Adjust BLS75 factoring -- 70+ digit numbers do some more ECM before QS; do QS on larger numbers
ECPP changes for 1.03, step 1
ECPP changes for 1.03, step 2
Update TODO
Merge branch 'master' of github.com:danaj/Math-Prime-Util-GMP
Add changes text
Don't do QS for numbers over 90 digits -- too much memory
Add some text to Changes
Fix is_prime shortcut test for divisibility by 2 and 5
cast signed comparisons
ECPP: min degree a little larger, hard min of 7 instead of 6
Update to next release
Update for release
Add valuation, invmod, and is_pseudoprime
Go back to /40 for minH
Add binomial function
binomial tests and docs
Simplify negative k code
Split is_prob_prime second gcd from 20k into 10k or 40k. Speedup for 128-4000 bit inputs
Clean up next/prev_prime; handle multiples of 7/11/13/17/19/23 natively
Tiny speedup for is_prob_prime
TODO note about power tests in factoring
Fix iterator for very large inputs, raise sizes slightly
Comment out PP from benchmark
Don't do the Fermat pretest in ECPP factoring primality tests
Add gcdext and vecsum
Update version number
Update release date
UL -> UVCONST for gcd_ui pretest
Wrap a 64-bit test number in quotes
Bump version number
Move factor loop out of XS.xs and into factor.c
Do our own trial factoring
Prevent hang if B1 is very low
Make qs_factor exportable
Have HOLF return 0 if the gcd resulted in no factor
Merge all specific factor routines in XS.xs
Alloc 10 factors at a time; don't use mpz_remove in trial loop
Add moebius
Reduce some overhead in factor loop
Add ranged moebius and liouville
Add Euler totient and Jordan totient
Tests and documentation for moebius,liouville,totient,jordan_totient
Add carmichael_lambda
Add znorder
Fix compiler warnings
Add znprimroot
Note testing of primality pre-tests
Fix factoring mpz_get_ui issue
Update description
Update README description
use UL instead of UVCONST for GMP ui func
Fix spelling and Paul Zimmermann's name
standalone ecpp expression parsing
Change 64-bit UL to two multiplied 32-bit ULs
Update to standalone ecpp
Make standalone ecpp version 1.04
Fat comma isn't fat if left arg begins with a digit
Bump version number
David Steinbrunner (1):
typo fix
Neil Bowers (1):
Reformatted as per CPAN::Changes::Spec
gregor herrmann (1):
Imported Upstream version 0.23
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libmath-prime-util-gmp-perl.git
More information about the Pkg-perl-cvs-commits
mailing list