[libmath-prime-util-perl] annotated tag upstream/0.68 created (now 1ebd1d6)
gregor herrmann
gregoa at debian.org
Sun Nov 5 17:36:15 UTC 2017
This is an automated email from the git hooks/post-receive script.
gregoa pushed a change to annotated tag upstream/0.68
in repository libmath-prime-util-perl.
at 1ebd1d6 (tag)
tagging 6f2a31dc7c31a2b55bde4a78eedd0ceab61e51d0 (commit)
replaces upstream/0.65
tagged by gregor herrmann
on Sun Nov 5 18:27:43 2017 +0100
- Log -----------------------------------------------------------------
Upstream version 0.68
Dana Jacobsen (108):
Add random_semiprime
random_semiprime: do first exp in native float
Better random_semiprime for unrestricted 65+ bit
Latest ramanujan prime limits
Speedup for is_carmichael
Slight savings for Fermat base 2
Split out random_semiprime and random_unrestricted_semiprime
overflow base changes, and Frobenius-Underwood test update
New Ei changed twin prime count results a bit
Hensel lifting in Montgomery inverse
Remove unused variable
Latest Ramanujan prime bound update.
New nth prime bounds and updated Ramanujan prime bounds
Rewrite sieve marking
Updates to Ramanujan prime bounds
Update Ramanujan prime bounds and Mersenne test limit
Push func defines above all includes
Change is_strong_pseudoprime for base=n case
New sieve
Change is_strong_pseudoprime bahavior for PP
Fix extra paren in big endian path
Better and faster initial limits for count and nth bounds
chinese objectifies bigint result
Consistent Lucas parameter selection. Add stronger Lucas test to primality.c.
Move entropy_bytes, srand, csrand to XS
Add forderange
Speedup for PP forderange
Add subfactorial example to doc
Update Mersenne exponent
Fallback entropy using timers
When calling PP csrand or srand, also call GMP seed
Add numtoperm, permtonum, randperm
If urandomm64 input is within 32-bit, pass to urandomm32.
Add shuffle
Rewrite PP vecany, vecall, etc. as much simpler forms
Spelling and coverage
git should ignore entropy .o file
Bump version number
More detailed comment about algorithm
Allow bigint rank when calling numtoperm
Don't call GMP lucas sequence with old version
compiler warnings
Work around old Perl and Math::BigInt::Lite
Doc change
Date in Changes file
Add array prototype to shuffle
Don't use XS objectify with 5.8 or earlier. Call PP directly.
Clean up todo
Increase const_int range
Move Chacha static context up one level (to csprng)
Move CSPRNG to per-thread context and remove locks
Make sure we do a good seed on boot and every thread
We auto-seed in XS, but not PP. Put back explicit seed on module load for PP.
Add correct argument for 32-bit use
Compiler warnings
Doc updates for per-thread changes
Document secure configuration option
Add is_square function
Validate n in is_power, includes necessary string->bigint conversion
lastfor in XS
Add lastfor to PP, tests for lastfor
Work around Perl internal error
Docs and simple formultiperm test
Add another lastfor test
doc updates
Add is_polygonal
Bump version
Compiler / XS warnings
Date in Changes
Add factorialmod
randperm with k << n faster and less memory.
Comment change
For prime partitions, pull min/max values into primes.
factorialmod can use by-primes method for faster results
Faster factorialmod for composite m and big n
Tests for factorialmod
Optimize derangement code
Use segmented sieve in factorialmod, 50% faster
numtoperm stays in XS for large n
Add POPs command to remove dual return values for unknown bigint classes
Speedup for string popcount
Allow Math::GMPq and Math::AnyNum as bigint classes
Factor speedups
Small improvement to addmod
Use our own multi-base Montgomery Miller-Rabin
Remove second perfect square pre-test on x86-64
factorialmod with montmath used more often
Updated factoring code: add Lehman, update SQUFOF, update factor() recipe
Remove rax constraint
Slightly faster HOLF, add HOLF and Lehman to factor loop.
Tests, spelling, exports, add lehman_factor, add factoring harness tests
Updated mont_inverse
Update factor recipe for faster speed on random inputs rather than semiprimes
Split out 63-bit and 64-bit mulredc
Update pbrent to use new mulredcs for better speed
Add and use new loop start/stop functions from PP
Fix formultiperm and lastfor
Don't force memory access for mulmod asm. Faster and smaller.
Add is_fundamental
Add Pari's ispowerful as an example
forcomb {} n; iterates over all k 0 .. n. Power set of n. Matches Pari/GP and Perl6.
Bump version number
exp_mangoldt pure perl make sure r is not a bigint
pminus1_factor and ecm_factor were mismatched after adding lehman_factor
Work around Perl 5.6.2 in stirling
Work around 5.6.2 brokenness
Silence compiler warning
Bump Changes for release
gregor herrmann (1):
New upstream version 0.68
-----------------------------------------------------------------------
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-perl.git
More information about the Pkg-perl-cvs-commits
mailing list