Findafactor

Exactly in the (elusive) case of finding a solution in Gaussian elimination, there was a segmentation fault. This has been fixed.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.1.3...v5.1.4

sha1sum results:
77e275817b0d9f4b17c5c62a69ecc8db0a24f94d FindAFactor-5.1.4-cp310-cp310-manylinux_2_35_x86_64.whl
5ea50f27f1238163e452cef8eaa2c6ce3d993cfd FindAFactor-5.1.4-cp312-cp312-manylinux_2_39_x86_64.whl
d2036bb3ff53e660d5f9525adb0cf25895005a59 FindAFactor-5.1.4-cp312-cp312-win_amd64.whl
49ab90190cbb8c4fe56c57e5175e435a64d78dd5 FindAFactor-5.1.4-cp313-cp313-macosx_13_0_x86_64.whl
05871a08d3b9eb9fddd6244f2a9095d649c64d35 FindAFactor-5.1.4-cp313-cp313-macosx_14_0_arm64.whl
3f06791bdfe7166cb2266db1f52fba8baa385fd6 FindAFactor-5.1.4-cp313-cp313-macosx_15_0_arm64.whl
fc27dfa8fc38e0290546d2abcd05d916adb9b419 FindAFactor-5.1.4-cp38-cp38-manylinux_2_31_x86_64.whl
ee7e9f26d4c081028a9852b990e7a643d97fab52 findafactor-5.1.4.tar.gz

Gaussian elimination has been debugged to (hopefully) correspond with the form here: https://github.com/NachiketUN/Quadratic-Sieve-Algorithm/blob/7462d39e29a8c95dc61d2cccab2944866c5441f2/src/main.py#L199

An early exit condition has also been added: if the sieve directly returns a perfect square residue, check immediately if it is a congruence of squares. If so, we will find a solution in the sieving phase.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.1.2...v5.1.3

sha1sum results:
5f653b54147e8eceaf0580baeae089a195472f9d FindAFactor-5.1.3-cp310-cp310-manylinux_2_35_x86_64.whl
f23dd8d846132d07880eb99d4304688d381ea243 FindAFactor-5.1.3-cp312-cp312-manylinux_2_39_x86_64.whl
5448ac6446f85dbdefa4c115fb1ff2c400cd9f3e FindAFactor-5.1.3-cp312-cp312-win_amd64.whl
d8bcfde9052dfaf60e0e11a08e824f45b7a5920c FindAFactor-5.1.3-cp313-cp313-macosx_13_0_x86_64.whl
fc81be709574c024fcd3703fab281a23198b0889 FindAFactor-5.1.3-cp313-cp313-macosx_14_0_arm64.whl
65e33330467c94b34cc256b3c494b06072ae6081 FindAFactor-5.1.3-cp313-cp313-macosx_15_0_arm64.whl
690440ec307e98ed393ad9c471afe8999ed2b684 FindAFactor-5.1.3-cp38-cp38-manylinux_2_31_x86_64.whl
c0505b8801bfe533f0afbeb1c0895c3975587dbb findafactor-5.1.3.tar.gz

The developer passed over the code to refactor for human-readability and performance.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.1.1...v5.1.2

sha1sum results:
2294c308dcaf2b506f73bf6a90b28c901a91010f FindAFactor-5.1.2-cp310-cp310-manylinux_2_35_x86_64.whl
936e9a97603e46f8b90641eb988e7782aa45d872 FindAFactor-5.1.2-cp312-cp312-manylinux_2_39_x86_64.whl
6692e887e60702f299c586ebd7d0b78fa084cc12 FindAFactor-5.1.2-cp312-cp312-win_amd64.whl
f63a7953a5a0eb3a4a9c7bf86c397c13596824ad FindAFactor-5.1.2-cp313-cp313-macosx_13_0_x86_64.whl
f5b3d9f1e8576208740ad6c4a4ca3e76d4dd79a9 FindAFactor-5.1.2-cp313-cp313-macosx_14_0_arm64.whl
997e3735a956672ef4b70e50e706adb5f3b40ccc FindAFactor-5.1.2-cp313-cp313-macosx_15_0_arm64.whl
3eb094336bed45ef0fabdc268f9fc85180767d05 FindAFactor-5.1.2-cp38-cp38-manylinux_2_31_x86_64.whl
29884195d03b45ff07d2a42c818358a396096ee1 findafactor-5.1.2.tar.gz

This release debugs the Gaussian elimination algorithm. (`FACTOR_FINDER` mode might be considered experimental, so far, but we're at least extremely close to stability and maybe "completion.")

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.1.0...v5.1.1

sha1sum results:
c474b70fe4908376f7762d594694656b4027899a FindAFactor-5.1.1-cp310-cp310-manylinux_2_35_x86_64.whl
66dc4134e73db3e7b9df603d20c68c7556441df4 FindAFactor-5.1.1-cp312-cp312-manylinux_2_39_x86_64.whl
803bcecd847495db158a58697258b45e8efd4244 FindAFactor-5.1.1-cp312-cp312-win_amd64.whl
6f1016e12e02b5d008e684342ae659610c6c3456 FindAFactor-5.1.1-cp313-cp313-macosx_13_0_x86_64.whl
b49da52d751d58808cc5e19b316a28dd4eb494d8 FindAFactor-5.1.1-cp313-cp313-macosx_14_0_arm64.whl
958f6b7898207c010a598464d1fdc875efc0ed60 FindAFactor-5.1.1-cp313-cp313-macosx_15_0_arm64.whl
99c272403ca7941fc09e8f7197fcc0613ed61e5c FindAFactor-5.1.1-cp38-cp38-manylinux_2_31_x86_64.whl
3d6867b07898c994f39f80d2e320ea835bcf884c findafactor-5.1.1.tar.gz

A bug has been fixed that prevented finding smooth numbers as intended. Also, I promised well-tuned defaults; however, the best settings very often depend on the exact number being factored. By design, the smoothness bound, sieving bound, and smooth number limit cutoff were all scaled such that accepted or defensible heuristics should be the default at a value of `1.0`: these **are** the **most-general** best defaults. We **could** tune for a specific scale, like 64-bit semiprimes, but why is the software specifically being used for 64 bits? I am still carefully considering default tuning, but maybe the best we can do **in general** reflects the fact that simple defaults like `1.0` multipliers **already** match best general advice, but **expect to tune in any specific case on basis of your knowledge of the theory**.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.0.6...v5.1.0

sha1sum results:
bad5d7a6b3d8949a404e3a191da7d45f72154276 FindAFactor-5.1.0-cp310-cp310-manylinux_2_35_x86_64.whl
c2295de00d285a87434b9e50b032337680d9b4de FindAFactor-5.1.0-cp312-cp312-manylinux_2_39_x86_64.whl
eb82a928e6671a82699e79adfdd06814463afb6d FindAFactor-5.1.0-cp312-cp312-win_amd64.whl
4c4ef7cfe4ecb39f63462c53a6a981aa353e18db FindAFactor-5.1.0-cp313-cp313-macosx_13_0_x86_64.whl
cc79fb2817600bc1811513553c9ef29b173a877d FindAFactor-5.1.0-cp313-cp313-macosx_14_0_arm64.whl
a7cd10002f4368b05b08e7a9e040f1e4bd61cedd FindAFactor-5.1.0-cp313-cp313-macosx_15_0_arm64.whl
2a53c382143b6d9b4c2092bc95d161e0fbb06c4f FindAFactor-5.1.0-cp38-cp38-manylinux_2_31_x86_64.whl
edbd0dfb682548edbbedc8bcc0f96e300a9e4591 findafactor-5.1.0.tar.gz

The previous releases relied on finding exactly two rows with identical factorization parity matrices for their residue `% N`, to form a perfect square residue; instead, we now allow solutions formed from an arbitrary number of rows.

Also, there is no need to worry about removing the "coercion" into perfect squares, in the sieve: **after** we find a perfect-square **residue**, if the left-hand side of the congruence of squares equation is not a perfect square **already** (which it might be), at least, **nothing is hurt by squaring the final result**, which guarantees a perfect square on the left-hand side of the equation, as well as the right-hand side after the rows are solved for a perfect-square residue.

What's Changed
* Row solver by WrathfulSpatula in https://github.com/vm6502q/FindAFactor/pull/7


**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v5.0.5...v5.0.6


sha1sum results:
d9b5af1567243c75dfffae1be0b6debc33abe30f FindAFactor-5.0.6-cp310-cp310-manylinux_2_35_x86_64.whl
0c6e8c137e8fb086467233fc7935a137872cb8d6 FindAFactor-5.0.6-cp312-cp312-manylinux_2_39_x86_64.whl
a72f45ae687ddbb9e4d694f3caec53241f9411ba FindAFactor-5.0.6-cp312-cp312-win_amd64.whl
282b06a5e7eba7cc3db48ec093002cdbad17d452 FindAFactor-5.0.6-cp313-cp313-macosx_13_0_x86_64.whl
226f0ba195c4aecde3a3f1d88770cd71a706ce9b FindAFactor-5.0.6-cp313-cp313-macosx_14_0_arm64.whl
36b651a3056dcd39d37cc68606658ab60927147a FindAFactor-5.0.6-cp313-cp313-macosx_15_0_arm64.whl
1bc2179ca7d7a5cca05ef38d705e5cf505c25fa4 FindAFactor-5.0.6-cp38-cp38-manylinux_2_31_x86_64.whl
3276815970a6d566dd926c314039a44e9cab6474 findafactor-5.0.6.tar.gz