Changes commited to PARI GIT during the last two weeks

doc typos

2	2	src/functions/number_fields/idealprimedec

fix problem in ZC_nfvalrem introduced in b1932a22

we can't remove p-contents when newx is not NULL. We could modify the
function interface to set *newx = x * (tau/p)^v1 / p^v2 with
v_pr(x) = v1 + e(pr/p) * v2; this would be more efficient. Don't bother for now.

1	1	src/basemath/base3.c

paridecl.h: add missing entries in pariinl.h

12	0	src/headers/paridecl.h

test-rnfkummer, test-bnrclassfield: report error detail

1	3	src/test/32/bnrclassfield
5	3	src/test/in/bnrclassfield
5	2	src/test/in/rnfkummer

kill warning

1	1	src/basemath/qfsolve.c

serhypergeom: add GC

7	1	src/basemath/hypergeom.c

streamline serhypergeom

14	20	src/basemath/hypergeom.c

cleanup hypergeom + add tests

55	50	src/basemath/hypergeom.c
24	11	src/test/32/hypergeom
6	2	src/test/in/hypergeom

HC  13- allow hypergeom(N, D, t_SER)

1	0	CHANGES
76	8	src/basemath/hypergeom.c
18	4	src/test/32/hypergeom
5	1	src/test/in/hypergeom

add test

2	1	src/test/32/ideal
4	0	src/test/in/ideal

21- missing GC in nfeltval + better algorithm when valuation is huge

1	0	CHANGES
7	0	src/basemath/base3.c

qfrsolve_normform: search for forms in both direction

14	10	src/basemath/Qfb.c
54	57	src/test/32/qfbsolve
3	1	src/test/in/qfbsolve

qfeval(Qfb(1,0,1),1): fix error message

1	0	src/basemath/gen3.c

cleanup ellintegralbmodel (+ kill warning)

9	10	src/basemath/elliptic.c

isprincipalall: better test for "large" generators if nf_GENMAT

9	3	src/basemath/buch2.c

isprincipalall: if nf_GENMAT don't reconstruct from embeddings is result is huge

3	1	src/basemath/buch2.c
2	5	src/test/32/rnfkummer

improve gcomposev

17	4	src/basemath/elliptic.c

cleanup ellintegralbmodel

2	2	src/basemath/elliptic.c

New libpari functions ZC_divexactu, ZM_divexactu

2	0	CHANGES
1	0	COMPAT
4	1	doc/usersch5.tex
5	1	src/basemath/ZV.c
2	1	src/headers/paridecl.h

New libpari function RgXQV_factorback

1	1	CHANGES
4	0	doc/usersch5.tex
6	0	src/basemath/RgX.c
1	0	src/headers/paridecl.h

ellintegralbmodel: check input

8	4	src/basemath/elliptic.c

further qfminimize cleanups / optimizations

8	8	src/basemath/qfsolve.c

improve critical loop in qfminimize

avoid many useless operations of type * p / p (operate on matrix parts)

33	19	src/basemath/qfsolve.c

cleanup qfminimize

44	55	src/basemath/qfsolve.c

ZX_mul: change back tuning to using mulii most of the time

it's generically faster, even when degrees differ a lot, at least for
polynomials with balanced coefficients. E.g.,

pol(d, b) = my(v = vector(d+1)); v[1] = [-2^b,2^b]; random(Pol(v));

Much slower for sparse polynomials but it can't be helped :-(

22	4	src/basemath/ZX.c

ellQ_factorback: take expected height as input.

52	11	src/basemath/elliptic.c

qfminimize: Use fast multiplication algorithm

81	32	src/basemath/qfsolve.c

qfminimize: clean up and add progress report

113	107	src/basemath/qfsolve.c

add sanity check in is_group()

6	2	src/basemath/galconj.c
4	1	src/test/32/galoischartable
5	0	src/test/in/galoischartable

config/DOC_Make.SH: set LC_ALL=C for sort

1	1	config/DOC_Make.SH

doc: alg_get_basis: fix typo albasis->algbasis

1	1	doc/usersch6.tex

Flj_changepointinv_pre: fix formula

4	3	src/basemath/Fle.c

FpE_ellweil/tatepairing: call Fle_ellweil/tatepairing

17	2	src/basemath/FpE.c

New libpari functions Fle_weilpairing, Fle_tatepairing

1	0	CHANGES
4	0	doc/usersch7.tex
167	0	src/basemath/Fle.c
2	0	src/headers/paridecl.h

kill warning

2	2	src/basemath/mftrace.c

cleanup mfdihedralcommon

19	19	src/basemath/mftrace.c

split mfdihedralall in 2 independent functions (N and [1,N])

28	31	src/basemath/mftrace.c

Bump year to 2021

2	2	README
1	1	doc/appa.tex
2	2	doc/parimacro.tex
1	1	doc/refmacro.tex
1	1	src/gp/gp.c

add diagnostics

9	2	src/basemath/mftrace.c

ZX_mul, ZXQXM_mul: more conservative tunings

3	3	src/basemath/ZX.c

split_ii: add early abort

3	2	src/basemath/mftrace.c

mfsplit: don't call polred if [K:Q]> 30 (was [K:Q(chi)])

1	1	src/basemath/mftrace.c

change ZX_mul and ZXQM_mul tunings

avoid interpolation when input have very different sizes

21	5	src/basemath/RgV.c
15	8	src/basemath/ZX.c
1	0	src/headers/paridecl.h

fix typo in ZM_mul tuning + missing timer initialization

1	1	src/basemath/ZV.c
1	0	src/basemath/mftrace.c

change ZM_mul tunings

- disable ZM_mul_fast when inputs have very different bitsizes
- check SW bound with MIN bitsizes (not max)

25	50	src/basemath/ZV.c

whitespace edits

21	41	src/basemath/ZV.c

ZM_mul: cleanup + whitespace edits

11	23	src/basemath/ZV.c

replace mf1basis by an even more heuristic version

- reduce Schaeffer use for composite levels (use known coefficients up
  to Sturm bound and apply Tp instead of using p * Sturm bound)
- fix the "Sturm bound" for prime levels
- save memory (avoid matsermul and treat each column individually with GC)
- ZabM_ker: replace final check by probabilistic test mod p
mf1basis: use Schaeffer again in all cases
- add diagnostics

3	0	CHANGES
24	9	src/basemath/alglin1.c
122	96	src/basemath/mftrace.c
7	4	src/test/32/mf
4	0	src/test/in/mf