Changes commited to PARI GIT during the last two weeks

remove obsolete TODO entry

0	3	TODO

ZpXQX_liftroot_vald: ZX_Z_divexact: test if u t_INT first

2	1	src/basemath/Hensel.c

159- padicappr('x,Mod(0,'t^2+1)+O(7^10)) -> SEGV [#2036]

1	0	CHANGES
3	3	src/basemath/Hensel.c
1	0	src/test/32/padic
1	0	src/test/in/padic

tune: Add dichotomic search

61	24	src/test/tune.c

cleanups ellsympow + whitespace edits

26	20	src/basemath/lfunutils.c

82- factorback([;]) is now invalid

* [;] is not the empty factorization, factor(1) = matrix(0,2) is
* better type checking for extended ideals [ avoid various SEGVs ]

2	0	CHANGES
31	25	src/basemath/base4.c
1	1	src/basemath/bnflog.c
5	4	src/basemath/buch2.c
1	1	src/basemath/buch3.c
2	2	src/basemath/buch4.c
6	6	src/basemath/kummer.c
0	1	src/basemath/polarit2.c
3	1	src/functions/number_fields/HEADER
5	1	src/test/32/ideal
4	0	src/test/in/ideal

update TODO

2	32	TODO

Rename SQRR_MULII_LIMIT to SQRR_SQRI_LIMIT

1	1	src/desc/deftune
2	2	src/headers/paritune.h
2	2	src/kernel/gmp/tune.h
1	1	src/kernel/none/mp_indep.c
1	1	src/kernel/none/tune-gen.h
2	2	src/kernel/none/tune.h
1	1	src/test/tune.c

Add new tuning parameter SQRR_MULII_LIMIT

2	1	src/desc/deftune
2	0	src/headers/paritune.h
4	2	src/kernel/gmp/tune.h
1	1	src/kernel/none/mp_indep.c
1	0	src/kernel/none/tune-gen.h
4	2	src/kernel/none/tune.h
7	3	src/test/tune.c

cleanup ellsympow_isabelian3

2	4	src/basemath/lfunutils.c

simplify ellminimaltwistcond + whitespace edit

7	7	src/basemath/elliptic.c

158- ellminimaltwist did not check that E is defined over Q

=> corruption

1	0	CHANGES
1	0	src/basemath/elliptic.c

fix clone leak in ellsympow + don't compute minimal twist if useless

refactor with lfunellmfpeters

20	12	src/basemath/lfunutils.c

ellsympow_goodred: remove unused argument

2	2	src/basemath/lfunutils.c

whitespace edits + remove impossible pari_err_BUG occurence

8	12	src/basemath/polmodular.c

add uu32toineg + replace setsigne(,-1) by call to utoineg or uu32toineg

5	0	doc/usersch5.tex
10	0	src/basemath/gen3.c
228	246	src/basemath/polmodular.c
1	0	src/headers/paridecl.h

lfun: derivative a a pole of gamma factor; avoid division by 0

6	3	src/basemath/lfun.c
2	1	src/test/32/lfun
1	0	src/test/in/lfun

Replace t_LFUN_ELLSYMSQ by t_LFUN_ELLSYMPOW

22	62	src/basemath/lfunutils.c
2	2	src/headers/paripriv.h

New GP function lfunellsympow

1	0	CHANGES
349	0	src/basemath/lfunutils.c
10	0	src/functions/l_functions/lfunsympow
1	0	src/headers/paridecl.h
1	1	src/headers/paripriv.h
8	1	src/test/32/lfun
25	1	src/test/32/lfuntype
8	0	src/test/in/lfun
3	1	src/test/in/lfuntype

etaquotype: handle arbitrary factorization matrices

don't assume t_VECSMALL compatibility

23	25	src/basemath/lfunutils.c
3	2	src/basemath/mftrace.c
1	1	src/headers/paridecl.h

fix ratpoints bench (missing labs)

2	2	src/modules/ratpoints.c

kill spurious warning

1	1	src/modules/genus2red.c

trivialize theta_j

3	13	src/modules/genus2red.c

typo in derivnum: gerepileupto -> gerepilecopy

1	1	src/language/sumiter.c

cleanup labelm3 & quadratic

11	14	src/modules/genus2red.c

remove R from struct igusa_p

8	9	src/modules/genus2red.c

cleanup polmini_zi*

11	22	src/modules/genus2red.c

cleanup tame_234_init

20	26	src/modules/genus2red.c

tame*: convert rational comparisons to integral

11	16	src/modules/genus2red.c

cleanup tame_567_init + equalis -> equaliu

36	45	src/modules/genus2red.c

genus2red: replace (rational) \theta by integral multiple

113	132	src/modules/genus2red.c

cleanup Ip initialization (move val[eps2] to val[8])

30	33	src/modules/genus2red.c

genus2red cleanups

- move Ip initalization to stable_reduction()
- cleanup tame() initialization

43	37	src/modules/genus2red.c

cleanup tame()

15	16	src/modules/genus2red.c

cleanup tame()

24	26	src/modules/genus2red.c

genus2red: refactor rational arithmetic in tame*

87	114	src/modules/genus2red.c

[libpari] ulcm

1	1	CHANGES
3	0	doc/usersch5.tex
2	2	src/basemath/base2.c
1	1	src/basemath/galconj.c
10	10	src/basemath/mftrace.c
2	3	src/basemath/perm.c
1	1	src/basemath/polclass.c
1	1	src/basemath/trans3.c
1	0	src/headers/paridecl.h
5	3	src/kernel/none/gcdll.c
4	6	src/modules/algebras.c
1	1	src/modules/subfield.c

cgcd -> ugcd

1	1	src/basemath/ZV.c
2	2	src/basemath/base1.c
2	2	src/basemath/ellanal.c
2	2	src/basemath/gen1.c
1	1	src/basemath/modsym.c
1	1	src/basemath/nffactor.c
1	1	src/basemath/polarit3.c
1	1	src/basemath/polclass.c
1	1	src/basemath/polmodular.c
22	21	src/modules/algebras.c
3	3	src/modules/ratpoints.c

rewrite sstoQ

11	5	src/headers/pariinl.h

reword comment + cgcd -> ugcd

6	12	src/basemath/galconj.c

cgcd -> ugcd + rename variable

4	5	src/basemath/FpXQX_factor.c

cgcd -> ugcd

36	42	src/basemath/mftrace.c

replace cgcd implementation by call to ugcd

2	11	src/kernel/none/gcdll.c

[libpari] ugcdiu, ugcdui

1	1	CHANGES
4	0	doc/usersch5.tex
1	1	src/basemath/buch1.c
1	1	src/basemath/buch4.c
4	8	src/basemath/ellanal.c
1	1	src/basemath/gen3.c
0	3	src/basemath/lfunutils.c
2	2	src/basemath/mftrace.c
5	9	src/basemath/polarit3.c
2	0	src/headers/paridecl.h
5	0	src/kernel/none/level1.h
1	1	src/modules/algebras.c
2	2	src/modules/krasner.c

allow mffrometaquo([24,1]) instead of mffrometaquo(Mat([24,1]))

same for lfunetaquo

7	1	src/basemath/lfunutils.c
4	1	src/functions/l_functions/lfunetaquo
5	1	src/functions/modular_forms/mffrometaquo
11	11	src/test/in/mf

rewrite etaquotype, unify with eta_NK/eta_holomorphic

- fix bad formulas [etacuspidal corrects the version in lfunetaquo]
- fix overflow potential
currently restricted to t_VECSMALL-compatible factorization matrix, could
support arbitrary factorization matrix. Worth it ?

95	64	src/basemath/lfunutils.c
16	67	src/basemath/mftrace.c
1	0	src/headers/paridecl.h
8	1	src/test/32/mf
7	0	src/test/in/mf

use addmulii

5	5	src/basemath/hnf_snf.c

hnflll: refactor swap decision

15	21	src/basemath/hnf_snf.c
1	1	src/basemath/mftrace.c

hnflll: remove HNFLLL_QUALITY-protected code

Dead code since 1998: never made any positive difference to enable it, with
alpha = 1/2, 3/4 or 1 (default). Simplify code.

2	16	src/basemath/hnf_snf.c

157- matkerint could be exponentially slow; guarantee polynomial time [#2034]

Drawback: slows down some classes of matrices, e.g.
  matkerint(matrix(50,55,i,j,random(10^5)))
time increases from 28ms to 1s. OTOH the matrix in #2034 goes down from
2 minutes to 4ms.

2	0	CHANGES
2	1	src/basemath/lll.c
8	8	src/test/32/linear

make bench identical on 32/64 bit archs

3	3	src/test/32/gamma
1	1	src/test/in/gamma

155- affrr(0.0, x) did not reset the exponent correctly

use min(expo(0.0), bit_accuracy(lg(x)))

Fixes: psi((3/2+0.E-38*I)+x/2+O(x^2)) -> diverges

2	0	CHANGES
7	5	src/kernel/none/mp_indep.c
11	11	src/test/32/bessel
3	1	src/test/32/gamma
4	4	src/test/32/polylog
1	0	src/test/in/gamma

150- gamma(0.0+1/2*x+O(x^2)) -> domain error in factorial

1	0	CHANGES
29	24	src/basemath/trans2.c
3	1	src/test/32/gamma
1	0	src/test/in/gamma

cleanup lfunetaquo / etaquotype / vecan_eta

44	64	src/basemath/lfunutils.c

173- [libpari] hclassno6, hclassno6u + update / fix docs

1	0	CHANGES
31	1	doc/usersch6.tex
0	1	src/functions/modular_forms/getcache
10	18	src/functions/number_theoretical/qfbclassno

fix idealchinese doc

1	1	src/functions/number_fields/idealchinese

qfbhclassno: fix doc

14	4	src/functions/number_theoretical/qfbhclassno

use usqrt

3	3	src/basemath/polmodular.c

cleanup modpoly_pickD_primes

10	18	src/basemath/polmodular.c

clone leak in consttabh

4	3	src/basemath/mftrace.c

polmodular_db_add_level: better fix

in case the obtained x is db[L] [impossible in current code, but...]

4	2	src/basemath/polmodular.c

81- memory leak in polclass() [ see also two preceding commits ]

1	0	CHANGES
6	4	src/basemath/polmodular.c

calc_primes_for_discriminants: replace 2 newblocks by a single one

12	15	src/basemath/polmodular.c

fix one clone leak in polmodular [2 remaining]

4	3	src/basemath/polmodular.c

polmodular_db_add_level: remove hack assuming INV_J == 0

1	2	src/basemath/polmodular.c

[ecpp] cleanup D_polclass + polmodular_db_init_allinv + fix clone leak

9	14	src/basemath/ecpp.c

alias0: use cgetg_block

1	2	src/language/anal.c

149- implementation of newblock was incompatible with getheap(). Now we
     always include the 'block' size in the header.

2	0	CHANGES
2	1	src/headers/paripriv.h
4	3	src/language/init.c
1	1	src/test/in/analyz
1	1	src/test/in/elliptic
1	1	src/test/in/graph
1	1	src/test/in/linear
1	1	src/test/in/nfields
1	1	src/test/in/number
1	1	src/test/in/polyser
1	1	src/test/in/program
1	1	src/test/in/sumiter
1	1	src/test/in/trans

whitespace edit

0	1	src/basemath/trans3.c

fix mspetersson in weight 1/2

3	2	src/basemath/mftrace.c

mfinit(weight 1/2 with joker) treated incorrectly

1	1	src/basemath/mftrace.c

cleanup QabV_tracerel

4	5	src/basemath/mftrace.c

ecpp: cleanup find_EP [remove unused argument, save a couple copies]

8	16	src/basemath/ecpp.c

172- [libpari] Fp_center_i, FpX_center_i

1	0	CHANGES
7	0	doc/usersch5.tex
3	3	src/basemath/FlxqE.c
8	0	src/basemath/FpX.c
1	1	src/basemath/QX_factor.c
1	1	src/basemath/aprcl.c
3	3	src/basemath/base2.c
3	3	src/basemath/ecpp.c
1	1	src/basemath/ellpadic.c
2	3	src/basemath/ellsea.c
5	5	src/basemath/galconj.c
1	1	src/basemath/gen3.c
1	1	src/basemath/hyperell.c
5	5	src/basemath/nffactor.c
1	1	src/basemath/polarit3.c
2	0	src/headers/paridecl.h
4	0	src/headers/pariinl.h

mfeval: wrong result when \alpha != 0

2	2	src/basemath/mftrace.c
3	2	src/test/32/mf
2	1	src/test/in/mf

fix possible overflow in vecpowug / vecpowuu

2	2	src/basemath/lfun.c

171- [libpari] umuluu_le

1	0	CHANGES
4	0	doc/usersch5.tex
2	2	src/basemath/dirichlet.c
3	7	src/basemath/mftrace.c
1	0	src/headers/paridecl.h
9	0	src/kernel/none/level1.h
5	14	src/modules/stark.c

remove polsymmodp, use FpX_Newton

8	37	src/basemath/base2.c
0	1	src/headers/paripriv.h

document RgXn_inv_i

3	0	doc/usersch5.tex

RgXn_inv_i: without GC

11	8	src/basemath/RgX.c
1	1	src/basemath/elliptic.c
1	1	src/basemath/gen3.c
3	2	src/basemath/lfunutils.c
1	0	src/headers/paridecl.h

148- missing GC in lfuntheta

1	0	CHANGES
3	1	src/basemath/lfun.c

rename ANSI_COLOR_* -> ANSI_*

36	36	src/basemath/ecpp.c

fix compilation

1	1	src/basemath/ecpp.c

fix nf_get_zk to treat old format correctly

2	1	src/headers/pariinl.h
1	0	src/test/32/nf
1	0	src/test/in/nf

cleanup N_downrun

14	18	src/basemath/ecpp.c

cleanup N_downrun

20	22	src/basemath/ecpp.c

change prefix NDinfomqg -> NDmqg

16	16	src/basemath/ecpp.c

step1: don't modify Dinfo in place, pass it on to step2

26	26	src/basemath/ecpp.c

ecpp: cleanup tunevec use

17	25	src/basemath/ecpp.c

D_find_discsqrt: s can't be NULL at the end

1	1	src/basemath/ecpp.c

remove counters A1, A6, D2

4	19	src/basemath/ecpp.c

whitespace edit + extra else for clarity

3	3	src/basemath/ecpp.c

cleanup D_find_discsqrt

15	27	src/basemath/ecpp.c

fix typo in diagnostic (D is now a long)

1	1	src/basemath/ecpp.c

NDinfomqg_get_D now returns a long (old NDinfomqg_get_longD, now deleted)

19	28	src/basemath/ecpp.c

cleanup NDinfomqgJ_find_EP

30	29	src/basemath/ecpp.c

refactor NDinfomqgJ_find_EP

30	43	src/basemath/ecpp.c

isprime: lower threshold between APRCL / ECPP

1	1	src/basemath/prime.c

cleanup FpX_classtower_oneroot + fix typ degpol(P)> 1 [instead of>= 1]

16	23	src/basemath/ecpp.c

FpX_classtower_oneroot: use tower factorization

This is faster when the coefficients are large

23	13	src/basemath/ecpp.c

maxsqrt^2 -> maxdisc

1	1	src/basemath/ecpp.c

fix formatting in idealred documentation

2	2	src/functions/number_fields/idealred

kill warnings

6	4	src/basemath/ecpp.c

unused variable

0	1	src/basemath/ecpp.c

N_downrun_NDinfomq: handle return value of gen_0

Normally impossible when not at toplevel, but costs nothing. And more
robust against tuning changes

3	3	src/basemath/ecpp.c

N_downrun_NDinfomq: remove useless recursive GC

replace by global GC once the function has returned

4	3	src/basemath/ecpp.c

Dmqvec_slice is such that expiN> expi(q) => remove dead code

2	7	src/basemath/ecpp.c

add missing macro ecpp_param_get_tune + fix comment

8	6	src/basemath/ecpp.c

cleanup ecpp_param_set

remove obscure macros with non-trivial dependencies, create 'param' in
one shot.  Move primelist one level up [was enclosed in a t_VEC]

9	36	src/basemath/ecpp.c

rewrite ecpp_disclist_init to save on memory (rough factor 2)

60	48	src/basemath/ecpp.c
2	2	src/test/32/isprime
1	1	src/test/in/isprime

ecpp_disclist_init: reduce memory use

22	32	src/basemath/ecpp.c
2	2	src/test/32/isprime
1	1	src/test/in/isprime

ecpp_disclist_init: remove useless maxsqrt argument

3	4	src/basemath/ecpp.c

ecpp_disclist_init: don't recompute primelist, no need for indexlist

10	46	src/basemath/ecpp.c

ecpp: further cleanups

11	27	src/basemath/ecpp.c

ecpp: cleanup batch factorization

28	74	src/basemath/ecpp.c

cleanup sort_disclist

3	5	src/basemath/ecpp.c

ecpp: cleanup timer_record

3	11	src/basemath/ecpp.c

cleanup comments

10	22	src/basemath/ecpp.c

cleanup ecpp_primelist_init

23	42	src/basemath/ecpp.c