Changes commited to PARI GIT during the last two weeks

cleanup gen_ellgroup

55	51	src/basemath/bb_group.c

improve doc

3	1	doc/usersch5.tex

improve doc

1	1	doc/usersch5.tex

gen_ellgroup: faster algorithm (From Martin Tot Bagi)

27	6	src/basemath/bb_group.c

gen_order: use recursive algorithm by Márton Tot Bagi

1	0	AUTHORS
1	0	CHANGES
34	26	src/basemath/bb_group.c

22- psi(x close to a nonpositive integer) -> error

1	0	CHANGES
6	3	src/basemath/lerch.c
6	1	src/test/32/gamma
3	0	src/test/in/gamma

move QXQ_charpoly [#2591] fix to its proper place

QXQ_charpoly arguments are documented as QX + ZX and this is easy
and cheaper to ensure in the caller.

Move fix to RgXQ_charpoly_fast with slight inefficency (extra
gerepileupto) in the silly case where the T argument was *not* a ZX.
Cheaper than a (systematic!) Q_primpart in any case, and a priori never
exercised if the caller of RgXQ_charpoly is sane.

I would go so far as calling RgXQ_charpoly with a non-ZX argument T as a
bug if it occurs through a libpari call (OK through a random call to
GP's charpoly: we can't prevent users from being inefficient).

9	1	src/basemath/polarit2.c
1	1	src/basemath/polarit3.c

QXQ_minpoly: Handle QX denominator too. [#2591]

1	0	CHANGES
1	1	src/basemath/polarit3.c
9	0	src/test/in/charpoly

rename Sunits -> SUnits

9	9	src/basemath/buch2.c

HC  19- mfsymboleval could return wrong sign [#2586]

2	1	CHANGES

CHANGES: fix numbering

2	2	CHANGES

15- control-C could lead to segfault later [#2496]

5	4	CHANGES