Changes commited to PARI GIT during the last two weeks

fix Mod(0,1)*x^0 == x^0

2	1	src/basemath/gen2.c
2	0	src/test/32/equal
2	1	src/test/in/equal

Pol(O(17^-1),v0)/Pol([Mod(9,23),O(23^-3)],v0) produced an illegal t_RFRAC

3	1	src/basemath/gen1.c
4	1	src/test/32/err
1	0	src/test/in/err

obscure gequalm1(t_RFRAC with t_QFB coeff) bug [#2393]

30	19	src/basemath/gen2.c
2	1	src/test/32/equal
2	0	src/test/in/equal

obscure bug t_SER z == 1 but z - 1 != 0

z=(Mod(2,17)*v1^-2+O(v1^-1))/(Mod(5,12)*v1^-4+O(v1))
sign(z) was incorrectly set to 1

1	0	src/basemath/gen1.c
3	1	src/test/32/ser
3	1	src/test/in/ser

hyperellchangecurve: add test [fix SEGC] + improve doc

1	0	doc/refcard-ell.tex
4	4	src/basemath/hyperell.c
8	1	src/functions/elliptic_curves/hyperellchangecurve
7	1	src/test/32/hyperelldisc
2	0	src/test/in/hyperelldisc

add / fix sanity checks in hyperellred + improve doc

1	0	doc/refcard-ell.tex
15	17	src/basemath/hyperell.c
20	15	src/functions/elliptic_curves/hyperellred
13	1	src/test/32/hyperelldisc
6	0	src/test/in/hyperelldisc

RgM_Rg_[add|sub] -> shallow equivalent

1	1	src/basemath/buch2.c
1	1	src/basemath/modsym.c
1	1	src/modules/thue.c

stack corruption in RgM_Rg_add and RgM_Rg_sub

Symptom : gerepile error in
\y
default(breakloop,0)
install("gequal1","iG");
install("gequalm1","iG");
install("gequal0","iG");
tstgenall(z)=
{
  if((z==1)!=(iferr(z-1,E,1,errname(E)!="e_BUG")==0),error(z));
  if((z==-1)!=(iferr(z+1,E,1,errname(E)!="e_BUG")==0),error(z));
  if(gequal0(z)!=(iferr(z,E,1,errname(E)!="e_BUG")==0),error(z));
  if(gequal1(z)!=(iferr(z-1,E,1,errname(E)!="e_BUG")==0),error(z));
  if(gequalm1(z)!=(iferr(z+1,E,1,errname(E)!="e_BUG")==0),error(z));
}
V=[Mod(0*v0,Pi*I*v0^0),Mat([Mat(Mod(0,1))])];
for(i=1,2,tstgenall(V[i]))

The extra cgetg would be inspected from higher gerepileupto code and
crash because its components have not been initialized

2	4	src/basemath/RgV.c

add modsym tests [#2392]

3	2	src/basemath/modsym.c
19	1	src/test/32/modsym
6	0	src/test/in/modsym

quadhilbert/quadray are in stark.c, not buch1.c (James Rickards) [#2391]

2	2	src/headers/paridecl.h

print( Mat([Mat(2)]) ) -> Mat([Mat(2)])

instead of Mat(Mat(2)) [which yields Mat(2)]

2	1	src/language/es.c
5	1	src/test/32/mat
2	0	src/test/in/mat

mscuspidal: the Tp matrix is rational, make it integral first

4	2	src/basemath/modsym.c

simplify test

7	10	src/test/in/ideal

handle correctly idealchineseinit() structures when RHS has denominators

fix doc: only works when denominators are coprime to modulus in the
'init' case

21	9	src/basemath/base3.c
2	1	src/functions/number_fields/idealchinese
1	1	src/modules/algebras.c
3	1	src/test/32/ideal
11	0	src/test/in/ideal

Use stack_sprintf instead of stack_malloc + sprintf

Not worth the trouble of estimating the string length

1	3	src/basemath/nflist.c
1	2	src/basemath/nflistQT.c
1	2	src/modules/elldata.c
2	6	src/modules/galois.c
1	2	src/modules/galpol.c

div_ser: fix scalarser usage (forbid prec < 0)

2	1	src/basemath/gen1.c

serconvol(O(x^-2),O(23)*O(x^-2)) -> crash

4	1	src/basemath/bibli2.c
2	1	src/test/32/ser
1	0	src/test/in/ser

 O(v3^-1)/(1+O(23)+I) -> crash

8	4	src/basemath/gen1.c
2	0	src/test/32/ser
2	0	src/test/in/ser

gcharnewprec_i: gcopy -> shallowcopy

2	1	src/basemath/grossenchar.c

ratpoints: get_2adic_info: make sure num_bits is initialized

5	2	src/modules/ratpoints.c

gchar_identify_i: avoid useless gcopy

18	18	src/basemath/grossenchar.c

test gcharidentify

3	1	src/test/32/gchar
4	0	src/test/in/gchar

fix gcharidentify, do not change input

1	1	src/basemath/grossenchar.c

pop_lex: fix memory leak

1	2	CHANGES
1	1	src/language/eval.c

Fincke_Pohst_ideal: move test for maxtry_ELEMENT in the (!fl) branch

if fl != 0, we are "skipfirst-ing" and not trying an element

1	1	src/basemath/buch2.c

fix comment

1	1	src/test/in/bnf

make ldata_eulerf static to lfunutils.c

if a private "fast" variant is needed later, we shall export it again
(and fix the interface; taking an 'ldata' would be ok, an ldata_get_an()
is weird for an exported function with that name)

1	0	src/basemath/lfunutils.c
0	1	src/headers/paridecl.h

off-by-1 error in vecsmall2vec2 [#2390]

4	3	src/basemath/pclgp.c
2	1	src/test/32/subcyclopclgp
1	0	src/test/in/subcyclopclgp

parforeach: fix memory leak

1	0	src/language/eval.c

whitespace edits

10	11	src/headers/pariinl.h

pclgp.c: replace Flx_recip1  by Flx_recip1_inplace

Reduce memory use by two for this example
subcycloiwasawa(x^2-29127,2,19)

45	10	src/basemath/pclgp.c

algebras [from James Rickards]: fix inefficiency in backtrackfacto

11	1	src/modules/algebras.c
2	1	src/test/32/alghasse
4	1	src/test/in/alghasse

document rfrac_to_ser_i

1	0	CHANGES
4	0	doc/usersch5.tex

pari.f90: add example of garbage collecting

25	6	examples/pari.f90

New libpari function get_avma

Useful for languages that does not provide access to TLS variables

3	0	doc/usersch5.tex
1	0	src/headers/paridecl.h
2	0	src/kernel/none/level1.h

cleanup mfgaloisprojrep / search_A5

15	17	src/basemath/mftrace.c
2	0	src/basemath/nflist.c

kill warning

1	1	src/basemath/nflist.c

improve doc

3	2	src/functions/elliptic_curves/ellL1
4	2	src/functions/elliptic_curves/ellanalyticrank