Bug numbers refer to the BTS at http://pari.math.u-bordeaux.fr/Bugs/

 [last column crossreferences current development release 2.18.1]
Done for version 2.17.3 (released 24/11/2025):
  Fixed
    1- sumnumrat & friends: use full precision estimates to avoid failing with
        domain error when "close" to boundary                             [F17]
    2- zeta([3,2,1]) -> oo-loop                                           [F18]
BA  3- parforeach([1..1000],j,vector(10^5,i,i)) -> stack overflow         [F19]
    4- mfpetersson() incorrect when one form not cuspidal                 [F20]
    5- lift_to_frac: use of cgiv could corrupt PARI stack                 [F22]
    6- sumeulerrat / prodeulerrat fail in cases where they shouldn't [#2620]
       e.g., sumeulerrat(1/(p^2-5),1.1), prodeulerrat(1+1/(p^2-p-23),0.6) [F23]
    7- incgam(0,10^-50) -> GEN corruption                                 [F25]
BA  8- 12*harmonic(1) -> 12/1                                             [F26]
    9- hyperellratpoints(P, d) could miss points when P(-d) = 0           [F27]
   10- obscure bnrclassfield bug [internally creating an incorrect ideal
       factorization whose prime ideals were not sorted]                  [F28]
BA 11- qfeval(Qfb(2,3,4),[1,2;3,4]).disc was wrong [#2630]                [F29]
   12- bnfsunit: could fail to detect non-S-units                         [F30]
BA 13- lfunan: overflow error with large bad primes.                      [F32]
   14- mftonew: fix problem with mfbd forms [#2619]                       [F34]
   15- lfunthetainit: lost accuracy for large shifts.                     [F35]
HC 16- hypergeom([8,8.5],7,-0.4) was wrong [#2628]                        [F36]
HC 17- lfunmfspec(L) divided by zero if period vanishes                   [F37]
BA 18- fforder(,factor(1)) -> crash [#2618]                               [F38]
   19- mateigen must not call qfjacobi on non-real matrices [#2636]       [F39]
BA 20- ^C\n^C in parfor could crash GP                                    [F40]
   21- solve() iteration could incorrectly abort and return 0.0 [#2638]   [F41]
   22- zeta(1/2+10^22*I) -> precision too low                             [F42]
BA 23- lfuninit(lfunshift(1,3,1)) was incorrect [#2631]                   [F43]

Done for version 2.17.2 (released 01/03/2025):
  Fixed
BA   1- ellrootno over number fields could be slow [#2595]                 [F2]
BA   2- matsolvemod could fail [#2604]                                     [F3]
BA   3- nfcompositum(nfinit(y-2),x^2-y,x^3-x-1) -> error [#2605]           [F4]
     4- Mod(a,A) == Mod(b,B) was checking a == b && A == B instead         [F6]
        of the documented Mod(a,A) - Mod(b,B) == 0 [F6]
     5- normlp(, oo): support components equal to oo or -oo                [F8]
     6- sqrtn(27+O(3^4),3) -> 1 + O(3) (wrong valuation in Qp_sqrtn_ram)  [F10]
     7- vectorsmall(5, i, 2^63-1).normfu -> SEGV                          [F11]
AP   8- alginit(nf,[a,b]) would reject a in basis form                    [F13]
BA   9- for(i=1,10^6,bernpol(5,1)) -> "no more variables" [#2611]         [F14]
BA  10- Flx_mod_Xn1(T,n), deg T >= 2*n -> wrong result                    [F15]
    11- forprime(p=1099505335069-100000,1099505335069+10000,) -> crash    [F16]

 [last column crossreferences current development release 2.18.0]
Done for version 2.17.1 (released 24/12/2024):
  Fixed
BA   1- incorrect GP proto for ellpadicfrobenius, mftocoset, qfcvp [#2575] [F1]
AP   2- fix idealismaximal for ideal p*ZK [Sage issue 38766]               [F2]
AP   3- segfault in algsubalg when input is linearly dependent             [F3]
AP   4- fix alglatinter trivial intersection (misuse of matkermod(,1))     [F4]
BA   5- ellrank could try many more curves than requested by effort [#2577][F5]
     6- wrong result in issquare(Mod(a, composite)) [#2578]                [F6]
     7- gener_FpXQ_local(t_POL of degre 1) -> oo loop [#2581]              [F7]
     8- GC error when squaring integer 2x2 matrices [#2582]                [F8]
AP   9- algebras: fix splitting field for CSA, algtomatrix / algcharpoly
        were sometimes wrong by an automorphism of the center              [F9]
BA  10- paridecl.h could conflict with complex.h [#2583]                  [F10]
    11- forprime(p=524288,1048576,1) -> crash [#2584]
        forprime(p=1048607,1048617,1) -> oo loop                          [F11]
    12- printf("%10.6f", tiny rational) -> undefined result [#2576]       [F12]
AP  13- alginit(nf,[a,b]) was only allowing a,b integral over the power basis
        instead of arbitrary algebraic integers.                          [F13]
HC  14- zeta not continous around 1/2+3000*I [#2585]                      [F14]
    15- missing type checks for change of variables in ellchangecurve,    [F15]
        ellchangepoint, ellchangepointinv
BA  16- control-C could lead to segfault later [#2496]                    [F16]
    17- [libpari] missing normalization in FpX_mulu (when p is not prime) [F18]
HC  18- mfsymboleval could return wrong sign [#2586]                      [F19]
    19- missing type checks in factorback [#2590]                         [F20]
BA  20- charpoly(Mod(x,x^20+1/2)) -> crash [#2591]                        [F21]
    21- psi(x close to a nonpositive integer) -> error                    [F22]
BA  22- polresultant(t_POL with RFRAC coeffs) could be incorrect [#2588]  [F24]
BA  23- qflll(non-maximal rank) -> potential oo-loop [#2587]              [F25]

  Added
BA   1- [loongarch64] add loongarch64 assembly kernel

Done for version 2.17.0 (released 30/09/2024):
  Fixed
     1- rare SEGV in sumnumrat [#2566]
     2- gp -p N did not check for N < 2^32
     3- nflist([6,8],12167) -> SEGV [#2567]
     4- nflist([6,1],1229312), nflist([6,10],36192256) -> duplicate fields
AP   5- variable leak in algsplit
     6- mfslashexpansion: possible extra rational factor [#2572]
     7- nfeltsign: possible wrong result

  Added
BA   1- [libpari] Flm_sqr

  Changed
BA   1- qflll(,1), qfllgram(,1): disable FLATTER algorithm.
     2- tunings of forprime over "short" intervals (use sieve more frequently)
