rnfpolred bug

[Bill Allombert <allomber@math.u-bordeaux.fr>]

Hello PARI-dev,

It seems that rnfpolred has a bug when used with bnf:

? rnfpolred(nfinit(polcyclo(3,a)),x^3-19);
time = 30 ms.
? rnfpolred(bnfinit(polcyclo(3,a)),x^3-19);
  ***   user interrupt after 1mn, 41,460 ms.

Also 
? ?rnfkummer
rnfkummer(bnr,{subgroup},{deg=0}): bnr being as output by bnrinit, finds a
relative equation for the class field corresponding to the module in bnr and
the given congruence subgroup (the ray class field if subgroup is omitted).
deg can be zero (default), or positive, and in this case the output is the
list of all relative equations of degree deg for the given bnr.

The last sentence is not accurate. We should specify 
``with the same conductor as (bnr, subgroup)''.

Cheers,
Bill.