Bug with projective roots with -sublat and -adjust-strategy 2
Imported issue: Initially reported by @gaudry in https://gforge.inria.fr/tracker/?group_id=2065&aid=21593
las sometimes fails with ""p=3 does not divide"" with -sublat 3 and -adjust-strategy 2.
The bug disappears with -powlim0 1 -powlim1 1, and I vaguely remember having investigated
up to the point where a power of 3 is pushed in the bucket-sieving part, which is not supported.
Example to reproduce:
/tmp/cado-nfs-build-master/sieve/las -A 19 -sublat 3 -poly /tmp/c60.poly -q0 92000 -nq 1 -lim0 78682 -lim1 111342 -lpb0 18 -lpb1 19 -mfb0 17 -mfb1 38 -fb /tmp/c60.roots.gz -t 1 -adjust-strategy 2
with attached files.