Integrate[q2^eps Log(q2+m2) (q2+m2)^-1,{q2,0,Infinity}]
gives
0, if -1<Re[eps]<0 && Re[m2]>0
which is wrong.
If ones adds
Assumptions -> {m2 > 0, -1 < eps < 0}
then it is correct.
Thanks Mariusz. I hope they may fix it quickly.
Yes It's a bug. I send to Wolfram Support.
Regards M.I.