Chemical abundances and abundance ratios measured in gas and stars in galaxies provide precious information about the modes and time-scales of the assembly of cosmic structures. Yet, the nucleogenesis and chemical evolution of elements heavier than helium are dictated mostly by the physics of the stars and the shape of the stellar mass spectrum. Recent work by our team demonstrated that the abundance ratios of specific CNO isotopes give a unique glimpse into the shape of the galaxy-wide stellar initial mass function (IMF), especially so in powerful starburst galaxies. Underlying uncertainties in stellar evolution and nucleosynthesis theory, however, may undermine our conclusions. In this talk I will discuss the implications of the adoption of new stellar yields coming from rotating massive star models. I will first implement the new yield set in a well-grounded chemical evolution model for the Milky Way. Then, the calibrated model will be adapted to the case of submillimeter galaxies observed at high redshifts. I will show that our conclusion that a top-heavy IMF has to be preferred in starbursts still holds, unless very high (most probably, too high) star formation rates are assumed to characterize the galaxies in our high-redshift sample.