SW2: Make the cents offsets behave

[frostburn]

pinkanberry — Today at 3:08 AM

if you can forgive the frivolous feature request: when you specify generators manually via cents and do smth like span lattice, you often get these trollishly long floating point artifacts for example "span lattice" with 704.71 57.42, 12x2, gives this: https://sevish.com/scaleworkshop/?n=Lattice%20(12%2C2%20of%20704.71%2057.42)&l=1F1SPS1l.16_1FgSPS3t0.2p_1FgSPS3um.13_1F2SPS135.16_1F2SPS14q.2z6wn49hd_1FwSPS4w6.13_1FwSPS4xr.29_1F4SPS26a.2c_1F4SPS27w.q_1F1sSPS5zb.29_1F1sSPS60x.tclsm611_1F8SPS39g.q_1F8SPS3b1.1w_1F1SPSjk.1z_1F1SPSl6.d_1FgSPS4cl.1w_1FgSPS4e7.1_1F2SPS1mq.d_1F2SPS1ob.1j_1FwSPS5fr.1_1FwSPS5hc.1g_1F4SPS2pv.1j_1F4SPS2rg.3fs49jx8j_2F1&version=2.4.1 this is just painful to see and i feel like it shouldnt be too hard in principle to get them to round correctly here is the corrected version for comparison: https://sevish.com/scaleworkshop/?n=Lattice%20(12%2C2%20of%20704.71%2057.42)&l=1F1SPS1l.16_1FgSPS3t0.2p_1FgSPS3um.13_1F2SPS135.16_1F2SPS14q.2c_1FwSPS4w6.13_1FwSPS4xr.29_1F4SPS26a.2c_1F4SPS27w.q_1F1sSPS5zb.29_1F1sSPS60x.n_1F8SPS39g.q_1F8SPS3b1.1w_1F1SPSjk.1z_1F1SPSl6.d_1FgSPS4cl.1w_1FgSPS4e7.1_1F2SPS1mq.d_1F2SPS1ob.1j_1FwSPS5fr.1_1FwSPS5hc.1g_1F4SPS2pv.1j_1F4SPS2rg.2p_2F1&version=2.4.1 probably the easiest fix is to round to 1 cent / 10^n for some acceptably large value of n that is significantly far away from the max precision

[frostburn]

I am very sorry, but fixing this in v2 series is not possible. The issue you're referring to was the main reason why in Scale Workshop 3 cents are now associated with the octave and why real values no longer have rational parts.

[osmiumic]

oh that's alright, i just meant fixing it in general it's not too important that it's sooner rather than later

[frostburn]

Related to #593