I am new about these things, but wouldn't just moving to fixed computation really fix this and a large number of forsee-able side-channel attacks? It sounds lik

I am new about these things, but wouldn't just moving to fixed computation really fix this and a large number of forsee-able side-channel attacks? It sounds like the libgcrypt devs chose specifically not to do this because they thought it wasn't necessary.