Questions: automated trial sieving

[Dubslow]

So, as a byproduct of other work that's going on, YAFU has the potential to implement a fully featured test sieving mechanism. So I'm curious: those who would use it, what features would you most desire?

Should it be an option to nfs(), e.g. `"nfs(num)" -nt file1,file2,file3`? Would you prefer it to be its own function, e.g. `"trialsieve(num, file1, file2, ...)"`? (In each case, the arguments would be names of job files with different poly/params.) What options should there be?

Some things I can think of that would need configurability are qrange, qstart, and maybe even give a list of qstarts to properly trial sieve the whole sieving range. Another option may be to trial sieve the polys in parallel, though that wouldn't be enabled by default due to potential gotchas. (Edit: I mean one jobfile per thread.)