the last missing curve with norm-conductor < 2000(!) (I think)

[Jonathan Bober]

I think that this was the only curve that we were missing: [ 0, phi +
1, 0, -6*phi + 7, -7*phi + 3 ].

Lassina found it using "a search routine that has been implemented in
Magma by Steve Donnelly based on an idea of Elkies" (I don't know what
that means) after I was told him that this was the only curve we were
missing and that I hadn't got his method to work on it.

In fact, I can find this curve immediately using the sieving code that
we have (those coefficients are pretty small), so in retrospect we
should have found this curve rather easily. I don't know why we
didn't.

[R. Andrew Ohana]

The version of the sieving code from last summer had a lot of bugs in it causing to frequently miss curves it should have found. I fixed many of these bugs, but I don't think I ever ran it past 1831.