nix: update bytestring-strict-builder to 0.4.5.6

[robx]

This comes with a rewrite rule that hopefully allows us to avoid
a copy of the request body.

[steve-chavez]

Hm, would this change be mutually exclusive with the change in #2333? That one seemed to give better results.

[robx]

The rule doesn't trigger, yet, unfortunately. If it worked, it would be equivalent as a first step, and allow us to avoid a second copy if we manage to read the body into a strict bytestring directly.

The other advantage of this approach is that it's something that doesn't require potentially unwanted upstream changes, compare discussion in nikita-volkov/hasql#146 and nikita-volkov/bytestring-strict-builder#11.

But if this doesn't work out quickly, we can definitely go with the forked version of hasql (#2333) momentarily and follow up after.

[robx]

The rewrite rule turned out not to trigger for us, closing this.