Error with batch and batch_deterministic feature combo

[pinkforest]

This PR depends on

• Propagate CtOption and ZERO from curve/4.0 #239

Currently the error if both are activated is a random one with cfg()'s that didn't expect someone would use both.

Also simplified cfg() in batch.rs

This creates proper error if both are activated.

foobar@foobar-MS-7C77:~/cc/ed25519-dalek$ cargo test --all-features

   Compiling ed25519-dalek v1.0.1 (/home/foobar/cc/ed25519-dalek)
error: Features batch and batch_deterministic are mutually exclusive.
   --> src/lib.rs:236:1
    |
236 | compile_error!("Features batch and batch_deterministic are mutually exclusive.");
    | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

[tarcieri]

Is there any reason to keep batch instead of replacing it entirely with batch_deterministic?

[pinkforest]

to keep rand_core optional I guess ?

Also:

• Merlin dep is broken on big-endian machines #228 (comment)

deterministic one is unstable ? - but it also speaks about merlin ? (I didn't investigate deeply yet 🙈 )

EDIT - Oh there is also this thread yes:

• Fix batch build #220 (comment)

So shall I get rid of batch here instead ?

[tarcieri]

I think that’s definitely worth discussing

[pinkforest]

Closed in favor of #220 (all sorts of batch fixes) and #235 (compile error)