Use new hex-conservative crate

[tcharding]

Use the newly released hex-conservative crate, by doing the following:

• Depend on hex-conservative in bitcoin and hashes
• Re-export hex-conservative as hex from both crate roots.
• Remove all the old hex code from hashes
• Remove all the old hex code from internals
• Remove the now unused internals::prelude
• Fix all the import statements (makes up the bulk of the lines changes in this patch)

[apoelstra]

I think we should rename it to Iter and yeah, we should export it at the crate root.

[apoelstra]

Nice! This looks good now.

[tcharding]

Thanks man, looks sweet huh.

[tcharding]

This will need the lock files updating still, hopefully #1916 will be in by then so msrv pinning stuff is resolved.

[apoelstra]

I think we should do the hex stuff before any other releases.

[tcharding]

Something is wrong with the core2 build and I can't seem to see it, does anything cry out?

git diff @~ hashes/Cargo.toml
diff --git a/hashes/Cargo.toml b/hashes/Cargo.toml
index 0516af97..ac8591b9 100644
--- a/hashes/Cargo.toml
+++ b/hashes/Cargo.toml
@@ -14,9 +14,10 @@ exclude = ["tests", "contrib"]
 
 [features]
 default = ["std"]
-std = ["alloc", "internals/std"]
-alloc = ["internals/alloc"]
+std = ["alloc", "internals/std", "hex/std"]
+alloc = ["internals/alloc", "hex/alloc"]
 serde-std = ["serde/std"]
+core2 = ["actual-core2", "hex/core2"]
 
 [package.metadata.docs.rs]
 all-features = true
@@ -24,12 +25,15 @@ rustdoc-args = ["--cfg", "docsrs"]
 
 [dependencies]
 internals = { path = "../internals", package = "bitcoin-internals", version = "0.1.0" }
+hex = { package = "hex-conservative", version = "0.1.0", default-features = false }
 
-core2 = { version = "0.3.0", default_features = false, optional = true }
 schemars = { version = "0.8.3", optional = true }
 # Only enable this if you explicitly do not want to use "std", otherwise enable "serde-std".
 serde = { version = "1.0", default-features = false, optional = true }
 
+# Do NOT use this feature! Use the `core2` feature instead.
+actual-core2 = { package = "core2", version = "0.3.0", default-features = false, optional = true }
+

[apoelstra]

@tcharding did you update Cargo-recent.lock and Cargo-minimal.lock after making those changes to Cargo.toml?

[tcharding]

Yeah I did, I couldn't get hashes to build even with cargo check locally so I figured it was something brain dead and I was just blind to it. I'll play more with it today.

[tcharding]

FTR the solution was

extern crate actual_core2 as core2;

But yesterday I hade extern crate core2;

[apoelstra]

Awesome, glad CI is passing.

I'm temped though to get hex-conservative 0.1.1 out first (with the new hex! macro and fixing the fmt impls to respect {:8})

[tcharding]

Agree, mind if I call it v0.2.0? I've given up understanding Rust semver rules for under v1.0, just feels too trivial increasing only the patch number :)

[apoelstra]

LGTM. May be a bit easier to review split into multiple commits, one for each crate, and a final one which deletes the hex stuff from -internals.

[apoelstra]

ACK bd8a74b

[tcharding]

Can do!

[tcharding]

I had a go but we can't split it up without removing the workspace patch of hashes because of the hash_newtype! macro. I could remove that as an initial patch, do each crate separately, then add it back in as a final patch, that kind of defeats the purpose of having each patch build though because git bisecter's would not expect such a change half way through bisecting.

[tcharding]

I pulled the internals changes out into a separate patch, its all red :)

[tcharding]

And I removed the unnecessary actual-core2 dependency trick from bitcoin because its unneeded since we only ever expect users to use "no-std" to activate the "core2" dependency - added an initial patch to this PR that comments the "core2" feature as such.

[apoelstra]

ACK bb8bd16

[tcharding]

Draft because we may be able to make the hex-conservative library available as hex after all - @stevenroose prompted me to look into it just now.

[tcharding]

The hex lib rename thing got closed so this is good to go.

[tcharding]

@sanket1729 would you mind taking a look at this if you get a chance please? The green in the diff is almost all import statements and error type renames. The red is 1000+ lines of mostly removing all the stuff that has been released already in hex-conservative - I don't think you need to put much thought into those lines.

One thing stand out as non-obvious, the implementation of FromHex in internal_macros has a lot of code removed and we call through to FromHex which has exactly the same code as that removed, might have to open the editor to confirm that.

Thanks man.

[sanket1729]

utACK bb8bd16

[tcharding]

Oh my goodness we can merge this now, this has been a loooooong time coming :)