Add no_std support.

[yuguorui]

You may need to accept https://github.com/rusticata/rusticata-macros/pull/3 first to use macro newtype_enum.

[yuguorui]

Seems some checks are not passed, I will check that.

[chifflier]

Hi, Thanks for your PR! I am interesting in adding no_std support, and originally I thought Cow would be a problem, and it's nice to have a solution. The most important error seems to be:

Error:  --> src/ber/serialize.rs:9:11
  |
9 | use core::io::Write;
  |           ^^ could not find `io` in `core`

Thanks!

[yuguorui]

Please note that other upstream repo (such as oid-registry) may not compile due to current changes, I am working on it.

[chifflier]

Thanks for the PR update The rustfmt error can be ignored, I'll fix it later

One question: how can we test that the crate is no_std-compliant? (and its dependencies?) I'd like to add that to the CI, to ensure no_std support is not broken by later commits

[yuguorui]

Thanks for the PR update The rustfmt error can be ignored, I'll fix it later

One question: how can we test that the crate is no_std-compliant? (and its dependencies?) I'd like to add that to the CI, to ensure no_std support is not broken by later commits

I think this could be checked by cargo test --no-default-features? But as I said at the beginning, in this case we will encounter errors related to newtype_enum.

$ cargo test --no-default-features
   Compiling der-parser v5.1.0 (/mnt/data1/Code/der-parser)
error[E0433]: failed to resolve: could not find `std` in the list of imported crates
  --> src/ber/ber.rs:47:1
   |
47 | / newtype_enum! {
48 | | impl debug BerTag {
49 | |     EndOfContent = 0x0,
50 | |     Boolean = 0x1,
...  |
83 | | }
84 | | }
   | |_^ could not find `std` in the list of imported crates
   |
   = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info)

[yuguorui]

Please note that other upstream repo (such as oid-registry) may not compile due to current changes, I am working on it.

Now, oid-registry can be compiled directly without any modification.

[chifflier]

I think this could be checked by cargo test --no-default-features?

Indeed, I don't know why I was confused.

But as I said at the beginning, in this case we will encounter errors related to newtype_enum.

I encountered this one while testing another crate. I'll update rusticata-macros to provide a no_std-friendly version of the macro. I'm taking care of this one.

BTW, it's really annoying to have to add extern crate core, this makes providing an impl in the macro for both std and no_std quite complex :/

[chifflier]

I encountered this one while testing another crate. I'll update rusticata-macros to provide a no_std-friendly version of the macro. I'm taking care of this one.

Fixed in (released) rusticata-macros 3.1.0

[chifflier]

I'm merging this now, and will add some commits (rustfmt, add no_std to CI, and maybe update the exports for alloc). Thanks for your contribution!