Remove error chain

[adeschamps]

This turned out to be really easy to do, so merge it or close as you prefer. Since it came in #43 that the cc crate no longer compiles on 1.28, I removed the error-chain dependency, which was transitively including cc. All the errors produced by this crate are just strings anyway. I replaced it with a similar Error and ErrorKind pair, and implemented std::error::Error.

I think this is unlikely to break anyone's code, but it's technically a breaking change.

[sebest]

any chance to merge this change?

error_chains seems to be not really maintain anymore (beyond maintenance), one issue that I'm facing is that error_chains is not 'Sync'.

See the following threads for more context:

• Require that errors are Send/Sync/'static. rust-lang-deprecated/error-chain#241
• support wrapping error_chain Error's ? / document that Send+Sync are required on top of std::error::Error dtolnay/anyhow#72

Moving to std::error::Error (or thiserror) would solve this issue.

Thank you!

[zslayton]

This is a good change; thanks for putting it together and sorry it fell off my radar. If you could rebase it against master I'd be happy to merge it.

[sebest]

@adeschamps : can you rebase/merge this pull request?

Otherwise should we close this pull request and create a new one?

[zslayton]

Otherwise should we close this pull request and create a new one?

Instead of rebasing you can merge from master and push another commit to this PR. I'm going to squash/merge it anyway, so the end result will be the same.

[sebest]

@zslayton : can I push a commit to this PR even if I'm not the initial author?

[zslayton]

@zslayton : can I push a commit to this PR even if I'm not the initial author?

I don't believe the permissions will allow you to. I think GitHub would expect you to fork the PR branch, make some changes, and then open another PR to merge your changes back into @adeschamps' original branch.

If you'd like to clone the PR branch and open a new PR, we can work on getting yours merged instead.

[sebest]

@zslayton I created a new PR that rebase this one on master: #65

[zslayton]

Closed in favor of #65, which contains the same changes. Thanks @sebest, @adeschamps!