feat(serde-up) Support serde 0.7 #32
Conversation
|
Hmm, tests passed on my machine. Will look into the failure later. |
|
The build is now passing. The serde dependency will still need to be updated to 0.7 once it is published. |
|
Tests are passing with latest master of serde |
| ErrorCode::MissingField(ref field) => write!(f, "missing field \"{}\"", field), | ||
| ErrorCode::TrailingCharacters => "trailing characters".fmt(f), | ||
| ErrorCode::UnexpectedEndOfHexEscape => "unexpected end of hex escape".fmt(f), | ||
| ErrorCode::Length(ref len) => write!(f, "incorrect value length {}", len), |
There was a problem hiding this comment.
These changes are already contained in PR #11 and are likely to conflict if that PR is merged.
|
Rebased on latest json#master and tested with latest serde#master. There's one test failing right now that I don't have time to chase down; It looks like the |
|
I also stumbled across that issue. Look at this https://github.com/tomprogrammer/json/commit/5eb789dc4d8ad4db013b3f8f9881b6009a50d69c commit. You can see I introduced a new ("{\"unknown\":[]}", Error::Syntax(ErrorCode::UnknownVariant("unknown".to_string()), 1, 10)), |
|
The interaction between your and my PR are a bit unfortunate though. I don't know how to solve this best. |
|
@erickt just mentioned
So my branch may be moot. Still, any commit adding support for 0.7 will have similar conflicts. |
|
This has actually landed since serde 0.7 was released! |
3 participants
NOT READY this needs changes before merging; Cargo.toml spec is hard coded to a git rev
Mappings from visit_* to de/serialize_* should be self explanatory. [serde(deny_unknown_fields)] was added to a test struct since it was checking for denial. A couple of error types were added to support new
serde de::value::Error variants. Finally, there are a bunch of throwaway changes pinning the serde version to a 0.7 commit. These should be removed once 0.7 lands in favor of a crates.io semver spec.