Miniserde

Prototype of a data structure serialization library with several opposite design goals from Serde.

As a prototype, this library is not a production quality engineering artifact the way Serde is. At the same time, it is more than a proof of concept and should be totally usable for the range of use cases that it targets, which is qualified below.

[dependencies]
miniserde = "0.1"

Version requirement: rustc 1.56+

Example

use miniserde::{json, Serialize, Deserialize};

#[derive(Serialize, Deserialize, Debug)]
struct Example {
    code: u32,
    message: String,
}

fn main() -> miniserde::Result<()> {
    let example = Example {
        code: 200,
        message: "reminiscent of Serde".to_owned(),
    };

    let j = json::to_string(&example);
    println!("{}", j);

    let out: Example = json::from_str(&j)?;
    println!("{:?}", out);

    Ok(())
}

Here are some similarities and differences compared to Serde.

Similar: Stupidly good performance

Seriously this library is way faster than it deserves to be. With very little profiling and optimization so far and opportunities for improvement, this library is on par with serde_json for some use cases, slower by a factor of 1.5 for most, and slower by a factor of 2 for some. That is remarkable considering the other advantages below.

Similar: Strongly typed data

Just like Serde, we provide a derive macro for a Serialize and Deserialize trait. You derive these traits on your own data structures and use json::to_string to convert any Serialize type to JSON and json::from_str to parse JSON into any Deserialize type. Like serde_json there is a Value enum for embedding untyped components.

Different: Minimal design

This library does not tackle as expansive of a range of use cases as Serde does. Feature requests are practically guaranteed to be rejected. If your use case is not already covered, please use Serde.

The implementation is less code by a factor of 12 compared to serde + serde_derive + serde_json, and less code even than the json crate which provides no derive macro and cannot manipulate strongly typed data.

Different: No monomorphization

There are no nontrivial generic methods. All serialization and deserialization happens in terms of trait objects. Thus no code is compiled more than once across different generic parameters. In contrast, serde_json needs to stamp out a fair amount of generic code for each choice of data structure being serialized or deserialized.

Without monomorphization, the derived impls compile lightning fast and occupy very little size in the executable.

Different: No recursion

Serde depends on recursion for serialization as well as deserialization. Every level of nesting in your data means more stack usage until eventually you overflow the stack. Some formats set a cap on nesting depth to prevent stack overflows and just refuse to deserialize deeply nested data.

In miniserde neither serialization nor deserialization involves recursion. You can safely process arbitrarily nested data without being exposed to stack overflows. Not even the Drop impl of our json Value type is recursive so you can safely nest them arbitrarily.

Different: No deserialization error messages

When deserialization fails, the error type is a unit struct containing no information. This is a legit strategy and not just laziness. If your use case does not require error messages, good, you save on compiling and having your instruction cache polluted by error handling code. If you do need error messages, then upon error you can pass the same input to serde_json to receive a line, column, and helpful description of the failure. This keeps error handling logic out of caches along the performance-critical codepath.

Different: Infallible serialization

Serialization always succeeds. This means we cannot serialize some data types that Serde can serialize, such as Mutex which may fail to serialize due to poisoning. Also we only serialize to String, not to something like an i/o stream which may be fallible.

Different: JSON only

The same approach in this library could be made to work for other data formats, but it is not a goal to enable that through what this library exposes.

Different: Structs and unit variants only

The miniserde derive macros will refuse anything other than a braced struct with named fields or an enum with C-style variants. Tuple structs are not supported, and enums with data in their variants are not supported.

Different: No customization

Serde has tons of knobs for configuring the derived serialization and deserialization logic through attributes. Or for the ultimate level of configurability you can handwrite arbitrarily complicated implementations of its traits.

Miniserde provides just one attribute which is rename, and severely restricts the kinds of on-the-fly manipulation that are possible in custom impls. If you need any of this, use Serde -- it's a great library.

License