#specification-language #parser #rpsl #routing #policy #property-testing #whois

rpsl-rs

A Routing Policy Specification Language (RPSL) parser with a focus on speed and correctness

3 stable releases

2.0.0 Nov 3, 2024
1.0.1 Dec 26, 2023

#329 in Parser implementations

Download history 4/week @ 2024-09-23 25/week @ 2024-09-30 63/week @ 2024-10-28 48/week @ 2024-11-04

111 downloads per month

MIT license

490KB
2K SLoC

rpsl-rs

A Routing Policy Specification Language (RPSL) parser with a focus on speed and correctness.

⚡️ 130-250x faster than other parsers
📰 Complete implementation for multiline RPSL values
💬 Able to parse objects directly from whois server responses
🧠 Low memory footprint by leveraging zero-copy
🧪 Robust parsing of any valid input ensured by Property Based Tests

Docs | Performance

Usage

Parsing RPSL objects

A string containing an object in RPSL notation can be parsed to an Object using the parse_object function.

use rpsl::parse_object;

let role_acme = "
role:        ACME Company
address:     Packet Street 6
address:     128 Series of Tubes
address:     Internet
email:       rpsl-rs@github.com
nic-hdl:     RPSL1-RIPE
source:      RIPE

";
let parsed = parse_object(role_acme).unwrap();

The returned Object allows access to the attributes contained within in form of Attributes.

println!("{:#?}", parsed);

Object(
    [
        Attribute {
            name: Name("role"),
            value: SingleLine(Some("ACME Company")),
        },
        Attribute {
            name: Name("address"),
            value: SingleLine(Some("Packet Street 6")),
        },
        Attribute {
            name: Name("address"),
            value: SingleLine(Some("128 Series of Tubes")),
        },
        Attribute {
            name: Name("address"),
            value: SingleLine(Some("Internet")),
        },
        Attribute {
            name: Name("email"),
            value: SingleLine(Some("rpsl-rs@github.com")),
        },
        Attribute {
            name: Name("nic-hdl"),
            value: SingleLine(Some("RPSL1-RIPE")),
        },
        Attribute {
            name: Name("source"),
            value: SingleLine(Some("RIPE")),
        },
    ]
)

Objects created from RPSL text use string references that point to attributes and their values instead of copying them.

role:           ACME Company ◀─────────────── &"role"    ───  &"ACME Company"
address:        Packet Street 6 ◀──────────── &"address" ───  &"Packet Street 6"
address:        128 Series of Tubes ◀──────── &"address" ───  &"128 Series of Tubes"
address:        Internet ◀─────────────────── &"address" ───  &"Internet"
email:          rpsl-rs@github.com ◀───────── &"email"   ───  &"rpsl-rs@github.com"
nic-hdl:        RPSL1-RIPE ◀───────────────── &"nic-hdl" ───  &"RPSL1-RIPE"
source:         RIPE ◀─────────────────────── &"source"  ───  &"RIPE"

This is what makes rpsl-rs performant and memory efficient, since no additional allocation is required during parsing.

Each Attribute can be accessed by its index and has a name and value.

println!("{:#?}", parsed[1]);

Attribute {
    name: Name("address"),
    value: SingleLine(Some("Packet Street 6")),
}

Since RPSL attribute values can either be single- or multiline, two different variants are used to represent them. See Attribute and parse_object for more details and examples.

Parsing a WHOIS server response

WHOIS servers often respond to queries by returning multiple related objects. An example ARIN query for AS32934 will return with the requested ASNumber object first, followed by its associated OrgName:

$ whois -h whois.arin.net AS32934
ASNumber:       32934
ASName:         FACEBOOK
ASHandle:       AS32934
RegDate:        2004-08-24
Updated:        2012-02-24
Comment:        Please send abuse reports to abuse@facebook.com
Ref:            https://rdap.arin.net/registry/autnum/32934


OrgName:        Facebook, Inc.
OrgId:          THEFA-3
Address:        1601 Willow Rd.
City:           Menlo Park
StateProv:      CA
PostalCode:     94025
Country:        US
RegDate:        2004-08-11
Updated:        2012-04-17
Ref:            https://rdap.arin.net/registry/entity/THEFA-3


To extract each individual object, the parse_whois_response function can be used to parse the response into a Vec containing all individual Objects within the response. Examples can be found in the function documentation.

Optional Features

The following cargo features can be used to enable additional functionality.

  • simd (enabled by default): Enables the Winnow simd feature which improves string search performance using simd.
  • serde: Enables Object serialization using Serde.
  • json: Provides JSON serialization of an Object using Serde JSON.

MSRV Policy

This project requires the minimum supported Rust version to be at least 6 months old. As long as this requirement is met, the MSRV may be increased as necessary through a minor version update. For the currently configured MSRV, please check Cargo.toml.

Contributing

Contributions of all sizes that improve rpsl-rs in any way, be it DX/UX, documentation, performance or other are highly appreciated. To get started, please read the contribution guidelines. Before starting work on a new feature you would like to contribute that may impact simplicity, reliability or performance, please open an issue first.

License

The source code of this project is licensed under the MIT License. For more information, see LICENSE.

Dependencies

~1–1.7MB
~36K SLoC