//! Security headers on every response (oxygenhub#222), and the endpoint

//! the browser reports Content-Security-Policy violations to.

//!

//! The policy is sent as `Content-Security-Policy-Report-Only` while the

//! pages still carry inline scripts and handlers: browsers report what the

//! policy would block, the hub logs it, and nothing breaks. Switching to

//! enforcing is one constant ([`ENFORCE`]) once the reports are quiet.

//! Responses that set their own policy (report frames, #127; a job's

//! HTML artifact) keep theirs, and may be framed as they say.

//! Strict-Transport-Security is deliberately not sent yet: browsers keep

//! it for months, so it waits for the site owner's go.

use axum::http::{header, HeaderName, HeaderValue};

use axum::response::Response;

use std::collections::HashMap;

use std::sync::Mutex;

/// Where violation reports go.

pub const REPORT_PATH: &str = "/csp-report";

/// Enforce the policy rather than only report it.

const ENFORCE: bool = false;

/// What the pages may load: their own scripts (and WebAssembly, for the

/// backdrop), styles (inline too: style attributes are everywhere), images

/// from anywhere over https (READMEs show badges), and forms posted only to

/// the hub (sign-in providers are reached by links, not forms).

const POLICY: &str = "default-src 'self'; script-src 'self' 'wasm-unsafe-eval'; style-src 'self' 'unsafe-inline'; \

img-src 'self' data: https:; font-src 'self'; connect-src 'self'; media-src 'self'; object-src 'none'; \

base-uri 'self'; form-action 'self'; frame-ancestors 'none'; report-uri /csp-report; report-to csp";

const PERMISSIONS: &str = "camera=(), microphone=(), geolocation=(), payment=(), usb=(), serial=(), bluetooth=(), \

midi=(), magnetometer=(), gyroscope=(), accelerometer=(), display-capture=(), browsing-topics=(), interest-cohort=()";

/// Adds the headers a response does not already carry.

pub async fn security_headers(req: axum::extract::Request, next: axum::middleware::Next) -> Response {

    let mut res = next.run(req).await;

    let h = res.headers_mut();

    let own_policy = h.contains_key(header::CONTENT_SECURITY_POLICY);

    let mut set = |name: HeaderName, value: &'static str| {

        h.entry(name).or_insert(HeaderValue::from_static(value));

    };

    set(header::X_CONTENT_TYPE_OPTIONS, "nosniff");

    set(header::REFERRER_POLICY, "strict-origin-when-cross-origin");

    set(HeaderName::from_static("permissions-policy"), PERMISSIONS);

    if !own_policy {

        set(header::X_FRAME_OPTIONS, "DENY");

        set(HeaderName::from_static("reporting-endpoints"), "csp=\"/csp-report\"");

        let name = if ENFORCE { header::CONTENT_SECURITY_POLICY } else { header::CONTENT_SECURITY_POLICY_REPORT_ONLY };

        set(name, POLICY);

    }

    res

}

/// One violation, as either report format gives it.

#[derive(Debug, PartialEq)]

pub struct Violation {

    pub page: String,

    pub directive: String,

    pub blocked: String,

    pub source: String,

}

/// The violations in a report body: the older `application/csp-report`

/// (`{"csp-report": {...}}`) or the Reporting API's list.

pub fn violations(body: &[u8]) -> Vec<Violation> {

    let Ok(v) = serde_json::from_slice::<serde_json::Value>(body) else { return vec![] };

    let s = |v: &serde_json::Value, keys: &[&str]| {

        keys.iter().find_map(|k| v.get(*k).and_then(|x| x.as_str().map(String::from).or_else(|| x.as_u64().map(|n| n.to_string()))))

    };

    let one = |r: &serde_json::Value, page: &[&str], dir: &[&str], blocked: &[&str], file: &[&str], line: &[&str]| Violation {

        page: s(r, page).unwrap_or_default(),

        directive: s(r, dir).unwrap_or_default(),

        blocked: s(r, blocked).unwrap_or_default(),

        source: match (s(r, file), s(r, line)) {

            (Some(f), Some(l)) => format!("{f}:{l}"),

            (Some(f), None) => f,

            _ => String::new(),

        },

    };

    if let Some(r) = v.get("csp-report") {

        return vec![one(

            r,

            &["document-uri"],

            &["effective-directive", "violated-directive"],

            &["blocked-uri"],

            &["source-file"],

            &["line-number"],

        )];

    }

    v.as_array()

        .map(|list| {

            list.iter()

                .filter(|r| r.get("type").and_then(|t| t.as_str()) == Some("csp-violation"))

                .filter_map(|r| r.get("body"))

                .map(|b| one(b, &["documentURL"], &["effectiveDirective"], &["blockedURL"], &["sourceFile"], &["lineNumber"]))

                .collect()

        })

        .unwrap_or_default()

}

/// Each distinct violation is logged once an hour, so a busy page does not

/// flood the log.

static SEEN: Mutex<Option<HashMap<String, i64>>> = Mutex::new(None);

const LOG_EVERY_MS: i64 = 60 * 60_000;

fn first_in_a_while(key: String, now_ms: i64) -> bool {

    let mut seen = SEEN.lock().unwrap();

    let seen = seen.get_or_insert_with(HashMap::new);

    if seen.len() > 1000 {

        seen.retain(|_, t| now_ms - *t < LOG_EVERY_MS);

    }

    match seen.get(&key) {

        Some(t) if now_ms - t < LOG_EVERY_MS => false,

        _ => {

            seen.insert(key, now_ms);

            true

        }

    }

}

/// `POST /csp-report`: log what the policy would block.

pub async fn csp_report(body: axum::body::Bytes) -> axum::http::StatusCode {

    let clip = |s: &str| s.chars().filter(|c| !c.is_control()).take(200).collect::<String>();

    for v in violations(&body) {

        // The page's path only: a query can hold anything.

        let page = v.page.split(['?', '#']).next().unwrap_or("").to_string();

        let key = format!("{}|{}|{}", v.directive, v.blocked, v.source);

        if first_in_a_while(key, crate::db::now_ms()) {

            tracing::warn!("csp would block: {} {} on {} ({})", clip(&v.directive), clip(&v.blocked), clip(&page), clip(&v.source));

        }

    }

    axum::http::StatusCode::NO_CONTENT

}

#[cfg(test)]

mod tests {

    use super::*;

    #[test]

    fn both_report_formats_are_read() {

        let old = br#"{"csp-report":{"document-uri":"https://hub.example/a?b","violated-directive":"script-src-elem","blocked-uri":"inline","source-file":"https://hub.example/a","line-number":12}}"#;

        assert_eq!(

            violations(old),

            vec![Violation {

                page: "https://hub.example/a?b".into(),

                directive: "script-src-elem".into(),

                blocked: "inline".into(),

                source: "https://hub.example/a:12".into()

            }]

        );

        let new = br#"[{"type":"csp-violation","url":"https://hub.example/x","body":{"documentURL":"https://hub.example/x","effectiveDirective":"img-src","blockedURL":"http://cdn.example/i.png"}},{"type":"deprecation","body":{}}]"#;

        let v = violations(new);

        assert_eq!(v.len(), 1);

        assert_eq!((v[0].directive.as_str(), v[0].blocked.as_str(), v[0].source.as_str()), ("img-src", "http://cdn.example/i.png", ""));

        assert!(violations(b"not json").is_empty());

    }

    #[test]

    fn a_violation_is_logged_once_an_hour() {

        let k = || "script-src|inline|test-only".to_string();

        assert!(first_in_a_while(k(), 1_000));

        assert!(!first_in_a_while(k(), 2_000));

        assert!(first_in_a_while(k(), 1_000 + LOG_EVERY_MS));

    }

}