ngit/

mbox_parser.rs

//! Parser for mbox-format git patch content.
//!
//! This module is a **fallback path** used only when nostr patch events are
//! missing optional tags (`author`, `committer`, `description`,
//! `parent-commit`). When those tags are present they always take precedence —
//! see [`crate::git::RepoActions::apply_patch_chain`].
//!
//! ## Why hand-rolled rather than a library?
//!
//! Neither libgit2 (via the `git2` crate) nor gitoxide (`gix`) exposes a
//! mailinfo-style parser. libgit2's email API is output-only
//! (`git_email_create_from_commit`); there is no `git_mailinfo` equivalent. The
//! gitoxide monorepo has no `gix-patch` crate, not even as a placeholder. No
//! production-quality standalone Rust mbox/git-patch parser crate exists.
//!
//! The genuinely hard parts of RFC 2822 parsing (header folding, RFC 2047 MIME
//! encoded-words for non-ASCII author names and subjects) are delegated to the
//! `mailparse` crate. The git-specific overlay (mbox envelope line, `[PATCH]`
//! prefix stripping, commit-message body extraction up to the `---` diffstat
//! separator) is implemented here, matching the behaviour of `git am`'s
//! `patchbreak()` function in `mailinfo.c`.
//!
//! ## If edge cases are reported
//!
//! If real-world patches produce incorrect metadata through this parser, the
//! escape hatch is to shell out to `git mailinfo` directly:
//! ```text
//! git mailinfo /tmp/msg /tmp/patch < input.patch
//! ```
//! This prints `Author:`, `Email:`, `Subject:`, `Date:` to stdout and writes
//! the commit body to `/tmp/msg`. Since ngit already requires `git` in PATH (it
//! is a git plugin), this is always available. It is not the primary approach
//! because it requires two temp files and a process spawn per patch, which is
//! acceptable cost but unnecessary given that most patches in the ngit `pr/`
//! flow will have the optional nostr tags and never reach this code.
//!
//! ## Known limitation: `---` in commit message body
//!
//! The `---` line that separates the commit message from the diffstat is
//! ambiguous when the commit message itself contains `---` (e.g. Markdown
//! horizontal rules). This parser stops at the first `---`-only line, matching
//! git am's own behaviour — `git am` has the same limitation and documents it.
//! This is not a bug we can fix without lookahead into the diff structure.
//!
//! ## Commit ID from mbox envelope
//!
//! The SHA1 in the mbox `From <sha1> <date>` envelope line is extracted but
//! **must not be assumed correct**. libgit2 generates this ID from the commit
//! object, but if the original commit was GPG-signed, or if the patch was
//! generated by a different tool, the reconstructed commit (applied via
//! `apply_to_tree` + `commit_create_buffer`) will have a different OID.
//! The `commit` nostr tag is the authoritative source for commit identity when
//! present.

use anyhow::{Context, Result, bail};
use chrono::DateTime;
use mailparse::{MailHeaderMap, parse_headers};

#[derive(Debug, Clone, PartialEq)]
pub struct PatchMetadata {
    pub commit_id: String,
    pub author_name: String,
    pub author_email: String,
    pub author_timestamp: i64,
    pub author_offset_minutes: i32,
    pub committer_timestamp: Option<i64>,
    pub subject: String,
    pub body: String,
}

pub fn parse_mbox_patch(content: &str) -> Result<PatchMetadata> {
    let commit_id = extract_commit_id_from_mbox(content)?;
    let (author_name, author_email) = extract_author_from_from_header(content)?;
    let (author_timestamp, author_offset_minutes) = extract_date_from_header(content)?;
    let committer_timestamp = None;
    let subject = extract_subject(content)?;
    let body = extract_commit_message_body(content)?;

    Ok(PatchMetadata {
        commit_id,
        author_name,
        author_email,
        author_timestamp,
        author_offset_minutes,
        committer_timestamp,
        subject,
        body,
    })
}

/// Extract the SHA1 from the mbox `From <sha1> <date>` envelope line.
///
/// **This value should not be assumed correct for the reconstructed commit.**
/// If the original commit was GPG-signed, or the patch was generated by a
/// different tool (e.g. `git format-patch` vs libgit2), the commit recreated
/// by applying this patch via `commit_create_buffer` will have a different OID.
/// Use the `commit` nostr event tag as the authoritative commit identity when
/// present.
fn extract_commit_id_from_mbox(content: &str) -> Result<String> {
    if !content.starts_with("From ") {
        bail!("patch does not start with 'From ' - not a valid mbox format");
    }

    let first_line = content.lines().next().context("patch content is empty")?;

    let parts: Vec<&str> = first_line.split_whitespace().collect();
    if parts.len() < 2 {
        bail!("mbox 'From ' line does not contain a commit id");
    }

    Ok(parts[1].to_string())
}

/// Extract the header section from the mbox content (everything after the first
/// line up to the first blank line that ends the headers).
fn extract_header_section(content: &str) -> &str {
    // Skip the mbox envelope line ("From <sha> <date>"), then pass the rest
    // to mailparse which understands where headers end.
    let after_envelope = content
        .find('\n')
        .map(|pos| &content[pos + 1..])
        .unwrap_or("");
    // Return only up to (and including) the blank line that terminates headers,
    // so mailparse doesn't try to parse the diff body.
    let header_end = after_envelope
        .find("\n\n")
        .map(|pos| pos + 2)
        .unwrap_or(after_envelope.len());
    &after_envelope[..header_end]
}

fn extract_author_from_from_header(content: &str) -> Result<(String, String)> {
    let header_bytes = extract_header_section(content).as_bytes();
    if let Ok((headers, _)) = parse_headers(header_bytes) {
        if let Some(from_value) = headers.get_first_value("From") {
            return parse_from_header_value(&from_value);
        }
    }

    // Fallback: manual search
    let from_line = content
        .lines()
        .find(|line| line.starts_with("From:"))
        .context("patch does not contain a 'From:' header")?;

    let from_value = from_line
        .strip_prefix("From:")
        .context("failed to strip 'From:' prefix")?
        .trim();

    parse_from_header_value(from_value)
}

fn parse_from_header_value(value: &str) -> Result<(String, String)> {
    if let Some(start) = value.find('<') {
        if let Some(end) = value.find('>') {
            let email = value[start + 1..end].to_string();
            let name_part = value[..start].trim();
            let name = name_part.trim_matches('"').trim().to_string();
            return Ok((name, email));
        }
    }

    if value.contains('@') {
        let email = value.trim().to_string();
        let name = email.split('@').next().unwrap_or("unknown").to_string();
        return Ok((name, email));
    }

    bail!("could not parse From header: {}", value)
}

fn extract_date_from_header(content: &str) -> Result<(i64, i32)> {
    let date_line = content
        .lines()
        .find(|line| line.starts_with("Date:"))
        .context("patch does not contain a 'Date:' header")?;

    let date_value = date_line
        .strip_prefix("Date:")
        .context("failed to strip 'Date:' prefix")?
        .trim();

    parse_rfc2822_date(date_value)
}

fn parse_rfc2822_date(value: &str) -> Result<(i64, i32)> {
    let parsed = DateTime::parse_from_rfc2822(value)
        .context(format!("failed to parse RFC2822 date: {}", value))?;

    let timestamp = parsed.timestamp();
    let offset_minutes = parsed.offset().local_minus_utc() / 60;

    Ok((timestamp, offset_minutes))
}

fn extract_subject(content: &str) -> Result<String> {
    // Use mailparse to handle RFC 2047 encoded-words and RFC 2822 header folding.
    let header_bytes = extract_header_section(content).as_bytes();
    if let Ok((headers, _)) = parse_headers(header_bytes) {
        if let Some(subject_value) = headers.get_first_value("Subject") {
            return Ok(cleanup_subject(&subject_value));
        }
    }

    // Fallback: manual single-line extraction.
    let subject_line = content
        .lines()
        .find(|line| line.starts_with("Subject:"))
        .context("patch does not contain a 'Subject:' header")?;

    let subject_value = subject_line
        .strip_prefix("Subject:")
        .context("failed to strip 'Subject:' prefix")?
        .trim();

    Ok(cleanup_subject(subject_value))
}

fn cleanup_subject(subject: &str) -> String {
    let mut result = subject.to_string();

    loop {
        let trimmed = result.trim();

        if trimmed.starts_with("Re:") || trimmed.starts_with("re:") {
            result = trimmed[3..].trim().to_string();
            continue;
        }

        if let Some(stripped) = trimmed.strip_prefix(':') {
            result = stripped.trim().to_string();
            continue;
        }

        if trimmed.starts_with('[') {
            if let Some(end) = trimmed.find(']') {
                result = trimmed[end + 1..].trim().to_string();
                continue;
            }
        }

        break;
    }

    result
}

fn extract_commit_message_body(content: &str) -> Result<String> {
    let mut in_body = false;
    let mut body_lines: Vec<String> = Vec::new();
    let mut found_first_content = false;

    for line in content.lines() {
        if !in_body {
            if line.is_empty() {
                in_body = true;
            }
            continue;
        }

        if line.starts_with("diff --git ")
            || line.starts_with("Index: ")
            || line.starts_with("--- ")
            || line.starts_with("From ")
        {
            break;
        }

        if line.starts_with("---") && line.trim().eq("---") {
            break;
        }

        // The email signature separator is exactly "-- " (dash dash space, nothing
        // after). Lines that merely start with "-- " followed by other text are
        // body content.
        if line == "-- " {
            break;
        }

        if !found_first_content && line.trim().is_empty() {
            continue;
        }

        found_first_content = true;
        body_lines.push(line.to_string());
    }

    while body_lines.last().is_some_and(|l| l.trim().is_empty()) {
        body_lines.pop();
    }

    Ok(body_lines.join("\n").trim().to_string())
}

pub fn extract_description_from_patch(content: &str) -> Result<String> {
    let subject = extract_subject(content)?;
    let body = extract_commit_message_body(content)?;

    if body.is_empty() {
        Ok(subject)
    } else {
        Ok(format!("{}\n\n{}", subject, body))
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    fn sample_patch() -> String {
        "\
From 431b84edc0d2fa118d63faa3c2db9c73d630a5ae Mon Sep 17 00:00:00 2001
From: Joe Bloggs <joe.bloggs@pm.me>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] add t2.md

This is the commit message body.

It can have multiple lines.

---
 t2.md | 1 +
 1 file changed, 1 insertion(+)
 create mode 100644 t2.md

diff --git a/t2.md b/t2.md
new file mode 100644
index 0000000..a66525d
--- /dev/null
+++ b/t2.md
@@ -0,0 +1 @@
+some content1
\\ No newline at end of file
--
libgit2 1.9.2

"
        .to_string()
    }

    #[test]
    fn parse_commit_id() {
        let patch = sample_patch();
        let result = extract_commit_id_from_mbox(&patch).unwrap();
        assert_eq!(result, "431b84edc0d2fa118d63faa3c2db9c73d630a5ae");
    }

    #[test]
    fn parse_author() {
        let patch = sample_patch();
        let (name, email) = extract_author_from_from_header(&patch).unwrap();
        assert_eq!(name, "Joe Bloggs");
        assert_eq!(email, "joe.bloggs@pm.me");
    }

    #[test]
    fn parse_author_with_quoted_name() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: \"John (nickname) Doe\" <john.doe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: test

Body
";
        let (name, email) = extract_author_from_from_header(patch).unwrap();
        assert_eq!(name, "John (nickname) Doe");
        assert_eq!(email, "john.doe@example.com");
    }

    #[test]
    fn parse_author_email_only() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: john.doe@example.com
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: test

Body
";
        let (name, email) = extract_author_from_from_header(patch).unwrap();
        assert_eq!(name, "john.doe");
        assert_eq!(email, "john.doe@example.com");
    }

    #[test]
    fn parse_date() {
        let patch = sample_patch();
        let (timestamp, offset) = extract_date_from_header(&patch).unwrap();
        assert_eq!(timestamp, 0);
        assert_eq!(offset, 0);
    }

    #[test]
    fn parse_date_with_timezone() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0500
Subject: test

Body
";
        let (timestamp, offset) = extract_date_from_header(patch).unwrap();
        assert_eq!(timestamp, -18000);
        assert_eq!(offset, 300);
    }

    #[test]
    fn parse_subject() {
        let patch = sample_patch();
        let subject = extract_subject(&patch).unwrap();
        assert_eq!(subject, "add t2.md");
    }

    #[test]
    fn parse_subject_with_patch_prefix() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH v2 3/5] fix: important bug

Body
";
        let subject = extract_subject(patch).unwrap();
        assert_eq!(subject, "fix: important bug");
    }

    #[test]
    fn parse_subject_with_re_prefix() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: Re: [PATCH] fix: important bug

Body
";
        let subject = extract_subject(patch).unwrap();
        assert_eq!(subject, "fix: important bug");
    }

    #[test]
    fn parse_subject_folded_rfc2822() {
        // RFC 2822 header folding: continuation lines start with whitespace.
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] fix: this is a very long commit message subject line
 that has been folded across two lines by RFC 2822 rules

Body
";
        let subject = extract_subject(patch).unwrap();
        assert_eq!(
            subject,
            "fix: this is a very long commit message subject line that has been folded across two lines by RFC 2822 rules"
        );
    }

    #[test]
    fn parse_subject_mime_q_encoded() {
        // RFC 2047 Q-encoding: =?UTF-8?q?...?=
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] =?UTF-8?q?fix=3A_add_=E2=9C=93_check?=

Body
";
        let subject = extract_subject(patch).unwrap();
        // Q-decoded: "fix: add ✓ check"
        assert_eq!(subject, "fix: add \u{2713} check");
    }

    #[test]
    fn parse_subject_mime_b_encoded() {
        // RFC 2047 B-encoding: =?UTF-8?b?...?= (base64)
        // "fix: résumé" base64 encoded
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] =?UTF-8?b?Zml4OiByw6lzdW3DqQ==?=

Body
";
        let subject = extract_subject(patch).unwrap();
        // B-decoded: "fix: résumé"
        assert_eq!(subject, "fix: r\u{e9}sum\u{e9}");
    }

    #[test]
    fn parse_body() {
        let patch = sample_patch();
        let body = extract_commit_message_body(&patch).unwrap();
        assert_eq!(
            body,
            "This is the commit message body.\n\nIt can have multiple lines."
        );
    }

    #[test]
    fn parse_body_empty() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: test

---
 file.txt | 1 +
diff --git a/file.txt b/file.txt
";
        let body = extract_commit_message_body(patch).unwrap();
        assert_eq!(body, "");
    }

    #[test]
    fn parse_body_stops_at_exact_email_sig_separator() {
        // "-- " (dash dash space, nothing after) is the email sig separator.
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] test

This is the body.
-- 
libgit2 1.9.2

diff --git a/file.txt b/file.txt
";
        let body = extract_commit_message_body(patch).unwrap();
        assert_eq!(body, "This is the body.");
    }

    #[test]
    fn parse_body_does_not_stop_at_double_dash_with_text() {
        // "-- some text" must NOT be treated as an email sig separator.
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] test

This is the body.
-- some CLI flag description
More body text.

---
diff --git a/file.txt b/file.txt
";
        let body = extract_commit_message_body(patch).unwrap();
        assert_eq!(
            body,
            "This is the body.\n-- some CLI flag description\nMore body text."
        );
    }

    #[test]
    fn parse_full_metadata() {
        let patch = sample_patch();
        let metadata = parse_mbox_patch(&patch).unwrap();

        assert_eq!(
            metadata.commit_id,
            "431b84edc0d2fa118d63faa3c2db9c73d630a5ae"
        );
        assert_eq!(metadata.author_name, "Joe Bloggs");
        assert_eq!(metadata.author_email, "joe.bloggs@pm.me");
        assert_eq!(metadata.author_timestamp, 0);
        assert_eq!(metadata.author_offset_minutes, 0);
        assert_eq!(metadata.committer_timestamp, None);
        assert_eq!(metadata.subject, "add t2.md");
        assert_eq!(
            metadata.body,
            "This is the commit message body.\n\nIt can have multiple lines."
        );
    }

    #[test]
    fn extract_description_combines_subject_and_body() {
        let patch = sample_patch();
        let description = extract_description_from_patch(&patch).unwrap();
        assert_eq!(
            description,
            "add t2.md\n\nThis is the commit message body.\n\nIt can have multiple lines."
        );
    }

    #[test]
    fn extract_description_subject_only() {
        let patch = "\
From abc123 Mon Sep 17 00:00:00 2001
From: Joe <joe@example.com>
Date: Thu, 1 Jan 1970 00:00:00 +0000
Subject: [PATCH] simple fix

---
 file.txt | 1 +
";
        let description = extract_description_from_patch(patch).unwrap();
        assert_eq!(description, "simple fix");
    }

    #[test]
    fn cleanup_subject_strips_patch_prefixes() {
        assert_eq!(cleanup_subject("[PATCH] test"), "test");
        assert_eq!(cleanup_subject("[PATCH v2] test"), "test");
        assert_eq!(cleanup_subject("[PATCH 1/3] test"), "test");
        assert_eq!(cleanup_subject("[PATCH v2 1/3] test"), "test");
        assert_eq!(cleanup_subject("Re: [PATCH] test"), "test");
        assert_eq!(cleanup_subject("re: test"), "test");
        assert_eq!(cleanup_subject(":test"), "test");
    }
}