gitks/diff/get_diff.rs

use std::collections::HashMap;

use crate::bare::GitBare;
use crate::diff::get_diff_stats::{diff_stats_for_range, push_diff_options};
use crate::error::{GitError, GitResult};
use crate::paginate;
use crate::pb::diff_file::ChangeType;
use crate::pb::{DiffFile, GetDiffRequest, GetDiffResponse};

/// Parsed entry from `git diff --raw -z`
struct RawDiffEntry {
    status: char,
    old_path: String,
    new_path: String,
    old_mode: u32,
    new_mode: u32,
    old_oid: String,
    new_oid: String,
    similarity: f64,
}

/// Type alias for diff raw output: (entries, numstat_map)
type DiffRawOutput = (Vec<RawDiffEntry>, HashMap<String, (u32, u32, bool)>);

impl GitBare {
    pub fn get_diff(&self, request: GetDiffRequest) -> GitResult<GetDiffResponse> {
        let base = match request.base.and_then(|s| s.selector) {
            Some(crate::pb::object_selector::Selector::Oid(oid)) => oid.hex.clone(),
            Some(crate::pb::object_selector::Selector::Revision(name)) => {
                crate::sanitize::validate_revision(&name.revision)?;
                name.revision.clone()
            }
            None => "HEAD".into(),
        };
        let head = match request.head.and_then(|s| s.selector) {
            Some(crate::pb::object_selector::Selector::Oid(oid)) => oid.hex.clone(),
            Some(crate::pb::object_selector::Selector::Revision(name)) => {
                crate::sanitize::validate_revision(&name.revision)?;
                name.revision.clone()
            }
            None => "HEAD".into(),
        };
        tracing::debug!(
            repo = %self.bare_dir.display(),
            base = %base,
            head = %head,
            "computing diff"
        );

        let options = request.options.as_ref();
        let want_patch = options.is_some_and(|o| o.include_patch);

        let (raw_entries, numstat_map) = self.diff_raw_and_numstat(&base, &head, options)?;

        let max_files = options.and_then(|o| (o.max_files > 0).then_some(o.max_files as usize));
        let overflow = max_files.is_some_and(|max| raw_entries.len() > max);
        let entries_to_build = max_files.map_or(raw_entries.as_slice(), |max| {
            &raw_entries[..raw_entries.len().min(max)]
        });

        let patch_map = if want_patch {
            self.diff_patch_batch(&base, &head, options)?
        } else {
            HashMap::new()
        };

        let mut files = Vec::with_capacity(entries_to_build.len());
        for entry in entries_to_build {
            let path = if !entry.new_path.is_empty() {
                &entry.new_path
            } else {
                &entry.old_path
            };
            let (additions, deletions, binary) = numstat_map
                .get(path)
                .map(|(a, d, b)| (*a, *d, *b))
                .unwrap_or((0, 0, false));

            let too_large = options.is_some_and(|o| {
                o.max_bytes > 0
                    && patch_map
                        .get(path)
                        .is_some_and(|p: &Vec<u8>| p.len() > o.max_bytes as usize)
            });
            let patch = patch_map
                .get(path)
                .map(|p| {
                    let max = options.map(|o| o.max_bytes as usize).unwrap_or(0);
                    if too_large && max > 0 {
                        p[..max].to_vec()
                    } else {
                        p.clone()
                    }
                })
                .unwrap_or_default();

            files.push(DiffFile {
                old_path: entry.old_path.clone(),
                new_path: entry.new_path.clone(),
                old_oid: if !entry.old_oid.is_empty()
                    && entry.old_oid != "0000000000000000000000000000000000000000"
                {
                    Some(self.oid_to_pb(&entry.old_oid))
                } else {
                    None
                },
                new_oid: if !entry.new_oid.is_empty()
                    && entry.new_oid != "0000000000000000000000000000000000000000"
                {
                    Some(self.oid_to_pb(&entry.new_oid))
                } else {
                    None
                },
                old_mode: entry.old_mode,
                new_mode: entry.new_mode,
                change_type: change_type(entry.status) as i32,
                binary,
                too_large,
                additions,
                deletions,
                hunks: Vec::new(),
                patch,
                similarity: entry.similarity,
            });
        }

        let stats = diff_stats_for_range(self, &base, &head, options)?;
        let (files, page_info) = paginate::paginate(&files, request.pagination.as_ref());

        Ok(GetDiffResponse {
            files,
            stats: Some(stats),
            page_info: Some(page_info),
            overflow,
        })
    }

    /// Single subprocess call that gets BOTH --raw and --numstat with -z.
    /// Returns parsed raw entries and a map of path → (additions, deletions, binary).
    ///
    /// Combined output format with -z (NUL-separated records):
    ///   :<src_mode> <dst_mode> <src_hash> <dst_hash> <status>\0<path>\0
    ///   (for R/C: ...\0<old_path>\0<new_path>\0)
    ///   Then numstat records: <add>\t<del>\t<path>\0
    fn diff_raw_and_numstat(
        &self,
        base: &str,
        head: &str,
        options: Option<&crate::pb::DiffOptions>,
    ) -> GitResult<DiffRawOutput> {
        let mut args = vec![
            "--git-dir".to_string(),
            self.bare_dir.to_string_lossy().into_owned(),
            "diff".into(),
            "--raw".into(),
            "--numstat".into(),
            "-z".into(),
        ];
        push_diff_options(&mut args, options);
        args.push(base.to_string());
        args.push(head.to_string());
        if let Some(options) = options
            && !options.pathspec.is_empty()
        {
            args.push("--".into());
            args.extend(options.pathspec.iter().cloned());
        }

        let result = duct::cmd("git", &args)
            .stdout_capture()
            .stderr_capture()
            .unchecked()
            .run()?;
        if !result.status.success() {
            return Err(GitError::CommandFailed {
                status_code: result.status.code(),
                stderr: String::from_utf8_lossy(&result.stderr).into_owned(),
            });
        }

        // Split by NUL — each record is NUL-terminated
        let records: Vec<&[u8]> = result.stdout.split(|b| *b == 0).collect();

        let mut raw_entries = Vec::new();
        let mut numstat_map: HashMap<String, (u32, u32, bool)> = HashMap::new();
        let mut i = 0;

        while i < records.len() {
            let record = records[i];
            if record.is_empty() {
                i += 1;
                continue;
            }

            if record.starts_with(b":") {
                // Raw meta record: ":<src_mode> <dst_mode> <src_hash> <dst_hash> <status_char>"
                // In older git: tab before status. In newer git: space before status.
                // The path(s) follow as separate NUL-terminated records.
                let record_str = String::from_utf8_lossy(record).into_owned();

                // Try tab separator first (older git), then space (newer git)
                let (meta, status_str) = if let Some((m, s)) = record_str.rsplit_once('\t') {
                    (m, s)
                } else if let Some((m, s)) = record_str.rsplit_once(' ') {
                    (m, s)
                } else {
                    i += 1;
                    continue;
                };

                let meta_parts: Vec<&str> = meta.split_whitespace().collect();
                let old_mode = meta_parts
                    .first()
                    .and_then(|s| u32::from_str_radix(s, 8).ok())
                    .unwrap_or(0);
                let new_mode = meta_parts
                    .get(1)
                    .and_then(|s| u32::from_str_radix(s, 8).ok())
                    .unwrap_or(0);
                let old_oid = meta_parts.get(2).unwrap_or(&"").to_string();
                let new_oid = meta_parts.get(3).unwrap_or(&"").to_string();

                let status = status_str.chars().next().unwrap_or('M');
                let similarity = status_str
                    .get(1..)
                    .and_then(|s| s.parse::<f64>().ok())
                    .unwrap_or(0.0);

                // Read path record(s) that follow the meta record
                let (old_path, new_path) = match status {
                    'R' | 'C' => {
                        // Two path records: old_path\0new_path\0
                        let op = if i + 1 < records.len() {
                            i += 1;
                            String::from_utf8_lossy(records[i]).into_owned()
                        } else {
                            String::new()
                        };
                        let np = if i + 1 < records.len() {
                            i += 1;
                            String::from_utf8_lossy(records[i]).into_owned()
                        } else {
                            String::new()
                        };
                        (op, np)
                    }
                    'A' => {
                        let p = if i + 1 < records.len() {
                            i += 1;
                            String::from_utf8_lossy(records[i]).into_owned()
                        } else {
                            String::new()
                        };
                        (String::new(), p)
                    }
                    'D' => {
                        let p = if i + 1 < records.len() {
                            i += 1;
                            String::from_utf8_lossy(records[i]).into_owned()
                        } else {
                            String::new()
                        };
                        (p, String::new())
                    }
                    _ => {
                        let p = if i + 1 < records.len() {
                            i += 1;
                            String::from_utf8_lossy(records[i]).into_owned()
                        } else {
                            String::new()
                        };
                        (p.clone(), p)
                    }
                };

                raw_entries.push(RawDiffEntry {
                    status,
                    old_path,
                    new_path,
                    old_mode,
                    new_mode,
                    old_oid,
                    new_oid,
                    similarity,
                });
            } else {
                // Numstat record: "<add>\t<del>\t<path>"
                let record_str = String::from_utf8_lossy(record);
                let parts: Vec<&str> = record_str.split('\t').collect();
                if parts.len() >= 3 {
                    let binary = parts[0] == "-" || parts[1] == "-";
                    let add = parts[0].parse().unwrap_or(0u32);
                    let del = parts[1].parse().unwrap_or(0u32);
                    let path = parts[2].to_string();
                    numstat_map.insert(path, (add, del, binary));
                }
            }
            i += 1;
        }

        Ok((raw_entries, numstat_map))
    }

    /// Single subprocess call to get patches for ALL files at once.
    /// Returns a map of path → patch bytes.
    fn diff_patch_batch(
        &self,
        base: &str,
        head: &str,
        options: Option<&crate::pb::DiffOptions>,
    ) -> GitResult<HashMap<String, Vec<u8>>> {
        let context = options
            .map(|o| o.context_lines.to_string())
            .unwrap_or_else(|| "3".into());
        let mut args = vec![
            "--git-dir".to_string(),
            self.bare_dir.to_string_lossy().into_owned(),
            "diff".into(),
            "--patch".into(),
            format!("--unified={context}"),
        ];
        if options.is_some_and(|o| o.include_binary) {
            args.push("--binary".into());
        }
        push_diff_options(&mut args, options);
        args.push(base.to_string());
        args.push(head.to_string());
        if let Some(options) = options
            && !options.pathspec.is_empty()
        {
            args.push("--".into());
            args.extend(options.pathspec.iter().cloned());
        }

        let result = duct::cmd("git", &args)
            .stdout_capture()
            .stderr_capture()
            .unchecked()
            .run()?;
        if !result.status.success() {
            return Err(GitError::CommandFailed {
                status_code: result.status.code(),
                stderr: String::from_utf8_lossy(&result.stderr).into_owned(),
            });
        }

        // Split combined patch output by "diff --git" headers
        let mut map = HashMap::new();
        let output = &result.stdout;
        let header = b"diff --git ";
        let mut chunks: Vec<&[u8]> = Vec::new();
        let mut pos = 0;

        // Find all header positions
        let mut header_positions = Vec::new();
        while let Some(idx) = output[pos..]
            .windows(header.len())
            .position(|w| w == header)
        {
            header_positions.push(pos + idx);
            pos = pos + idx + header.len();
        }

        for (i, &start) in header_positions.iter().enumerate() {
            let end = header_positions.get(i + 1).copied().unwrap_or(output.len());
            chunks.push(&output[start..end]);
        }

        for chunk in chunks {
            // Extract file path from "diff --git a/path b/path\n"
            let first_line_end = chunk
                .iter()
                .position(|&b| b == b'\n')
                .unwrap_or(chunk.len());
            let first_line = String::from_utf8_lossy(&chunk[..first_line_end]);
            if let Some(b_pos) = first_line.rfind(" b/") {
                let path = &first_line[b_pos + 3..];
                map.insert(path.to_string(), chunk.to_vec());
            }
        }

        Ok(map)
    }
}

fn change_type(status: char) -> ChangeType {
    match status {
        'A' => ChangeType::DiffFileChangeTypeAdded,
        'D' => ChangeType::DiffFileChangeTypeDeleted,
        'R' => ChangeType::DiffFileChangeTypeRenamed,
        'C' => ChangeType::DiffFileChangeTypeCopied,
        'T' => ChangeType::DiffFileChangeTypeTypeChanged,
        'U' => ChangeType::DiffFileChangeTypeUnmerged,
        _ => ChangeType::DiffFileChangeTypeModified,
    }
}