From 5e3adeadf40b29f63e0e8236e120aeae124ff82a Mon Sep 17 00:00:00 2001
From: Jeff Carr <jcarr@wit.com>
Date: Mon, 6 Oct 2025 19:35:26 -0500
Subject: gen 'git patch-id' from raw bytes

---
 patchid.go | 135 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 131 insertions(+), 4 deletions(-)

(limited to 'patchid.go')

diff --git a/patchid.go b/patchid.go
index 28c8313..72761a9 100644
--- a/patchid.go
+++ b/patchid.go
@@ -1,18 +1,18 @@
 package gitpb
 
 import (
+	"bufio"
 	"bytes"
-	"errors"
+	"crypto/sha1"
+	"encoding/hex"
 	"fmt"
-	"os"
+	"io"
 	"os/exec"
-	"path/filepath"
 	"strings"
 
 	"go.wit.com/log"
 )
 
-
 // git show 5b277e7686974d2195586d5f5b82838ee9ddb036 |git patch-id --stable
 // bf86be06af03b1a89ee155b214358362ec76f7b6 5b277e7686974d2195586d5f5b82838ee9ddb036
 // forged patch: "working on ping pong"
@@ -70,3 +70,130 @@ func (repo *Repo) FindPatchIdByHash(hash string) (string, error) {
 
 	return fields[0], nil
 }
+
+// ComputePatchID calculates the patch ID for a given patch file's content.
+// It mimics the behavior of the `git patch-id` command by normalizing the
+// patch content before hashing.
+//
+// Normalization rules:
+//  1. It ignores all lines until the first "--- " line is encountered.
+//  2. It ignores diff headers (lines starting with "---", "+++", "diff --git", "index").
+//  3. For content lines ('+', '-', or ' '), it keeps the first character and
+//     trims any trailing whitespace from the rest of the line before hashing.
+//  4. All other lines (like hunk headers "@@ ... @@") are ignored.
+func FindPatchIdByBytes(patchData io.Reader) (string, error) {
+	scanner := bufio.NewScanner(patchData)
+	var normalizedPatch bytes.Buffer
+	var inPatchBody bool
+
+	for scanner.Scan() {
+		line := scanner.Text()
+
+		// The patch content officially starts at the first `---` line.
+		// This helps skip email headers in patches generated by `git format-patch`.
+		if !inPatchBody && strings.HasPrefix(line, "---") {
+			inPatchBody = true
+		}
+
+		if !inPatchBody {
+			continue
+		}
+
+		// Skip headers and metadata lines
+		if strings.HasPrefix(line, "---") ||
+			strings.HasPrefix(line, "+++") ||
+			strings.HasPrefix(line, "index") ||
+			strings.HasPrefix(line, "diff --git") {
+			continue
+		}
+
+		// Process only the actual content lines (context, addition, deletion)
+		if len(line) > 0 && (line[0] == ' ' || line[0] == '+' || line[0] == '-') {
+			// Keep the first character (' ', '+', or '-')
+			firstChar := line[0]
+			content := line[1:]
+
+			// Trim trailing whitespace from the content part of the line
+			trimmedContent := strings.TrimRight(content, " ")
+
+			// Write the normalized line to our buffer for hashing
+			normalizedPatch.WriteByte(firstChar)
+			normalizedPatch.WriteString(trimmedContent)
+			normalizedPatch.WriteByte('\n')
+		}
+		// All other lines (e.g., "@@ ... @@") are ignored.
+	}
+
+	if err := scanner.Err(); err != nil {
+		return "", fmt.Errorf("error reading patch data: %w", err)
+	}
+
+	// Calculate the SHA-1 hash of the normalized patch content
+	hash := sha1.Sum(normalizedPatch.Bytes())
+	patchID := hex.EncodeToString(hash[:])
+
+	return patchID, nil
+}
+
+/*
+// --- Example Usage ---
+
+func main() {
+	// Example 1: Using a string as input
+	patchString := `
+From 2f42795678c7113f7437c04f56570c08b68850f4 Mon Sep 17 00:00:00 2001
+From: A Developer <dev@example.com>
+Subject: [PATCH] Example change to demonstrate patch-id
+
+---	a/README.md
++++ b/README.md
+@@ -1,3 +1,4 @@
+ # My Project
+
+-This is a sample project.
++This is a sample project.
++It has an additional line.
+`
+	stringReader := strings.NewReader(patchString)
+	patchID, err := ComputePatchID(stringReader)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error computing patch ID from string: %v\n", err)
+		return
+	}
+	fmt.Printf("Patch ID from string: %s\n", patchID)
+
+	// Example 2: Reading from a file
+	// Create a temporary patch file for demonstration
+	patchContent := []byte(patchString)
+	tmpfile, err := os.CreateTemp("", "example.*.patch")
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error creating temp file: %v\n", err)
+		return
+	}
+	defer os.Remove(tmpfile.Name()) // clean up
+
+	if _, err := tmpfile.Write(patchContent); err != nil {
+		fmt.Fprintf(os.Stderr, "Error writing to temp file: %v\n", err)
+		return
+	}
+	if err := tmpfile.Close(); err != nil {
+		fmt.Fprintf(os.Stderr, "Error closing temp file: %v\n", err)
+		return
+	}
+
+	// Open the file to pass to our function
+	file, err := os.Open(tmpfile.Name())
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error opening temp file: %v\n", err)
+		return
+	}
+	defer file.Close()
+
+	patchIDFromFile, err := ComputePatchID(file)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error computing patch ID from file: %v\n", err)
+		return
+	}
+	fmt.Printf("Patch ID from file:   %s\n", patchIDFromFile)
+}
+*/
-- 
cgit v1.2.3