From 0e489b52cafe7f4c26582cedc69c0a6cb97d243b Mon Sep 17 00:00:00 2001
From: Dmitry Vyukov <dvyukov@google.com>
Date: Sun, 13 Oct 2024 14:32:35 +0200
Subject: [PATCH] hunk: add context

Add Hunk.Context field.
Context is the optional part at the end of the @@ line.
Usually it contains the changed function name,
so it may be useful to e.g. understand that all
hunks belong to a single function.
---
 parser.go                         | 6 ++++--
 testdata/significant/add.diff     | 2 +-
 testdata/significant/add.diff.msg | 2 +-
 3 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/parser.go b/parser.go
index 70cdcdd..f547d5b 100644
--- a/parser.go
+++ b/parser.go
@@ -33,7 +33,7 @@ type ChangeList []ContentChange
 // Hunk is a line that starts with @@.
 // Each hunk shows one area where the files differ
 // Unified format hunks look like this:
-// @@ from-file-line-numbers to-file-line-numbers @@
+// @@ from-file-line-numbers to-file-line-numbers @@( context)?
 //
 //	line-from-either-file
 //	line-from-either-file…
@@ -45,6 +45,7 @@ type Hunk struct {
 	CountOld           int        `json:"count_old"`
 	StartLineNumberNew int        `json:"start_line_number_new"`
 	CountNew           int        `json:"count_new"`
+	Context            string     `json:"context,omitempty"`
 }
 
 func (changes *ChangeList) IsSignificant() bool {
@@ -57,7 +58,7 @@ func (changes *ChangeList) IsSignificant() bool {
 }
 
 func NewHunk(line string) (Hunk, error) {
-	namedHunkRegex := regexp.MustCompile(`(?m)^@@ -(?P<start_old>\d+),?(?P<count_old>\d+)? \+(?P<start_new>\d+),?(?P<count_new>\d+)? @@`)
+	namedHunkRegex := regexp.MustCompile(`(?m)^@@ -(?P<start_old>\d+),?(?P<count_old>\d+)? \+(?P<start_new>\d+),?(?P<count_new>\d+)? @@ ?(?P<context>.*)?`)
 	match := namedHunkRegex.FindStringSubmatch(line)
 	result := make(map[string]string)
 	for i, name := range namedHunkRegex.SubexpNames() {
@@ -86,6 +87,7 @@ func NewHunk(line string) (Hunk, error) {
 		CountOld:           countOld,
 		StartLineNumberNew: startLineNumberNew,
 		CountNew:           countNew,
+		Context:            result["context"],
 	}, nil
 }
 
diff --git a/testdata/significant/add.diff b/testdata/significant/add.diff
index b40e2de..f8ff1db 100644
--- a/testdata/significant/add.diff
+++ b/testdata/significant/add.diff
@@ -3,5 +3,5 @@ new file mode 100644
 index 0000000..7898192
 --- /dev/null
 +++ b/a.txt
-@@ -0,0 +1 @@
+@@ -0,0 +1 @@ test context
 +a
diff --git a/testdata/significant/add.diff.msg b/testdata/significant/add.diff.msg
index 67952c4..4826ec3 100755
--- a/testdata/significant/add.diff.msg
+++ b/testdata/significant/add.diff.msg
@@ -1 +1 @@
-significant diff &git_diff_parser.FileDiff{FromFile:"a.txt", ToFile:"a.txt", Type:"modify", IsBinary:false, NewMode:"100644", Hunks:[]git_diff_parser.Hunk{git_diff_parser.Hunk{ChangeList:git_diff_parser.ChangeList{git_diff_parser.ContentChange{Type:"add", From:"", To:"a"}, git_diff_parser.ContentChange{Type:"", From:"", To:""}}, StartLineNumberOld:0, CountOld:0, StartLineNumberNew:1, CountNew:1}}, BinaryPatch:[]git_diff_parser.BinaryPatch(nil)}
\ No newline at end of file
+significant diff &git_diff_parser.FileDiff{FromFile:"a.txt", ToFile:"a.txt", Type:"modify", IsBinary:false, NewMode:"100644", Hunks:[]git_diff_parser.Hunk{git_diff_parser.Hunk{ChangeList:git_diff_parser.ChangeList{git_diff_parser.ContentChange{Type:"add", From:"", To:"a"}, git_diff_parser.ContentChange{Type:"", From:"", To:""}}, StartLineNumberOld:0, CountOld:0, StartLineNumberNew:1, CountNew:1, Context:"test context"}}, BinaryPatch:[]git_diff_parser.BinaryPatch(nil)}
\ No newline at end of file