fix: handle Windows CRLF line endings in grep tool (#5948)

Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
lif 2025-12-23 12:26:15 +08:00 committed by GitHub
parent eab177f5e7
commit 5af35117db
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 114 additions and 3 deletions

View file

@ -240,7 +240,8 @@ export namespace Ripgrep {
if (done) break
buffer += decoder.decode(value, { stream: true })
const lines = buffer.split("\n")
// Handle both Unix (\n) and Windows (\r\n) line endings
const lines = buffer.split(/\r?\n/)
buffer = lines.pop() || ""
for (const line of lines) {
@ -379,7 +380,8 @@ export namespace Ripgrep {
return []
}
const lines = result.text().trim().split("\n").filter(Boolean)
// Handle both Unix (\n) and Windows (\r\n) line endings
const lines = result.text().trim().split(/\r?\n/).filter(Boolean)
// Parse JSON lines from ripgrep output
return lines

View file

@ -49,7 +49,8 @@ export const GrepTool = Tool.define("grep", {
throw new Error(`ripgrep failed: ${errorOutput}`)
}
const lines = output.trim().split("\n")
// Handle both Unix (\n) and Windows (\r\n) line endings
const lines = output.trim().split(/\r?\n/)
const matches = []
for (const line of lines) {

View file

@ -0,0 +1,108 @@
import { describe, expect, test } from "bun:test"
import path from "path"
import { GrepTool } from "../../src/tool/grep"
import { Instance } from "../../src/project/instance"
import { tmpdir } from "../fixture/fixture"
const ctx = {
sessionID: "test",
messageID: "",
callID: "",
agent: "build",
abort: AbortSignal.any([]),
metadata: () => {},
}
const projectRoot = path.join(__dirname, "../..")
describe("tool.grep", () => {
test("basic search", async () => {
await Instance.provide({
directory: projectRoot,
fn: async () => {
const grep = await GrepTool.init()
const result = await grep.execute(
{
pattern: "export",
path: path.join(projectRoot, "src/tool"),
include: "*.ts",
},
ctx,
)
expect(result.metadata.matches).toBeGreaterThan(0)
expect(result.output).toContain("Found")
},
})
})
test("no matches returns correct output", async () => {
await using tmp = await tmpdir({
init: async (dir) => {
await Bun.write(path.join(dir, "test.txt"), "hello world")
},
})
await Instance.provide({
directory: tmp.path,
fn: async () => {
const grep = await GrepTool.init()
const result = await grep.execute(
{
pattern: "xyznonexistentpatternxyz123",
path: tmp.path,
},
ctx,
)
expect(result.metadata.matches).toBe(0)
expect(result.output).toBe("No files found")
},
})
})
test("handles CRLF line endings in output", async () => {
// This test verifies the regex split handles both \n and \r\n
await using tmp = await tmpdir({
init: async (dir) => {
// Create a test file with content
await Bun.write(path.join(dir, "test.txt"), "line1\nline2\nline3")
},
})
await Instance.provide({
directory: tmp.path,
fn: async () => {
const grep = await GrepTool.init()
const result = await grep.execute(
{
pattern: "line",
path: tmp.path,
},
ctx,
)
expect(result.metadata.matches).toBeGreaterThan(0)
},
})
})
})
describe("CRLF regex handling", () => {
test("regex correctly splits Unix line endings", () => {
const unixOutput = "file1.txt|1|content1\nfile2.txt|2|content2\nfile3.txt|3|content3"
const lines = unixOutput.trim().split(/\r?\n/)
expect(lines.length).toBe(3)
expect(lines[0]).toBe("file1.txt|1|content1")
expect(lines[2]).toBe("file3.txt|3|content3")
})
test("regex correctly splits Windows CRLF line endings", () => {
const windowsOutput = "file1.txt|1|content1\r\nfile2.txt|2|content2\r\nfile3.txt|3|content3"
const lines = windowsOutput.trim().split(/\r?\n/)
expect(lines.length).toBe(3)
expect(lines[0]).toBe("file1.txt|1|content1")
expect(lines[2]).toBe("file3.txt|3|content3")
})
test("regex handles mixed line endings", () => {
const mixedOutput = "file1.txt|1|content1\nfile2.txt|2|content2\r\nfile3.txt|3|content3"
const lines = mixedOutput.trim().split(/\r?\n/)
expect(lines.length).toBe(3)
})
})