{"TrimRightFunc", TrimRightFunc, tc.rightOut},
		}
		for _, trimmer := range trimmers {
			actual := trimmer.trim([]byte(tc.in), tc.f.f)
			if actual == nil && trimmer.out != nil {
				t.Errorf("%s(%q, %q) = nil; want %q", trimmer.name, tc.in, tc.f.name, trimmer.out)
			}
			if actual != nil && trimmer.out == nil {
				t.Errorf("%s(%q, %q) = %q; want nil", trimmer.name, tc.in, tc.f.name, actual)
			}

//			log.Fatal(err)
//		}
//
//		biggerZip := makeZip("big.zip", bytes.NewReader(bigZip))
//		if err := os.WriteFile("/tmp/bigger.zip", biggerZip, 0666); err != nil {
//			log.Fatal(err)
//		}
//
//		biggestZip := makeZip("bigger.zip", bytes.NewReader(biggerZip))
//		if err := os.WriteFile("/tmp/biggest.zip", biggestZip, 0666); err != nil {
//			log.Fatal(err)
//		}
//	}
//

package bytes_test

import (
	. "bytes"
	"fmt"
	"internal/testenv"
	"io"
	"math/rand"
	"strconv"
	"testing"
	"unicode/utf8"
)

const N = 10000       // make this bigger for a larger (and slower) test
var testString string // test data for write tests
var testBytes []byte  // test data; same as testString but as a slice.

type negativeReader struct{}

	// Do this because Seek is often lazy about reporting errors; this will mask
	// the fact that the stream may be truncated. We can rely on the
	// io.CopyN done shortly afterwards to trigger any IO errors.
	var seekSkipped int64 // Number of bytes skipped via Seek
	if sr, ok := r.(io.Seeker); ok && n > 1 {
		// Not all io.Seeker can actually Seek. For example, os.Stdin implements

		t.Fatal(err)
	}
	if hdr.Linkname != longLinkname {
		t.Fatal("Couldn't recover long link name")
	}
}

func TestPaxNonAscii(t *testing.T) {
	// Create an archive with non ascii. These should trigger a pax header
	// because pax headers have a defined utf-8 encoding.
	fileinfo, err := os.Stat("testdata/small.txt")
	if err != nil {
		t.Fatal(err)
	}

	hdr, err := FileInfoHeader(fileinfo, "")