archive/zip: Error zip reader test

[gopherbot]

by iacob.campia:

By curiosity I was looking over the source code and spotted an error in the
/archive/zip/reader_test.go that when fixed, the test fails.

https://code.google.com/p/go/source/browse/src/pkg/archive/zip/reader_test.go#358


340:  func readTestFile(t *testing.T, zt ZipTest, ft ZipTestFile, f *File) {
       // ....

358:    size0 := f.UncompressedSize

        var b bytes.Buffer
        r, err := f.Open()
        if err != nil {
                t.Errorf("%s: %v", zt.Name, err)
                return
        }

367:    if size1 := f.UncompressedSize; size0 != size1 {
                t.Errorf("file %q changed f.UncompressedSize from %d to %d", f.Name, size0, size1)
        }

As seen in there, first it saves UncompresedSize value in size0 then compares it with
itself as size1 and not to the length of the uncompressed bytes.

If you fix it, the test will fail with:

--- FAIL: TestReader (0.01 seconds)
        reader_test.go:367: file "README" changed f.UncompressedSize from 4294967295 to 36

Yhe value of UncompresedSize is actually 4294967295 bytes and of the uncompressed data
36 bytes.

[adg]

Comment 1:

I don't see the error; the purpose of the test is to check that the UncompressedSize
field does not change after f.Open is called. This is because first UncompressedSize is
first populated by the directory header (at the end of the zip file), and then updated
as Open reads the actual file header (present just before the file body).
What's your "fix"?

Owner changed to @adg.

Status changed to WaitingForReply.

[gopherbot]

Comment 2 by iacob.campia:

f.Open() doesn't touch anything, it just calls flate decompress on the data.
I thought the purpose of the test was that the value of the UncompressedSize as written
in the headers is the same as the actual size of the uncompressed data.
The fix would be to move the check after the _, err = io.Copy(&b, r) and compare len(b)
with f.size() witch returns UncompresedSize64 if > 0 or UncompresedSize if not. and not
UncompressedSize which is useless for 64 archives because it contains 0xFFFFFF. This
test fails on zip64.zip

[adg]

Comment 3:

Ah. Originally, f.Open did re-read the header, but it seems it doesn't do that now.
The failure message you're seeing is for tests of ZIP64 files, which populates the
UncompressedSize64 field instead.
Fix: https://golang.org/cl/61650046

Status changed to Started.

[adg]

Comment 4:

This issue was closed by revision 413e28d.

Status changed to Fixed.