Alexandre Julliard julliard@winehq.org wrote:
@@ -752,13 +752,13 @@ static void check_tiff_format(IStream *stream, const WICPixelFormatGUID *format) } else if (IsEqualGUID(format, &GUID_WICPixelFormat2bppIndexed)) {
ok(width == 32, "wrong width %u\n", width);
ok(width == 16, "wrong width %u\n", width); ok(height == 2, "wrong height %u\n", height);
ok(bps == 1, "wrong bps %d\n", bps);
ok(bps == 2, "wrong bps %d\n", bps); ok(photo == 3, "wrong photometric %d\n", photo); ok(samples == 1, "wrong samples %d\n", samples);
ok(colormap == 6, "wrong colormap %d\n", colormap);
ok(colormap == 12, "wrong colormap %d\n", colormap);If the test succeeds both before and after this change, doesn't it mean that 2bpp is never actually tested? What's the point of this test?
It means that the original test is broken, and I had to fix it after adding actual support for it in Wine.
Well, the broken test is added in your 1/9 patch...
That's correct, and I didn't notice that until I've spotted it.
That also means that since 2bpp is not supported under Windows I missed that the test is broken while running the test under Windows.
Which is my point, if 2bpp is not supported on Windows what's the point of supporting it in Wine? And having a test for it?
It seems to me that the correct test would be to verify that 2bpp doesn't work, and then Wine should conform to that.
The thing is that I first added support for palette formats in Wine, and only after that added the tests and discovered that 2bpp format is not supported under Windows. Then I just decided to leave the implementation and mark the tests as broken. The reasoning that since there is an official GUID for 2bpp format then nothing prevents it to be implemented in the codec: even if Microsoft codecs don't support it other codecs may have this format supported.