Hi Adam,
This is mainly a brain-dump, there's nothing wrong the patch.
On Thu, Jul 10, 2008 at 12:44:55AM -0400, Adam Petaccia wrote:
dlls/gdiplus/tests/region.c | 38 ++++++++++++++++++++++++++++++++++++++ 1 files changed, 38 insertions(+), 0 deletions(-)
diff --git a/dlls/gdiplus/tests/region.c b/dlls/gdiplus/tests/region.c index 1b1b047..7b1eba5 100644 --- a/dlls/gdiplus/tests/region.c +++ b/dlls/gdiplus/tests/region.c @@ -361,6 +361,44 @@ todo_wine expect(Ok, status); status = GdipDeleteRegion(region); expect(Ok, status);
- status = GdipGetRegionDataSize(region, &needed);
- expect(Ok, status);
- ok(needed == 72, "Expected 72, got %d\n", needed);
- status = GdipGetRegionData(region, (BYTE*)buf, sizeof(buf), &needed);
- expect(Ok, status);
- expect_dword(buf, 64);
- trace("buf[1] = %08x\n", buf[1]);
- expect_dword(buf + 2, RGNDATA_MAGIC);
- expect_dword(buf + 3, 0);
- expect_dword(buf + 4, RGNDATA_PATH);
- expect_dword(buf + 5, 48);
- expect_dword(buf + 6, RGNDATA_MAGIC);
- expect_dword(buf + 7, 4);
- expect_dword(buf + 8, 0);
- expect_float(buf + 9, 5.6);
- expect_float(buf + 10, 6.2);
- expect_float(buf + 11, 7.2);
- expect_float(buf + 12, 8.9);
- expect_float(buf + 13, 8.1);
- expect_float(buf + 14, 1.6);
- expect_float(buf + 15, 5.6);
- expect_float(buf + 16, 6.2);
There's one more DWORD left here (buf + 17).
I realised after I sent in my last lot of tests that this is the BYTE array of PathPointType*s packed into DWORDs (hence the 0x81...... vs 0x01...... that you noted in an earlier patch).
As you've also noticed there are two ways of storing paths, if all the co-ords are short ints then they get stored that way and the 0x00004000 bit of buf + 8 is set. Otherwise the co-ords are stored as floats.
In addition, the 0x00002000 bit of that same DWORD corresponds to the FillMode of the path.
Anyway, good work!
Huw.