Bug#518295: GL_LINE_LOOP not rendered correctly on Radeon VE 7000 QY

2009-03-09 Thread BALLABIO GERARDO
Brice Goglin wrote:
 Now it would be nice to try with mesa 7.3 (currently in experimental).

I've been able to install mesa from experimental (packages: libdrm-intel1, 
libdrm2, libgl1-mesa-dev, libgl1-mesa-dri, libgl1-mesa-glx, mesa-common-dev) in 
a Lenny live cd environment. The bug still shows on my system. It seems that 
the issue is indeed specific to this graphics card model.

If you want me to test something, please let me know. Otherwise, don't worry 
too much. There aren't probably many such cards still being used. Myself, I can 
work around the bug in my programs by replacing GL_LINE_LOOPs with 
GL_LINE_STRIPs and repeating the start/end point (the extra point shouldn't 
affect performance very much). And I've been thinking of buying a new card 
already for some time, maybe this will make me resolve myself eventually.

Gerardo



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#468359: Can you reproduce #468359?

2008-12-26 Thread BALLABIO GERARDO
I tried with a Lenny livecd. I can confirm that half of the bug is still 
present. That is, the first test case (bug1.c) still gives the same incorrect 
result, while the second (bug2.c) now seems to work correctly.
I don't think I'm able to try with Sid or experimental, sorry.

Gerardo



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#468359: Image dimensions wrong after calling MagickCropImage

2008-02-28 Thread BALLABIO GERARDO
Package: libmagick9-dev
Version: 7:6.2.4.5.dfsg1-0.14

In a C program I'm writing that uses the imagemagick library API, when I
crop an image by calling MagickCropImage and then I manipulate it
further, I obtain incorrect results. As far as I can tell, the image
does get cropped (if I save it it's ok), but the geometry information in
the MagickWand is NOT changed to reflect the cropping -- i.e., Magick*
functions believe that the image still has the original size.

I'm attaching two test cases and an image that should serve as input for
them (but any other image 300x300 or larger should do).

In the first test, I crop the same image twice:

  MagickCropImage(wand, 300, 300, 50, 50);
  MagickCropImage(wand, 250, 250, 0, 0);

The resulting image should be 250x250 (from 50 to 300 of the original
image along both directions). Instead it's 200x200 (from 50 to 250). It
seems that the second crop is still applied to the original image, not
to the cropped one; the resulting image is the _intersection_ of the two
cropped regions.

In the second test, I crop a 50x10 portion of the image and then use a
PixelIterator and a PixelWand to loop over pixels. For each row the
program prints the line width (as set by PixelGetNextIterationRow): it
should be 50. Instead it's 500. Again, it seems that operations
following the crop are applied to the original image.

My system is a 32-bit PC running Etch. The kernel is the stock kernel in
Etch, same for libc and all packages that might be involved. I don't
know whether the problem is also present in Lenny or Sid.

Thanks
 Gerardo Ballabio


bug1.c
Description: bug1.c


bug2.c
Description: bug2.c
attachment: in.png