[cairo] 1.6: Current cairo-pdf test failures

[Carl Worth]

On Thu, 24 Jan 2008 12:43:35 -0800, Carl Worth wrote:
> Meanwhile, there are a *bunch* of PDF, PS, and SVG failures in the
> test suite right now that must be sorted out before release, (and I'll
> insert more fallbacks if necessary to make the test suite pass).

I started looking at "CAIRO_TEST_TARGET=pdf make test" on
master. There are 10 failures. One of them, (meta-surface-pattern),
just has a stale reference image, (using meta-surfaces for
_cairo_pdf_surface_create_similar makes its result even *better* than
the image-backend-generated reference image).

So, for the remaining 9 failures I started some git-bisect runs and
this is what I came up with:

First, this one

    bitmap-font
        12cc32b674d3f1b705ef1a66bbce1ef8c4488daf is first bad commit
        commit 12cc32b674d3f1b705ef1a66bbce1ef8c4488daf
        Author: Søren Sandmann <sandmann at redhat.com>
        Date:   Wed Jun 13 00:15:34 2007 -0400

            [pixman] Initial port to standalone pixman library.

Søren, would you be willing to take a look into this test case?

Then the failures for these 8 all git-bisect consistently:

    linear-gradient-reflect
    mask
    paint-source-alpha
    paint-with-alpha
    push-group
    rotate-image-surface-paint
    scale-source-surface-paint
    source-clip-scale
        1cdd11873b18ec55521ffd744ad12c619cf9b227 is first bad commit
        commit 1cdd11873b18ec55521ffd744ad12c619cf9b227
        Author: Adrian Johnson <ajohnson at redneon.com>
        Date:   Sat Aug 25 06:44:08 2007 +0930

            Add finer-grained fallback support to the PDF surface

            The stream handling has been changed to support writing the content to
            one or more group objects. Each page has a top level knockout
            group. The first operation in the knockout group paints another group
            containing the content. Fallback images are painted from the knockout
            group. This ensures that fallback images do not composite with any
            content under the image.

And that's another not-too-surprising place for git-bisect to land.

Now, this git-bisect run isn't actually the complete answer we
want. Instead, it points to us having missed some things. For example,
I looked closer at push-group and on master
push-group-pdf-argb32-out.png is missing some pieces of the
diamond. But on 1cdd1187 that part of the test passes perfectly.

Meanwhile, back on 1cdd1187, push-group-pdf-rgb24-out.png is failing,
(image coming out solid white). This one is tricky because it looks
like the PDF output might be just fine and that something is just
going wrong in the pdf2png stage.

So one approach to tracking this down better might be to identify a
good pdf2png, (against some fixed cairo version for example), and
bisect with that consistently, (rather than using the bisected cairo
version both to generate the PDF output and to convert it to a PNG).

If anyone else wanted to chase some of these down further, that would
be great!

-Carl
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20080127/68110800/attachment-0001.pgp