[cairo] Time for a release ?

[Heiko Lewin]

> Or am I misunderstanding
> you and you do not mean "admin work" as "keep GitLab running", but
> rather "make the test suite pass"?

Even with only testing image surfaces, the test-suite seems to deadlock:

TESTING pthread-show-text
pthread-show-text.image.argb32 [0x1]:    FAIL
pthread-show-text.image.rgb24 [0x1]:    ^C


"Making it pass" I would remove all thread-shared data-objects from
cairo. No "holdovers" or caching of font-faces and stuff which could be
done on an application-layer.
To me, that code always seemed broken beyond repair and violating the
principle of separation of concern, which is why I evade it as best as I
can.
However I won't spend any afford removing that stuff without any backing
of the idea as doing so might affect applications relying on that
functionality performance-wise.

Other than this many of the fails indicate a regression in font-hinting.
I don't know when that stuff did break, but at first glance most FAILs
differ in snapping letters differently.
The font-hinting or fixes for it were a more recent addition if I
remember correctly, however I cannot remember the test-suite passing at
any time,
so I'll not spend any effort on that one either - in the current state I
couldn't even tell if the tests or the code is wrong.


Might seem keen, but I'd wait on concrete reports from a test-release.
If there really is something broken, ppl will come and fix it or they won't.
"Free as in freedom..."

(Not) just to start a controversy
HL

...

TESTING text-antialias-gray
text-antialias-gray.image.argb32 [0x1]:    FAIL
text-antialias-gray.image.rgb24 [0x1]:    FAIL
text-antialias-gray: FAIL (image)

TESTING text-antialias-none
text-antialias-none.image.argb32 [0x1]:    FAIL
text-antialias-none.image.rgb24 [0x1]:    FAIL
text-antialias-none: FAIL (image)

TESTING text-antialias-subpixel
text-antialias-subpixel.image.argb32 [0x1]:    FAIL
text-antialias-subpixel.image.rgb24 [0x1]:    FAIL
text-antialias-subpixel: FAIL (image)

TESTING text-antialias-subpixel-rgb
text-antialias-subpixel-rgb.image.argb32 [0x1]:    FAIL
text-antialias-subpixel-rgb.image.rgb24 [0x1]:    FAIL
text-antialias-subpixel-rgb: FAIL (image)

TESTING text-antialias-subpixel-bgr
text-antialias-subpixel-bgr.image.argb32 [0x1]:    FAIL
text-antialias-subpixel-bgr.image.rgb24 [0x1]:    FAIL
text-antialias-subpixel-bgr: FAIL (image)

TESTING text-antialias-subpixel-vrgb
text-antialias-subpixel-vrgb.image.argb32 [0x1]:    FAIL
text-antialias-subpixel-vrgb.image.rgb24 [0x1]:    FAIL
text-antialias-subpixel-vrgb: FAIL (image)

TESTING text-antialias-subpixel-vbgr
text-antialias-subpixel-vbgr.image.argb32 [0x1]:    FAIL
text-antialias-subpixel-vbgr.image.rgb24 [0x1]:    FAIL
text-antialias-subpixel-vbgr: FAIL (image)

TESTING text-cache-crash
text-cache-crash.image.argb32 [0x1]:    PASS
text-cache-crash.image.rgb24 [0x1]:    PASS
text-cache-crash: PASS

TESTING text-glyph-range
text-glyph-range.image.argb32 [0x1]:    FAIL
text-glyph-range.image.rgb24 [0x1]:    FAIL
text-glyph-range: FAIL (image)

TESTING text-pattern
text-pattern.image.argb32 [0x1]:    FAIL
text-pattern.image.rgb24 [0x1]:    FAIL
text-pattern: FAIL (image)

TESTING text-rotate
text-rotate.image.argb32 [0x1]:    FAIL
text-rotate.image.rgb24 [0x1]:    FAIL
text-rotate: FAIL (image)

TESTING text-subpixel-positioning
text-subpixel-positioning.image.argb32 [0x1]:    XFAIL
text-subpixel-positioning.image.rgb24 [0x1]:    XFAIL
text-subpixel-positioning: XFAIL

TESTING text-transform
text-transform.image.argb32 [0x1]:    FAIL
text-transform.image.rgb24 [0x1]:    FAIL
text-transform: FAIL (image)

TESTING text-unhinted-metrics
text-unhinted-metrics.image.argb32 [0x1]:    FAIL
text-unhinted-metrics.image.rgb24 [0x1]:    FAIL
text-unhinted-metrics: FAIL (image)

...




On 11/16/20 5:21 PM, Uli Schlachter wrote:
> Hi,
>
> Am 16.11.20 um 15:31 schrieb Stuart Axon:
>> Is Meson the preferred way of building Cairo these days ?
> I am not sure. The "meson stuff" was only merged three months ago, so is
> quite new. I guess it works in the common settings, but I'd expect the
> "autotools stuff" to cause less problems.
>
>> After briefly looking at the script it's reassuring it seems to do quite a bit, though I noticed that LATEST still points at 1.16.0 for some reason so maybe something isn't quite right.
> I'm sorry, but I do not understand the question. Which LATEST are you
> talking about?
>
>> Does Cairo have CI anywhere?
> FDO's GitLab runs some builds:
> https://gitlab.freedesktop.org/cairo/cairo/-/pipelines
>
>> I guess the tests not all passing is still an issue there?
> Yup. It "builds stuff" and runs tests, but ignores the result of running
> cairo-test-suite by setting CAIRO_TEST_FORCE_PASS=1. That's better than
> nothing.
>
>> This seems like something people with more sporadic free time might be able to help with enabling.
> Sure, please test the meson setup and report problems on GitLab.
>
>> Is it worth looking into funding options to support admin work like this ?
> [...]
>
> Feel free. The current infrastructure runs on freedesktop.org. I am
> actually not sure how the funding works here. Or am I misunderstanding
> you and you do not mean "admin work" as "keep GitLab running", but
> rather "make the test suite pass"?
>
> Cheers,
> Uli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cairographics.org/archives/cairo/attachments/20201116/ffe9e9b9/attachment-0001.htm>