After applying patch <a class="issue-link js-issue-link" data-error-text="Failed to lo

I believe the issue is that at <a href="https://github.com/blend2d/blend2d/blob/master

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

Should be fixed by <a class="commit-link" data-hovercard-type="commit" data-hovercard-

Should be fixed by <a href="https://github.com/blend2d/blend2d/commit/e46

ctest of bl_test_unit failed on fedora s390x about blend2d HOT 5 CLOSED

topazus commented on July 22, 2024

ctest of bl_test_unit failed on fedora s390x

from blend2d.

Comments (5)

QuLogic commented on July 22, 2024

I believe the issue is that at https://github.com/blend2d/blend2d/blob/master/src/blend2d/pixelconverter.cpp#L545 the converter takes the last byte of the [PX]RGB32 value (I printed out srcAI which was 3), whereas the test fills the expected value to the MSB of a 32-bit value: https://github.com/blend2d/blend2d/blob/master/src/blend2d/image_test.cpp#L86 On a big-endian system, that would be the first byte (I'd expect srcAI to be 0). But I'm not sure where alphaShift gets set.

from blend2d.

kobalicek commented on July 22, 2024

Thanks a lot for the report and for the PR - I will definitely check this code out.

Unfortunately I have no big endian machine to test this on so I'm not that surprised that something is broken even when I really tried to write code that should be aware of the differences.

from blend2d.

kobalicek commented on July 22, 2024

@QuLogic You are correct here. I just looked into this and yeah the function bl_convert_a8_from_8888 is wrong on big endian machines.

from blend2d.

kobalicek commented on July 22, 2024

Should be fixed by e4656f4

from blend2d.

topazus commented on July 22, 2024