Open runlevel5 opened 7 months ago
Might be an upstream issue, cf. https://github.com/microsoft/vcpkg/pull/38250#discussion_r1570513595
@dg0yt it is plausible though I do not think it is an issue upstream. For example in Fedora Linux, the RPM spec file does not do anything special with configure
and make
part (ref https://src.fedoraproject.org/rpms/libpng/blob/rawhide/f/libpng.spec#_71) so the logic of upstream CMakeList must be correct. One way to verify that is that I use the /usr/lib64/libpng16.a
(belongs to libpng-static
package on Fedora 40 ppc64le) and the app does not complain anymore.
For example in Fedora Linux, the RPM spec file does not do anything special with
configure
andmake
part (ref https://src.fedoraproject.org/rpms/libpng/blob/rawhide/f/libpng.spec#_71) so the logic of upstream CMakeList must be correct.
No. Fedora (and the "additional context" in your initial post) use autotools/configure/make. This is independent of CMake. And that might explain why nobody on linux noticed the problem.
This is an automated message. Per our repo policy, stale issues get closed if there has been no activity in the past 180 days. The issue will be automatically closed in 14 days. If you wish to keep this issue open, please add a new comment.
Don't close this @github-actions
Describe the bug I ran into errors when compiling my Raytracing demo app. Please refer to Failure logs section for more details
Environment
To Reproduce Steps to reproduce the behavior:
Expected behavior No error
Failure logs
Additional context
That is likely caused by the lacking of VSX support. From what I could tell is that
libpng
does opt-in the VSX optimisation option, please refer to codesHowever it's seems to me it is not the case.
I guess the vcpkg port might have to handle that logic instead... (though I am not sold of that idea)
My ugly pseudocodes are kinda like this: