[jasper] freeglut is not a dependency in macOS

[david-antiteum]

Also remove freeimage:x64-osx from baseline as it now compiles

• What does your PR fix? Fixes issue #

Broken Qt 5 in macOS after commit f8165f7#diff-5df6afe35378f98a8c68a59e1faed145

[david-antiteum]

@JackBoosY Any idea about the issue on vcpkg-osx-PR-test? REGRESSION: ogre:x64-osx ?

[JackBoosY]

@david-antiteum We are fixing the problem that the error log cannot be extracted from the osx pipeline. After the repair, I will restart the osx test of this PR and provide the error log to you.

[JackBoosY]

Seems a header file could not be found:

/Users/vagrant/azure-agent/_work/2/s/buildtrees/ogre/src/c710e36bc0-fd78f2aff5/OgreMain/include/OgrePlatform.h:150:13: fatal error: 'Availability.h' file not found
#   include "Availability.h"
            ^~~~~~~~~~~~~~~~
1 error generated.

[david-antiteum]

My change is not related to Ogre and the missing header is a system header. No idea of what is going on.

[Neumann-A]

@JackBoosY ogre should probably use include <Availability.h> instead of include "Availability.h" since Availability.h seems to be a system header

[david-antiteum]

Ogre has not changed in 5 months and the related PR #9686 passes the vcpkg-osx-PR-test.

[david-antiteum]

Ah! I'm seems that Qt5 for macOS is broken in other parts:
Building package qt5-tools[core]:x64-osx...
...
Undefined symbols for architecture x86_64:
"_iconv_open", referenced from:
_FcSfntNameTranscode in libfontconfig.a(fcfreetype.c.o)
"_iconv", referenced from:
_FcSfntNameTranscode in libfontconfig.a(fcfreetype.c.o)
"_iconv_close", referenced from:
_FcSfntNameTranscode in libfontconfig.a(fcfreetype.c.o)

[Neumann-A]

That just means FONTCONFIG_LIBS in qt5-base must contain the libiconv libraries.

[Neumann-A]

Just a Note here: Qt was not build on osx due to regressions in a feature of a dependency of Qt which Qt does not even use. #9705 is running a working Qt CI build on OSX and should discover these kind of errors.

[vicroms]

/azp run

[Neumann-A]

Ogre has not changed in 5 months and the related PR #9686 passes the vcpkg-osx-PR-test.

Yeah 9686 passes the test but it does not build ogre. Your change unblocks the ogre build in CI and thus the regression happens.
From the error I would simply try to patch ogre to use include <Availability.h> and see what happens. GCC/Clang are a bit pedantic about the correct usage of "" vs <>

[JackBoosY]

@david-antiteum Can you add ogre:x64-osx=fail to VCPKG_PATH/scripts/ci.baseline.txt ?
Except this, LGTM.

Thanks.

[david-antiteum]

Done, ci.baseline.txt updated.

[JackBoosY]

LGTM, thanks for this PR!