Date: Wed, 23 Feb 2022 22:15:01 +0000 From: bugzilla-noreply@freebsd.org To: ports-bugs@FreeBSD.org Subject: [Bug 262109] Mk/Uses/python.mk: Improve CMake/Python integration Message-ID: <bug-262109-7788-meIrY91yFy@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-262109-7788@https.bugs.freebsd.org/bugzilla/> References: <bug-262109-7788@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D262109 --- Comment #4 from John Hein <jcfyecrayz@liamekaens.com> --- (In reply to Kubilay Kocak from comment #3) No, I am not aware of anything. The FindPython{2,3,}.cmake modules use different hint variables. And the CMakeLists.txt file in a particular soft= ware package could use any of them. In the case of graphics/libjxl, for instance, it conditionally uses any of = them depending on what it dynamically sees in the shebang header for a2x. In the case of multimedia/onevpl, it uses FindPython{Interp,Libs}.cmake (wh= ich wants the Python_ADDITIONAL_VERSIONS knob or FindPython.cmake (wants Python_EXECUTABLE) conditionally depending on detected cmake version. And = it expresses a preference for python3 by using FindPython3.cmake as a hint bef= ore trying the others. The modules that cmake bundles do a poor job of enforcing any sort of consistency. It's just a fact of life for that tool. And it _can_ change = in a later version of cmake without careful consideration for backward compatibility. Some of the main tool core features have a somewhat rigorous feature compatibility infrastructure, but many of the bundled modules are n= ot nearly as well... considered. --=20 You are receiving this mail because: You are on the CC list for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-262109-7788-meIrY91yFy>