tech-pkg archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: lang/ocaml generated PLIST

Seems like a sensible solution - the conditionals are a relic of the past that I've never really thought about removing; but now that you mention it, there certainly is no reason to keep them that I know of.

I was planning to migrate 4.12 from wip to main pkgsrc soonish, so I might make that change as well then, but if someone wants/needs to do it earlier, they're more than welcome to.



On 13 June 2021 09:09:54 CEST, nia <> wrote:
It seems to me that the OCaml PLIST contains too many target-specific
entries, and complex conditionals such as these:

# Optional components built only on certain platforms.
.if (${MACHINE_ARCH} == "i386") || (${MACHINE_ARCH} == "powerpc") || \
!empty(MACHINE_ARCH:M*arm*) || (${MACHINE_ARCH} == "aarch64") || \
(${MACHINE_ARCH} == "x86_64")
BUILD_TARGET+= opt opt.opt
TEST_PROGRAM+= ocamltest.opt
PLIST.ocaml-opt= yes
. if empty(MACHINE_PLATFORM:MDarwin-*-powerpc) && \
empty(MACHINE_PLATFORM:MSunOS-*-i386) && \
(${MACHINE_ARCH} != "aarch64")
PLIST.ocaml-prof= yes
. endif

# XXX This conditional is not complete. It should match the
# conditional in ocaml-x.y.z/configure.
.if (${MACHINE_ARCH} == "i386") || (${MACHINE_ARCH} == "x86_64") || \
(${MACHINE_ARCH} == "aarch64") || (${MACHINE_ARCH} == "powerpc") || \
. if !empty(MACHINE_PLATFORM:MLinux-*-*) || \
!empty(MACHINE_PLATFORM:MFreeBSD-*-*) || \
!empty(MACHINE_PLATFORM:MDragonFly-*-*) || \
!empty(MACHINE_PLATFORM:MNetBSD-*-*) || \
!empty(MACHINE_PLATFORM:MDarwin-*-*) || \
PLIST.ocaml-nat= yes
. endif

which are often incomplete and incorrect, leading to errors like this:

Should lang/ocaml be switched to a generated PLIST?

Sent from my Android device with K-9 Mail. Please excuse my brevity.

Home | Main Index | Thread Index | Old Index