[MacPorts] #59785: llvm ports opportunistically build ocaml binding even when +ocaml variant is not selected; leads to conflicts between them
MacPorts
noreply at macports.org
Thu Dec 5 05:52:10 UTC 2019
#59785: llvm ports opportunistically build ocaml binding even when +ocaml variant
is not selected; leads to conflicts between them
--------------------------------+----------------------
Reporter: cooljeanius | Owner: jeremyhu
Type: defect | Status: assigned
Priority: Normal | Milestone:
Component: ports | Version: 2.6.2
Resolution: | Keywords:
Port: llvm-7.0 llvm-9.0 |
--------------------------------+----------------------
Comment (by cooljeanius):
Replying to [comment:2 kencu]:
> the various bindings (ocaml, go, python) that come with llvm* are
detected opportunistically during a source installation, and are
unfortunately not as simple to turn off as you might expect them to be.
There should just be some simple cmake switches for them, and there are
switches that seem related, but they don't really effectively turn the
bindings off.
>
> And things, as I recall, changed in this area at certain points in the
llvm life cycle.
>
> Last time I looked into this about 6 months ago, it was looking to take
some cmake surgery to get this done. As these issues never seem to affect
the buildbot versions of clang/llvm, which are built in clean environments
where these go/ocaml etc bindings are never opportunistically detected, it
doesn't come up for most people.
>
> I see you're building +universal so that must be why you are seeing
this, and most don't.
>
> It is a bit of a pain, it should be fixed, upstream has declined to fix
it and says things are fine as they are (in the email exchange I read
about this last go round).
The llvm mailing lists have public archives, right? Do you have a link to
where in the archives this exchange occurred that I could read?
--
Ticket URL: <https://trac.macports.org/ticket/59785#comment:3>
MacPorts <https://www.macports.org/>
Ports system for macOS
More information about the macports-tickets
mailing list