Swift now functional (on FreeBSD)


(Davide C. C. Italiano) #1

Hi,
this is a quick heads-up for all the people interested in running
swift on FreeBSD.
I committed several fixes in the swift repo, but there was a bug in
FreeBSD runtime linker (rtld) which prevented swift (and the repl) to
work correctly in all but very basic scenarios.
After some discussion, and review, the rtld patch finally was
committed to FreeBSD-CURRENT.
https://svnweb.freebsd.org/base?view=revision&revision=294373

The only version supported for now is FreeBSD 11-CURRENT. The next
FreeBSD snapshot (which will come out next week) will contain the fix.
There are plans to merge back this commit to the stable branch at some
point, but that won't happen for FreeBSD 10.3, sorry.

This means that swift is somewhat usable for FreeBSD. There are a
*lot* of failures in the test suite but I guess it's about time to set
up a buildbot to catch at least regression failures.

Any feedback is greatly appreciated.

Thanks,

···

--
Davide


(Chris Lattner) #2

Great work, congrats! It is fantastic to see this progress,

-Chris

···

On Jan 20, 2016, at 12:06 PM, Davide Italiano via swift-dev <swift-dev@swift.org> wrote:

Hi,
this is a quick heads-up for all the people interested in running
swift on FreeBSD.
I committed several fixes in the swift repo, but there was a bug in
FreeBSD runtime linker (rtld) which prevented swift (and the repl) to
work correctly in all but very basic scenarios.
After some discussion, and review, the rtld patch finally was
committed to FreeBSD-CURRENT.
https://svnweb.freebsd.org/base?view=revision&revision=294373

The only version supported for now is FreeBSD 11-CURRENT. The next
FreeBSD snapshot (which will come out next week) will contain the fix.
There are plans to merge back this commit to the stable branch at some
point, but that won't happen for FreeBSD 10.3, sorry.

This means that swift is somewhat usable for FreeBSD. There are a
*lot* of failures in the test suite but I guess it's about time to set
up a buildbot to catch at least regression failures.

Any feedback is greatly appreciated.