Hi,
I couldn’t get a build of ROOT pushed through, though I am quite far. I decided to first try to get cling up. After burning quite a bit of carbon I have cling and the tests built. However I am getting some failures for the check-cling target:
O, and IIRC we had a problem with one of FreeBSD’s system APIs - putenv or similar… Again IIRC it prevented cling from finding its directories back. Now - I won’t get to look at this over the weekend; please ping me next week should you not hear back from me!
O, and IIRC we had a problem with one of FreeBSD’s system APIs -
putenv> or similar… Again IIRC it prevented cling from finding its
directories back. Now - I won’t get to look at this over the weekend;
please ping me next week should you not hear back from me!
So I’ve now got the build installed and cling up and running,
(rpaths work, glad I didn’t have to work through that) and yes, it
is having trouble finding its llvm directories. I at least have
something to look for now.
Thanks for the pointers. I suspect I will need to insert
tools/cling/lib/Interpreter/CIFactory.cpp into my brain in order to
fix this.
However, I was able to make some progress with the ‘ninja cling-test’
(or check-cling, doesn’t seem to matter) failures. I was able to make
a few of tests work by replacing the ‘lit’ generated -I include. For
instance ninja cling-test produces this sort of output:
FAIL: Cling :: Extensions/Lookup/ControlFlow.C (119 of 150)
******************** TEST ‘Cling :: Extensions/Lookup/ControlFlow.C’
FAILED ********************
Script:
Fails with that -I build include path, but it does work if I
substitute the installation include path, in my case,
/usr/local/pkg/cling/include. Reading the various ‘lit’ scripts, I
thought I might be able to fix this by exporting
INCLUDE=/usr/local/pkg/cling/include before running ninja cling-test,
but no luck. If you could point out where the lit test runner sets
it that might help me.