We are working to change the default version of GCC from 4.7 (which
is end of life) to 4.8.
This port is one of a handfull that blocks this migration since it fails
to build on FreeBSD/amd64 10 as follows:
ada -c -O2 -gnatn -g -gnatT3 -gnatws -gnat12 ada_semantic_tree-dependency_tree.adb
ada: internal compiler error: Killed (program gnat1)
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
gprbuild: *** compilation phase failed
gmake: *** [all] Error 4
has the full logs.
This bug report makes no sense to me. GPS is built with USES=ada, which is, by definition, lang/gcc-aux (based on gcc 4.9)
So trying to build an ada program with an ada free lang/gcc48 makes no sense. I think the report is invalid. I'm not sure what trickery was used to force gcc48 but it's likely invalid.
No trickery involved that I'd know of; this is the result of an -exp
run to update the default version of GCC (USE_GCC=yes, lang/gcc) from
GCC 4.7 to GCC 4.8.
According to https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=192025
this caused a regression for devel/gps. If you believe this is a false
positive, we can ignore/skip this failure report I guess.
with my luck it will be repeatable, but I have no idea how to approach a fix. The Ada stuff should all be built by the Ada compiler (same as now) and the C stuff should be built by base compiler (same as now) so the only change would be some like of dependency library requiring USE_GCC=any/yes and even then I don't see why ada compiler would care.
So I don't know what I can do at this point.
I relaunched the build and this time it succeeded: http://package23.nyi.freebsd.org/data/10amd64-default-PR192025/2014-07-26_21h30m19s/logs/gps-ide-22.214.171.124.log
That means we can close the ticket.
One thing that might be still good to do is rebase this from 4.7 to 4.8,
but, yes, the ticket as such was false alert.
rebase what exactly?
lang/gcc-aux is based on gcc49, not gcc47. If not gcc-aux, I don't know what you mean.