On Sun, Dec 31, 2017 at 04:06:39PM +0100, Björn Ketelaars wrote:
> On Sun 31/12/2017 15:38, Klemens Nanni wrote:
> > `make build' starts with bad make usage, builds almost everything
> > and eventually exits with one. Rerunning it again build some other stuff
> > and exits zero; this doesn't seem right.
> >
> > `make test' fails as well.
> >
> > See full logs attached for both targets.
>
> Klemens, I'm unable to reproduce the errors from your build. Could you check
> that your ports tree is up to date, and that the diff has been applied
> correctly? What is the result of 'make clean && make && make test'?
Latest snapshot and ports tree here on my amd64 X230.
The build broke due to MAKE_JOBS=4 on my machine, =1 fixes it; sorry for
that.
> My build log: http://sprunge.us/LZSY
> My test log: http://sprunge.us/JRFC
Both links yield 500 errors.
The test suite however still fails for both 1.3.2 and 1.3.3, see each
${WRKDIR}/test.log attached.
1.3.2 tests fail to allocate enough memory and I'm not sure about how
much it wants. This is what I ran them on:
$ sysctl hw.{phys,user}mem
hw.physmem=16844521472
hw.usermem=17000685568
$ df -h -t mfs
Filesystem Size Used Avail Capacity Mounted on
mfs:67852 5.8G 537M 5.0G 10% /tmp
$ make show=WRKOBJDIR
/tmp/pobj
Those 537M are before running 1.3.2, during tests it went up to about
4.6G FWIW.
It's not much but I didn't look further into these issues so far. Can
anyone else successfully run the test suits?
No comments:
Post a Comment