It looks like it is related somehow to having PartialTypeSignatures
enabled, having no other warnings/errors, and not going beyond parseSource.
Then it sometimes outputs the typed hole warnings, and sometimes not.
Alan
On Fri, May 15, 2015 at 10:37 AM, Alan & Kim Zimmerman
Interesting. The t10358.hs test is a direct clone of a number of others with just the filename to be processed changed.
I have a queued TODO to harmonise this into a single module.
In other words the build process for this test is identical to say t10357.hs, which runs just before t10358.
Alan
On Fri, May 15, 2015 at 10:13 AM, Erik de Castro Lopo < mle+hs@mega-nerd.com> wrote:
Joachim Breitner wrote:
Hi,
I observe this on travis:
Wrong exit code (expected 0 , actual 2 ) Stdout:
Stderr: collect2: ld terminated with signal 9 [Killed] make[3]: *** [t10358] Error 1
I've seen this just recently on these tests. Looking at /var/log/messages showed the `ld` process had been killed by the OOM killer.
My instance of this was on armhf/linux with only 2Gig of RAM and no swap. I have to recompile the kernel to get swap :-(.
Cheers, Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/