I have been running cvsup nightly to grab -current and -ports,
and noticed some strangeness with awk that seemed to start last
week sometime.
When building /usr/ports/lang/guile, the build exited with an
awk 'internal error' and a log on the console that awk had
exited on signal 6. To test my the
Tony,
On Mon, Aug 21, 2000 at 07:17:31PM -0700, Tony Fleisher wrote:
> I have been running cvsup nightly to grab -current and -ports,
> and noticed some strangeness with awk that seemed to start last
> week sometime.
>
> When building /usr/ports/lang/guile, the build exited with an
> awk 'intern
On Tue, 22 Aug 2000 12:27:01 +0200, Sheldon Hearn wrote:
> > awk: ./guile-snarf.awk:17: (FILENAME=- FNR=9680) fatal error: internal
> > error
> > Abort trap - core dumped
> > *** Error code 1
>
> I get the same thing here. Inspecting the core dump, one finds that the
> abort() happens in eval
I also was able to get it to run properly when not redirecting to a file.
I wonder if there might be some dependancy in the system that causes
different behavior with this when output is to a tty vs. a
filehandle.
I tried this with a copy of /bin/sh from 4.0-RELEASE, and had the same
result, so
On Thu, Aug 24, 2000 at 05:19:34PM +0200, Sheldon Hearn wrote:
> So I'm a bit stumped as far as formulating an easy How-To-Repeat is
> concerned. :-(
How about wedging a printenv into the makefile, before the call
to awk, so that you can re-create the environment when testing
it?
--
Andrew
T
On Mon, 21 Aug 2000 19:17:31 MST, Tony Fleisher wrote:
> awk: ./guile-snarf.awk:17: (FILENAME=- FNR=9680) fatal error: internal
> error
> Abort trap - core dumped
> *** Error code 1
Hi Tony,
Aharon Robbins has come back with an "official patch" that fixes the
problem. It has been committed i