It should now prompt for a value if makewidth is undefined w
an open -j argument was never the intended behavior - as it is not
particularly safe
Also the boost recipe has been converted to do $makewidth wide builds
with bjam now
-Tim
On 01/09/2014 11:26 AM, Tom Rondeau wrote:
On Thu, Jan 9, 2014 at 11:15 AM, Robert McGwier <rwmcgw...@gmail.com> wrote:
I did an update to pybombs two days ago, and immediately gnuradio got to
around 57% completion and just hung my Ubuntu 13.10 computer so hard, I
couldn't get it to respond to ctrl-alt-del for over half and hour.
I tried again after doing a force rebuild on gnuradio and it was worse. I
long-hit the power button to do a reset.
I checked the pybombs defaults and it is makewidth = 4, but it certainly
acts like it is ignoring this and doing
make -j
. Further symptoms include a nonstop, hard green LED on disk writes on a 8
GB system. I have never seen it actually go to a major use of virtual
memory (if that is what is happening) until now.
I went into src/gnuradio/build and did a manual make -j4 and it worked
perfectly, thus my diagnosis.
Bob
Yep, making with an unnumerated -j is really dangerous, even on very
high-end systems.
There was another report about issues with makewidth, so that needs
looking into.
Tom
_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
