On 2012-11-02 12:18- Andrew Ross wrote:
>
> Alan,
>
> I've now fixed you warnings. Turns out there was an API change to
> remove a redundant argument in 3.4.0.
That API change does sound like it could be the source of the
intermittent invalid pointer and backtrace result I got (just once) at
Alan,
I've now fixed you warnings. Turns out there was an API change to
remove a redundant argument in 3.4.0. The guilty code was in our
plplot_octave.i file so it is relatively easy to fix. I've tested
the changes with 3.2.4 and 3.6.2 and all seems fine. It does involve
an #if #else #endif sta
Alan,
I can confirm the warnings you see. This is is swig generated code so
I'll need to dig a bit deeper to see what the problem is.
I cannot reproduce you crash using make test_octave_psc on Debian
testing. This should be the same(?) version as you are using Alan.
I also found a number of o
Alan,
Thanks for the testing. I'll take a look. The bindings are now all swig
generated so it may be a swig issue as well. I've not had any issues up
to this stage building on Debian testing.
Andrew
On Thu, Nov 01, 2012 at 06:41:47PM -0700, Alan Irwin wrote:
> Hi Andrew:
>
> I just got this:
Hi Andrew:
I just got this:
time (nice -19 make -j8 install >& install.out)
*** glibc detected *** /usr/bin/octave: free(): invalid pointer:
0x02b0fa80 ***
=== Backtrace: =
/lib/x86_64-linux-gnu/libc.so.6(+0x75b76)[0x2acd45a95b76]
/lib/x86_64-linux-gnu/libc.so.6(cfree+0x6c)[0x
