Tue 10 Apr 2012 01:00:49 PM UTC, comment #1:
> I have to make some comments on bug #35933
Then please make them there, not here.
I'm closing this bug report.
> However, magically, after that, the plot works. It works,
> even if I do it as a fist instruction entering in Octave.
That is the nature of nasal demons. It is folly to attempt a careful description of nasal demons. Sometimes it segfaults, sometimes it runs out of memory, somtimes it wipes your hard drive, and sometimes it makes demons fly out of your nose.
|
Tue 10 Apr 2012 12:26:18 PM UTC, original submission:
Dear everybody,
I have to make some comments on bug #35933
I installed today Octave 3.6.1 from the prebulid zip.
As first instruction to test the program I tried the plot instruction, with a very simple plot([1:10]).
I got the message
panic: Segmentation violation -- stopping myself... etc.
I look in the bug tracker, finding bug #35933.
I tried all the libblas dll, both with and witout the instruction
graphics_toolkit gnuplot
with no result (a pain...).
I tried to test gnuplot alone: it works, but Octave still don't.
Before to give up, I decided to test some matrix instruction, expecting to confirm the problems with the libblas.dll. However with the commands
a = rand(30,30); b = inv(a);
I got no errors.
However, magically, after that, the plot works. It works, even if I do it as a fist instruction entering in Octave.
If it is important, I'm running win XP home edition, Vers. 2002, sp.3 processor athlon 64 3400+. At the time I decided to test the matrix instructions, I was probably running the pentium libblas.dll, but now I switched back on the default one, that works anyway.
|