Also a lazy_index bug (Re: Assignment fails, crash)

[Jarno Rajahalme]

I looked at the src/ov-lazy-idx.h|cc and that being new code (2010), therein 
lies a bug somewhere. To verify I tested this with 3.2.3, and there the results 
are identical.

Also, looking at octave_lazy_index::save_ascii for example, I have no idea why 
the type "lazy_index" should appear in the saved file, as the saved object is a 
octave_value created from idx_vector?

However, it might be mostly due to lazy_index that my code runs so much faster, 
so I would very much appreciate if someone would take a look at this..

  Jarno

On May 10, 2010, at 8:37 PM, Jarno Rajahalme wrote:

> Here is another bug, likely related to the other one. (both are now in 
> Savannah). Both should produce the same output.
> 
> $ octave
> GNU Octave, version 3.3.51+
> ...
> Octave was configured for "x86_64-apple-darwin10.3.0".
> ...
> octave:1> testdata(find([1:2] != 0)) = [1:2]
> testdata =  1
> octave:2> testdata(find([1:2] != 0)+0) = [1:2]
> testdata =
> 
>   1   2
> 
> octave:3>
> 
> 
> I would really appreciate if someone knowledgeable would comment on this. I 
> would hate to go back to 3.2.3 because of this, the dev branch seems to be 
> about 4x faster on my code...
> 
> I also got this, while reducing the bug from my code, but do not know how to 
> reproduce at this time:
> 
> panic: Segmentation fault -- stopping myself...
> attempting to save variables to `octave-core'...
> terminate called after throwing an instance of 'std::logic_error'
>  what():  basic_string::_S_construct NULL not valid
> panic: attempted clean up apparently failed -- aborting...
> Abort trap
> 
> 
> On May 10, 2010, at 7:58 PM, ext Jarno Rajahalme wrote:
> 
>> $ octave
>> GNU Octave, version 3.3.51+
>> ...
>> octave:1> 1(find([1 1] != 0))
>> ans =
>> 
>>   1.0000e+00   2.1617e-314
>> 
>> octave:2> 1(find([1 1] != 0)+0)
>> error: A(I): index out of bounds; value 2 out of bound 1
>> octave:2> 
>> 
>> The result on the prompt 1 above is clearly incorrect. Saving a result of 
>> the find above, and looking at the file tells that the result is a 
>> "lazy_index".
>> 
>> Any ideas?
>> 
>> Jarno
>> 
>> 
>