Re: incorrect dimension in axes::properties::calc_ticks_and_lims?

[Ben Abbott]

On Jan 25, 2011, at 3:41 AM, logari81 wrote:

> On Mon, 2011-01-24 at 20:06 -0500, Ben Abbott wrote:
>> On Jan 24, 2011, at 6:51 PM, logari81 wrote:
>> 
>>> On Mon, 2011-01-24 at 01:19 -0500, John W. Eaton wrote:
>>>> I noticed that the Matrix tmp_mticks might not be dimensioned
>>>> correctly in the function axes::properties::calc_ticks_and_lims.  The
>>>> code at the end of that function is:
>>>> 
>>>>   int n = is_logscale ? 9 : 4;
>>>>   Matrix tmp_mticks (1, n * tmp_ticks.numel ());
>>>> 
>>>>   for (int i = 0; i < tmp_ticks.numel ()-1; i++)
>>>>     {
>>>>       double d = (tmp_ticks (i+1) - tmp_ticks (i)) / (n+1);
>>>>       for (int j = 0; j < n; j++)
>>>>         {
>>>>           tmp_mticks (n*i+j) = tmp_ticks (i) + d * (j+1);
>>>>         }
>>>>     }
>>>>   mticks = tmp_mticks;
>>>> }
>>>> 
>>>> I see that the loop over I must go from 0 to tmp_ticks.numel()-2 since
>>>> we access tmp_ticks(i+1) in the loop, but this leaves the last N
>>>> elements of tmp_mticks uninitialized.  What is the intent here?
>>>> Should tmp_mticks be declared with
>>>> 
>>>>   Matrix tmp_mticks (1, n * (tmp_ticks.numel () - 1));
>>>> 
>>>> instead?
>>>> 
>>>> jwe
>>> 
>>> that was my mistake, I haven't tried your suggestion yet but I think you
>>> are right, it should be:
>>> 
>>> Matrix tmp_mticks (1, n * (tmp_ticks.numel () - 1));
>>> 
>>> BR
>>> 
>>> Kostas
>> 
>> Kostas,
>> 
>> Will you be submitting a changeset for this?
>> 
>> Ben
> 
> I have just tried this fix and of course works correctly as supposed.
> So, I could push a changeset for this, I am just a bit confused with
> mercurial, when working with different branches. Is there something I
> would take care of with respect to 3.4 branch?
> 
> Kostas

You can use "hg view" to make sure everything looks ok. If it does, please push.

Ben