Re: AspectRatio Patches [changeset]

[Ben Abbott]

On Oct 31, 2010, at 4:04 PM, logari81 wrote:

> On Sun, 2010-10-31 at 10:01 -0400, Ben Abbott wrote:
>> On Oct 31, 2010, at 7:42 AM, logari81 wrote:
>> 
>>> On Fri, 2010-10-29 at 14:58 +0800, Ben Abbott wrote:
>>>> On Oct 29, 2010, at 2:30 PM, John W. Eaton wrote:
>>>> 
>>>>> On 29-Oct-2010, Ben Abbott wrote:
>>>>> 
>>>>> | John, is it ok to push?
>>>>> 
>>>>> Yes, it seems OK to me.
>>>>> 
>>>>> Thanks,
>>>>> 
>>>>> jwe
>>>> 
>>>> Done.
>>>> 
>>>>    http://hg.savannah.gnu.org/hgweb/octave/rev/51ac3a08e53c
>>>> 
>>>> Ben
>>> 
>>> thank you for applying the changes.
>>> 
>>> I have investigated the problem with the wrong axes limits on page 5 of
>>> my test case pdf and found out that the case with both
>>> plotboxaspectratiomode and dataaspectratiomode set to manual was
>>> explicitly handled (probably incorrectly) for the gnuplot backend in
>>> plot/__go_draw_axes__.m
>>> This code should be removed since the same calculation now is already
>>> done (correctly) in src/graphics.cc .
>>> 
>>> So please consider the attached changeset.
>>> 
>>> As a side effect of this change, there is an important improvement of
>>> the wrong plotbox size with the gnuplot backend in my test cases. See
>>> the new comparison results of my tests here:
>>> 
>>> http://files.ubuntu-gr.org/forum/logari81/Bugs/comparison_aspectratios_update1.pdf
>>> 
>>> Best regards,
>>> 
>>> Kostas
>>> 
>>> <aspectratios1.changeset>
>> 
>> I've pushed your changeset.
>> 
>>      http://hg.savannah.gnu.org/hgweb/octave/summary
>> 
>> Thanks
>> Ben
>> 
> 
> Hi,
> 
> I have attached a last changeset related to the aspectratio
> calculations. A test case demonstrating the fixed behavior of "axis
> equal" can be found on page 3 of the following pdf:
> http://files.ubuntu-gr.org/forum/logari81/Bugs/comparison_aspectratios_update2.pdf
> 
> With this patch applied I am not aware of any further cases of
> incompatibility between matlab and the fltk backend of octave with
> respect to aspectratio settings.
> 
> Unfortunately the gnuplot backend is a bit more problematic. The best
> way of fixing its limitations is to implement real 3D dataaspect support
> upstream. Otherwise consider this
> http://gnuplot-tricks.blogspot.com/2009/08/changing-aspect-ratio-of-axes.html
> 
> I will try to stay informed about any bug or regression reports
> concerning aspect ratios and provide fixes if necessary.
> 
> Best regards,
> 
> Kostas
> 
> <octave_aspectratios_2.patch>

I was unable to apply your patch.

$ patch -p1 < /Users/bpabbott/Desktop/octave_aspectratios_2.patch 
patching file src/graphics.cc
Hunk #2 FAILED at 3911.
Hunk #3 FAILED at 3927.
Hunk #4 FAILED at 4575.
3 out of 4 hunks FAILED -- saving rejects to file src/graphics.cc.rej

Can you generate a mercurial changeset against the current sources?

Ben