axes limit calculations

[Ben Abbott]

On Aug 28, 2008, at 8:47 AM, Ben Abbott wrote:

>
> On Thursday, August 28, 2008, at 08:42AM, "Shai Ayal" <shaiay at gmail.com 
> > wrote:
>> On Thu, Aug 28, 2008 at 3:33 PM, Ben Abbott <bpabbott at mac.com> wrote:
>>> On Thursday, August 28, 2008, at 07:54AM, "Shai Ayal" <shaiay at gmail.com 
>>> > wrote:
>>>> On Thu, Aug 28, 2008 at 2:30 PM, Ben Abbott <bpabbott at mac.com>  
>>>> wrote:
>>>>
>>>>>
>>>>> On Aug 28, 2008, at 7:22 AM, Shai Ayal wrote:
>>>>>
>>>>>
>>>>>> On Thu, Aug 28, 2008 at 1:47 PM, Ben Abbott <bpabbott at mac.com>  
>>>>>> wrote:
>>>>>>
>>>>>> However, switching to fltk appears to work ... well  almost.  
>>>>>> The plot
>>>>>> doesn't render properly.
>>>>>>
>>>>>> Can you elaborate on how it renders?
>>>>>>
>>>>>
>>>>> Perhaps a picture says it best?
>>>>>
>>>>> octave:30> close all
>>>>> octave:31> backend('fltk')
>>>>> octave:32> x = 0:10;
>>>>> octave:33> plot(x)
>>>>>
>>>>> To underlying windows each contribute to the result in a rather  
>>>>> strange
>>>>> way. The resulting figure includes various sections of my email  
>>>>> client, and
>>>>> some things I don't recognize. In addition the following line is  
>>>>> displayed
>>>>> in the terminal.
>>>>>
>>>>> octave:34> ca=nan
>>>>>
>>>>> A picture is attached.
>>>>
>>>>
>>>> This looks like an OpenGL and/or fltk problem. The best way to  
>>>> debug this
>>>> would be to run the ftlk-opengl test programs gl_overlay and  
>>>> glpuzzle and
>>>> see if they work. You should have them in the test subdir of the  
>>>> fltk
>>>> sources. If you did not install fltk from source it's a problem :(
>>>>
>>>> Another possible solution: I know than on OSX, after compiling a  
>>>> fltk app,
>>>> you should run
>>>>
>>>> ftlk-config --post application_exe_file
>>>>
>>>> on your application_exe_file. I am not sure what is the exe file  
>>>> here --
>>>> octave or the fltk_backend.oct file. Maybe you should try both.  
>>>> Also, you
>>>> can just look at fltk-config (it's a bash script)  and see what -- 
>>>> post does
>>>> (in my linux PC it does nothing)
>>>
>>> I don't notice a change on my system either. As I'm using the aqua  
>>> variant, I issued the following commands ...
>>>
>>> $ fltk-config.aqua --post octave
>>> $ fltk-config.aqua --post fltk_backend.oct
>>>
>>> Trying the simple plot again produced a result similar to before  
>>> (different garbage but otherwise the same).
>>>
>>> I've got a full day and don't have time to dig deeper right not. I  
>>> hope to have some time this evening or over the weekend. When I  
>>> have a moment, I'll try the test programs again and report back.
>>>
>>> Ben
>>>
>>> octave:1> backend('fltk')
>>> octave:2> x = 0:10;
>>> octave:3> plot(x)
>>> octave:4> close all
>>> octave:5> backend('gnuplot')
>>> octave:6> quit
>>> panic: Segmentation fault -- stopping myself...
>>> attempting to save variables to `octave-core'...
>>> terminate called after throwing an instance of 'std::length_error'
>>> what():  basic_string::_S_create
>>> panic: attempted clean up apparently failed -- aborting...
>>>
>>
>> Ben -- I somehow have a growing suspicion that you are using an old
>> version of octave. If you have a full day ahead of you, it is a
>> wonderful opportunity to get the current source from hg and rebuild
>> (nice -19, assuming you have enough memory shouldn't bother you at
>> all)
>>
>> Shai
>
> Good point. My build was completed the evening of the 26th. As I  
> have fresh sources with me (about 12 hrs old) I can build those  
> while I do what pays my bills ;-)
>
> Ben

Ok I have a fresh build, but get the same result ... but with some  
warnings this time

octave:1> x = 0:10;
octave:2> backend('fltk')
octave:3> plot(x)
error: hggroup::defaults: not implemented
error: called from:
error:   /Users/bpabbott/Development/mercurial/octave-3-0-0/scripts/ 
plot/__plt2vv__.m at line 82, column 5
error:   /Users/bpabbott/Development/mercurial/octave-3-0-0/scripts/ 
plot/__plt2__.m at line 68, column 14
error:   /Users/bpabbott/Development/mercurial/octave-3-0-0/scripts/ 
plot/__plt1__.m at line 61, column 10
error:   /Users/bpabbott/Development/mercurial/octave-3-0-0/scripts/ 
plot/__plt__.m at line 83, column 10
error:   /Users/bpabbott/Development/mercurial/octave-3-0-0/scripts/ 
plot/plot.m at line 189, column 5

For reference the make script failed while working on the  
documentation (same error). However, "make check" went rather well  
with only one failure related to log2() and NaNs/Infs.

I've also noticed another problem ...

octave:1> backend('gnuplot')
octave:2> clear all
octave:3> backend('fltk')
octave:4> clear all
octave:5> panic: Segmentation fault -- stopping myself...
attempting to save variables to `octave-core'...
save to `octave-core' complete

Ben