Re: FLTK changesets [was: branching for release?]

[Ben Abbott]

On Jan 22, 2011, at 9:35 AM, Ben Abbott wrote:

> On Jan 22, 2011, at 7:38 AM, logari81 wrote:
> 
>> On Sat, 2011-01-22 at 12:35 +0100, logari81 wrote:
>>> On Fri, 2011-01-21 at 23:14 -0500, Ben Abbott wrote:
>>>> On Jan 21, 2011, at 4:39 PM, logari81 wrote:
>>>> 
>>>>> On Thu, 2011-01-20 at 20:01 -0500, Ben Abbott wrote:
>>>>>> On Jan 20, 2011, at 6:59 PM, Ben Abbott wrote:
>>>>>> 
>>>>>>> On Jan 19, 2011, at 3:58 PM, logari81 wrote:
>>>>>>> 
>>>>>>>> 1. Current version of the fltk backend doesn't show minor ticks/grid,
>>>>>>>> there is no bug report for this, but there is a patch:
>>>>>>>> https://mailman.cae.wisc.edu/pipermail/octave-maintainers/2011-January/022173.html
>>>>>>>> 
>>>>>>>> 2. a)Color of axis labels / b)Positioning of axis labels at right and
>>>>>>>> top:
>>>>>>>> https://savannah.gnu.org/bugs/?31800
>>>>>>>> a) there is a patch for the colors, waiting for evaluation
>>>>>>>> b) I can provide a patch for the positioning after the patch of (1) is
>>>>>>>> accepted/pushed 
>>>>>>>> 
>>>>>>>> 3. I would also like to apply this patch:
>>>>>>>> https://mailman.cae.wisc.edu/pipermail/octave-maintainers/2011-January/022108.html
>>>>>>>> but this also depend on the progress on (1)
>>>>>>>> 
>>>>>>>> 4. Axes Position/Outerpositioning/Tightinset synchronization
>>>>>>>> Probably will not do it for 3.4. I have a plan for fixing this but it 
>>>>>>>> is
>>>>>>>> quite a long of work still (also depends on (1) and (3))
>>>>>>>> 
>>>>>>>> 
>>>>>>>> Concerning the replacement of "backend" I don't have any fixed opinion.
>>>>>>>> Concerning replacing "fltk" I would vote against "fltk_opengl", I would
>>>>>>>> prefer one of "fltk" or "opengl".
>>>>>>>> 
>>>>>>>> Best regards
>>>>>>>> 
>>>>>>>> Kostas
>>>>>>> 
>>>>>>> I hope to look at the three changesets below in the next few days.
>>>>>>> 
>>>>>>>         1. minorticks-restructure-draw_axes.changeset
>>>>>>>         2. bug-31800-color.changeset
>>>>>>>         3. axes-ticks-positioning.changeset 
>>>>>>> 
>>>>>>> My intention is to run Soren's dump_demo script and make sure there are 
>>>>>>> no obvious reversions.
>>>>>>> 
>>>>>>> I'll respond back to this thread with what I find.
>>>>>>> 
>>>>>>> If there are no objections, and there are no reversions, I'll ask 
>>>>>>> Kostas to push.
>>>>>>> 
>>>>>>> Ben
>>>>>> 
>>>>>> Kostas,
>>>>>> 
>>>>>> Regarding minorticks-restructure-draw_axes.changeset
>>>>>> 
>>>>>> Please take a look at the demo for "grid"
>>>>>> 
>>>>>>  demo ("grid", 1)
>>>>>> 
>>>>>> I've attached the result I get using Matlab.
>>>>>> 
>>>>>> Essentially, when the minor grid is turned on the major one becomes a 
>>>>>> solid line.
>>>>>> 
>>>>>> Ben
>>>>> 
>>>>> Thank you Ben for the testing. I have just pushed the changeset for (2).
>>>>> I have also merged the patches for (1) and (3) into one and re-based the
>>>>> result relatively to the current status. You can find this combined
>>>>> patch in the attachment. The attached patch should additionally solve
>>>>> the positioning problem in Bug 31800 (demo("plotyy",1)).
>>>>> 
>>>>> Sorry for the big patch but recent whitespace changes make it very
>>>>> difficult to work with the older separate individual patches. This patch
>>>>> reduces gl-render.cc from 3441 to 3239 lines!!
>>>>> 
>>>>> Concerning the behavior of ML when minorgrid is enabled, I couldn't
>>>>> verify your results with the ML version, that I have access to. Probably
>>>>> is quite easy to make grid lines continuous when minor grid is enabled,
>>>>> but I would prefer to try it after the branching for 3.4 .
>>>>> 
>>>>> Please consider the attached patch for inclusion in 3.4.
>>>>> 
>>>>> Regards
>>>>> 
>>>>> Kostas
>>>>> <minorticks-restructure-draw_axes-bug-31800.changeset>
>>>> 
>>>> Kostas, 
>>>> 
>>>> I've run Soren's dump_demo's script using the FLTK backend, and don't see 
>>>> any regressions. Using FLTK on MacOS doesn't always give the same results 
>>>> as on Linux, but the link below compare the plots obtained via 
>>>> Octave/gnupot, Matlab 2010b, and Octave/FLTK. The files are in the process 
>>>> of being uploaded. The link should be function within 30 mins of the time 
>>>> of this email.
>>>> 
>>>>    https://files.me.com/bpabbott/kuq28y
>>>> 
>>>> Please push this changeset.
>>>> 
>>>> Ben
>>> 
>>> I have just pushed it. I will keep an eye on it, please report any
>>> problems or objections.
>>> 
>>> Kostas
>> 
>> I have noticed in your tests here:
>> http://homepage.mac.com/bpabbott/kostas2/compare_plots.html
>> that in plotyy_1_1 the tick marks of the second y axis will cover the
>> tick marks of the primary y axis at the left. Even if the problem has
>> surfaced now, I do not consider it as a problem in the backend but a
>> problem in plotyy.m, which leaves property 'box' set to on for both
>> primary and secondary axes. For comparison, plotyy in ML sets boxes of
>> both axes to off by default.
>> 
>> I believe this small issue should be fixed in plotyy.m like e.g. in the
>> attached file.
>> 
>> BR
>> 
>> Kostas
>> <plotyy.m>
> 
> I've pushed your change.
> 
>       http://hg.savannah.gnu.org/hgweb/octave/rev/e916491cbb99
> 
> Thanks
> Ben

While running Soren's dump_demos, I noticed that the printed output still 
colors both y-axes blue.

I'll file a bug report.

Ben