Re: Outerposition Patch

[Ben Abbott]

On Feb 12, 2011, at 11:41 PM, logari81 wrote:

> On Sat, 2011-02-12 at 12:46 -0500, Ben Abbott wrote:
>> On Feb 12, 2011, at 9:20 AM, logari81 wrote:
>> 
>>> On Fri, 2011-02-11 at 18:45 -0500, Ben Abbott wrote:
>>>> On Feb 10, 2011, at 7:04 PM, Ben Abbott wrote:
>>>> 
>>>>> On Feb 9, 2011, at 2:39 PM, Søren Hauberg wrote:
>>>>> 
>>>>>> ons, 09 02 2011 kl. 19:28 +0000, skrev bpabbott:
>>>>>> 
>>>>>>>> dump_demos.m and friends are becoming a valuable debugging tool. Can
>>>>>>>> we commit them to Savannah?
>>>>>>>> 
>>>>>>>> - Jordi G. H.
>>>>>>> 
>>>>>>> I like the suggestion. One problem is that the dump_demos.m script
>>>>>>> produces a script with syntax errors. Essentially, when the syntax is
>>>>>>> tweaked to work under Matlab, some features are missed. I'll take a
>>>>>>> look at the details. Maybe it is possible to have the resulting script
>>>>>>> run error free under Octave.
>>>>>>> 
>>>>>>> Ben
>>>>>> 
>>>>>> The problem is that some demos are written in Octave syntax (i.e. using
>>>>>> " for strings, "endfor" to stop loops and so forth). This required me to
>>>>>> make some hacks in the code to semi-automatically create matlab
>>>>>> compatible code; this is just a mess. It would be more simple if the
>>>>>> demos could be changed to be matlab compatible.
>>>>>> 
>>>>>> Søren
>>>>> 
>>>>> I've pushed a change to allow all demo scripts to be converted to Matlab 
>>>>> syntax using Soren's simple approach.
>>>>> 
>>>>>   http://hg.savannah.gnu.org/hgweb/octave/rev/a754c2d8a13f
>>>>> 
>>>>> Soren, if it's ok with you, I'll prepare a changeset to push dump_demos.m 
>>>>> as well as a script I wrote to produce your html web-page.
>>>>> 
>>>>> Ben
>>>> 
>>>> I have more work to do, but the link below illustrates the result I get 
>>>> now (tip 12437)
>>>> 
>>>>    http://homepage.mac.com/bpabbott/12437/compare_plots.html
>>>> 
>>>> Ben
>>>> 
>>> It seems that it is not clear 12437 but you have applied the patch I
>>> have sent here:
>>> https://mailman.cae.wisc.edu/pipermail/octave-maintainers/2011-February/022955.html
>>> 
>>> If this is true I see no regressions, should I push it?
>>> 
>>> Kostas
>> 
>> Yes the path is applied.
>> 
>> There are two regressions. They can both be seen in the subplot demos.
>> 
>> (1) See the yticklabel for subplot (3, 1, 1) which is created by demo 
>> ("subplot", 1)
>> 
> I cannot confirm this problem. On my machine both the plot on the window
> and the png output are correct.
> 
>> (2) The ylabels for all subplots are too tight. This can be see in either 
>> demo, demo ("subplot", 1:2)
>> 
> Yes you are right, but this is not a regression. ylabels always used to
> be too tight to the axes on the png output. If you consider the plot in
> the window they are not so tight. Actually this is a issue in gl2ps that
> I would like to file a bug report for, but it is not new.
> 
>> Ben
>> 
> Kostas

I've pulled again and applied "sync-position-outerposition-2a.changeset". I'm 
now getting the correct result under MacOS and Ubuntu.

        http://homepage.mac.com/bpabbott/12439/compare_plots.html

Please push this changeset.

Ben