Re: regression for legend()

[Rik]

On 11/23/2011 06:29 AM, address@hidden wrote:
> On Nov 23, 2011, at 8:17 AM, logari81 wrote:
>
>> > On 11/22/2011 10:36 PM, John W. Eaton wrote:
>>> >> I would like to focus on making the 3.6.0 release soon.
>>> >> 
>>> >> If you know of important problems that should be fixed, please make
>>> >> sure they are in the bug tracker and listed with a serverity of
>>> >> important or blocker.  That way we can sort the list of reported bugs
>>> >> by severity and work on the most important bugs first.  But please
>>> >> don't tag every problem that you would like to see fixed as a blocker,
>>> >> as that would not be not helpful.  Here are some guidelines to decide
>>> >> whether a problem should be tagged with blocker or important severity:
>>> >> 
>>> >>   * the problem is a regression (i.e., the feature worked in a
>>> >>     previous version of Octave)
>>> >> 
>>> >>   * the problem causes a crash that can be duplicated on multiple
>>> >>     systems and is something that is likely to affect many users.
>>> >> 
>>> >>   * the problem leads to incorrect results, can be duplicated on
>>> >>     multiple systems, and is likely to affect many users.
>>> >> 
>>> >> jwe
>> > I have just found a regression in the fltk backend. Legend text will not 
>> > show up with the current default branch:
>> > 
>> > octave:1> graphics_toolkit fltk
>> > octave:2> plot(0:5,0:5)
>> > octave:3> legend('test text')
>> > 
>> > if this is confirmed and no one else is available for fixing it, I will do 
>> > it before the 3.6 release.
>> > 
>> > Kostas
> Your example works for me on MacOS with fltk-1.1.10-aqua
>
> However, I notice that some of the diplaynames are missing when running 
> demos. This is not repeatable. When running the demos a second time, I see 
> different missing displaynames.
>
> Ben
However, I can confirm the problem on a Linux build from today (11/23/11)
so it is real for some platforms.

--Rik