Re: 3.6.0 release

[Michael Goffioul]

On Tue, Nov 29, 2011 at 7:59 PM, Michael D Godfrey
<address@hidden> wrote:
> On 11/29/2011 10:45 AM, John W. Eaton wrote:
>
> Yes.  I even did the bisecting twice because I couldn't believe that
> this could be the changset that caused it.  Then I tried the above
> change on the current sources with the same effect.
>
> | This must the symptom of something else (like uninitialized memory).
> | Forcing the use of the octave allocator is IMO not a solution.
>
> Absolutely, this can't be the fix.  But when I run with valgrind with
> and without the above change, nothing obvious about uninitialized
> memory pops up.  However, there are some messages related to
> fontconfig (I think) in both cases.  I'll check again and try to send
> details later.
>
> jwe
>
> I should have added two things:
> 1. I am using ftlk for these tests.
> 2. It would be good to run some of the other plot
>     tests to see if the results are changed by this "fix."

Something that's worth checking is to dump the tree of graphics
objects with their properties (including the hidden ones) and check
whether the text objects are present in the legend axes. This would
maybe help to determine why they are not visible.

I suspect it's due to the chain of events that happening when
something changes in an axes object (recomputing the limits...).

Michael.