Re: R: octave 3.0.4 rc5

[Marco Atzeri]

--- Gio 26/2/09, Rob Mahurin ha scritto:

> Da: Rob Mahurin 
> Oggetto: Re: R: octave 3.0.4 rc5
> A: "Rafael Laboissiere" 
> Cc: address@hidden
> Data: Giovedì 26 febbraio 2009, 19:34
> On Feb 26, 2009, at 11:37 AM, Rafael Laboissiere wrote:
> > * Rob Mahurin <address@hidden> [2009-02-26 01:43]:
> >> On Feb 26, 2009, at 1:22 AM, Marco Atzeri wrote:
> >>> I can not confirm your result, 3.0.4 rc5 +
> your patch (cygwin) still crash
> >>> on Tetsuro's exdiff.m and also on
> John's
> >>>
> http://jbrwww.che.wisc.edu/home/jbraw/chemreacfun/figures.html
> >> 
> >> I believe the 3.0.4 rc5 tarball came from revision
> 7e472cf19d4a in the
> >> 3-0-x repository.  Since then there have been two
> patches
> >> 
> >>
>       http://hg.tw-math.de/release-3-0-x/rev/489d0ba75c42 to
> graphics.cc
> >>
>       http://hg.tw-math.de/release-3-0-x/rev/e1779a6d4b7a to
> legends.m
> >> 
> >> in addition to the reversion of clf.m I posted to
> the list.  I did my
> >> testing with all three changes.
> > 
> > I applied the three mentioned patches to 3.0.4-rc5 and
> the jbraw plot crash
> > test run now without the axes handle errors on my
> Debian squeeze system.
> > 
> > However, after a given point (I think after fig-4-22)
> all the graphs become
> > superimposed.  Also, the clf() command becomes
> inneffective after example
> > fig-4-15 is executed.
> > 
> > Do you observe the same behavior?
> 
> 
> I do not, on OSX 10.4.11 and some version of Red Hat
> Enterprise.
> 
> There seem to be two patches against the current 3-0-x tip
> for the axis handle bug.  One is from me, reverting clf.m to
> a 2007 version,
> 
>       
> https://www-old.cae.wisc.edu/pipermail/octave-maintainers/2009-February/011100.html
> 
> and one from Marco Atzeri, deleting a figure's children
> explicitly,
> 
>       
> https://www-old.cae.wisc.edu/pipermail/octave-maintainers/2009-February/011107.html
> 
> Both of these appear to work equally well for me.  Marco
> Atzeri's patch keeps new functionality not in 3.0.3.
> 
> Just before sending, I see another patch from Rafael
> Laboissiere,
> 
>       
> https://www-old.cae.wisc.edu/pipermail/octave-maintainers/2009-February/011111.html
> 
> which looks like it would not handle figures with multiple
> axes correctly, consistent with Rafael's observations.
> 
> Which to apply seems like a call for the release manager.
> 
> Cheers,
> Rob
> 
> --Rob Mahurin

Hi Rob,
the patches are not mine but of Marco Caliari.
I just tested them.

I merged them in the attached to simplify the check.
I used it on top of 3.0.4-rc5, and I found no graphics 
problem.
Both Tatsuro's exediff and John's book pictures 
look fine with no handle crash.

Regards
Marco Atzeri




      Passa a Yahoo! Mail.

La webmail che ti offre GRATIS spazio illimitato, 
antispam e messenger integrato.
http://it.mail.yahoo.com/

full_caliari.patch
Description: Binary data