Re: Timeline for 3.4 release?

[Tatsuro MATSUOKA]

Hello

> I am switching to mingw32 gcc-4.5.0 as it is the first release of gcc 
> that does not require manual patching to get it to work properly (and 
> hopefully solves the segfault problem...)

I have not noticed the msys  update very recently. Thanks for the information.

I have rebuild dependencies by gcc-4.5.0 as possible and tried to build octave 
on the development
branch (the latest ChangeLog date 2010-05-18 ).

Some fails exits but no seg-fault occurred during the make check !!!!

Summary:

  PASS   6359
  FAIL     11

At this moment, I have not looked at the fails but octave seems to be far more 
stable than that built
in April.
Perhaps not only gcc update to 4.5.0 but also improvement of the octave and the 
gnulib result in the
improvement of stabilities.

I could run fltk backend very smoothly.
The 3d plot

X=linspace(-pi/2,pi/2,100);    
[XX, YY] = meshgrid (X);
ZZ=sin(XX.*XX+YY.*YY);
mesh(XX,YY,ZZ);

went very fast!!!


I have execute some scripts using in my lecture.   
The scripts tested went well without trouble. 

Especially speed of the animation of successive plots in fltk backend run quite 
fast than that in
gnuplot backend.

Super!!

Perhaps Benjamin will reports similar way in the near future.

The situation of windows build is much improved than than a month before.
Please go ahead further to prepare octave 3.4.

I will try to update the msys system and see the fails in make check in detail.

Regards

Tatsuro

--- Benjamin Lindner  wrote:

> 
> > It was reported on April-17. The situation might be improved by him but I 
> > do not know how it
> is
> > improved.
> >   
> I am switching to mingw32 gcc-4.5.0 as it is the first release of gcc 
> that does not require manual patching to get it to work properly (and 
> hopefully solves the segfault problem...)
> And I am taking a closer look at the mingw64 project which provides both 
> x86 and x86_64 runtime for gcc on windows platform (and the 
> corresponding binutils/gcc suite). This one looks very promising - I got 
> the i686-w64-mingw32 platform working here. If x86_64 also works this 
> will provide the possibility for 64bit versions of octave, which would 
> be cool.
> The Msys project is also currently updating and re-releasing the entire 
> environment, as they switched to gcc-3.4 (from gcc-2.9).
> So there is quite some movement ahead on the win platform here.
> 
> I wouldn't want windows to be the only reason from keeping an octave 
> release - so if you feel comfortable with it, please go ahead.
> The situation for building octave 3.3.x on windows is kind of the same 
> with the 3.2 release. Some things now work OOTB, others no longer. It 
> wasn't exactly straightforward with 3.2 and it's not with 3.4. But it's 
> doable. So please don't let you keep from a 3.4 release.
> 
> benjamin
> 


--------------------------------------
2010 FIFA World Cup News [Yahoo!Sports/sportsnavi]
http://pr.mail.yahoo.co.jp/southafrica2010/