octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Short report Cygwin Octave-2.9.13


From: Tatsuro MATSUOKA
Subject: Re: Short report Cygwin Octave-2.9.13
Date: Thu, 23 Aug 2007 09:13:34 +0900 (JST)

Dear David

It is my great honor for your inviting me octave-forge.

As you pointed out, '--disable-silj-exception' flag on GCC
is not comtatible standard cygwin install.
So my current distribution, the GCC configured with this
flag was basically prepared only for the octave.
In the octave preparation, 

export CC='/usr/local/gcc-3.4.4d/bin/gcc'
export CXX='/usr/local/gcc-3.4.4d/bin/g++'
   :
   :
./configure .....

The mkoctfile uses the absolute path to the GCC so that
the confliction with the standard GCC on the Cygwin is
avoided.

I found the following post from search engine, so I tried
gcc-3.4.4-3.
http://sourceware.org/ml/cygwin/2004-11/msg00007.html

As for the standalone package, do you mean the package
like the octave-forge 2.1.73 ?

In my opinion, it is questionable to distribute the
standalone packge like the the octave-forge 2.1.73.
The MSVC octave-2.9.13 is now officially distributed, I
think that most windows users use it.

The purpose of my binary distrubuion of the cygwin octave
is for daily users of cygwin like me.
(I currently uses the octave and the cygwin gnuplot. )

Consulting the history of octave-forge windows
distribution, I found the following 

octave-2.1.42-cygwin (2003-03-03 21:00) 
  octave-2.1.42-gnuplot-octaveforge-...
     :
     :

I think that this type of distribution will be reasonable.

However, to avoid confusion, the binary should not be
installed in /usr/bin or /usr/local/bin.
I think that it will be installed like
/usr/bin/octave-2.9.13.

In the directry, all packages, octave, gnuplot4.3, GCC,
the related library, octave-forge functions.

My idea is:

1. Basic Packege : octave and gnuplot4.3
2. Octave-forge binary package and its octave script
installer.
3. Developer kit : GCC and the related libraries and
included files.

Thank you again and it will be grateful for me if you will
give me the reply.

I am also waiting for other contributors opinion.

Sincerely,
T. Matsuoka

 
 

--- David Bateman <address@hidden> wrote:

> Hi Tatsuro,
> 
> Sure if you want to maintain a cygwin binary I see
> no issue. However
> does the "--disable-sjlj-exceptions" flag make it
> incompatible with a
> standard cygwin install? I believe it does and in
> that case you'd
> probably have to distribute a standalone package
> like the one on
> octave-forge. Its probably also an idea to include
> the compiler in this
> case as asking the users to recompile the gnu
> compiler suite with the
> "--disable-sjlj-exceptions" to be able to compile
> oct-files might be
> asking too much...
> 
> If you do this, I think it would be a good idea to
> host the binary on
> octave-forge..
> 
> Cheers
> David
> 
> 
> Tatsuro MATSUOKA wrote:
> > Today I have built gcc-3.4.4-3 configured with
> > --disable-sjlj-exceptions on Cygwin and buit
> octave-2.9.13
> > as a try.
> >
> > Exit bug problem was resolved.  The library used
> was still
> > constructed by the gcc-3.3.3-3 configured with 
> > --disable-sjlj-exceptions.
> >
> > Execute speed the lsode solving oregonator.m
> function is
> > alomost the same as the gcc-3.3.3-3 case. 
> However, the
> > the lsode solving oregonator.oct is about 10%
> faster than
> > the previous one.
> >
> > Therefore I would like to reconstruct the
> libraries by the
> > speciallly configured gcc-3.4.4-3 and the upload
> on the
> > web.
> >
> > Recently I have fount that Xming, which is the X
> server
> > constructed with the mingw.  Cygwin/X had not been
> > upadated for a long time.  However Xming is still
> updated
> > and its ability is good.
> > The gnuplot speed on xming is improved.
> >
> > At the moment the gnuplot x11 is more stable and
> faster
> > than the wgnuplot on wxt terminal.
> >
> > In my opnion, the octave on cygwin is still
> meaningful for
> > the Cygwin users.
> >
> > T. Matsuoka
> >
> >
> > T. Matsuoka.
> >    
> >
> > --------------------------------------
> > Easy + Joy + Powerful = Yahoo! Bookmarks x Toolbar
> > http://pr.mail.yahoo.co.jp/toolbar/
> >
> >
> >   
> 
> 
> -- 
> David Bateman                               
> address@hidden
> Motorola Labs - Paris                        +33 1
> 69 35 48 04 (Ph) 
> Parc Les Algorithmes, Commune de St Aubin    +33 6
> 72 01 06 33 (Mob) 
> 91193 Gif-Sur-Yvette FRANCE                  +33 1
> 69 35 77 01 (Fax) 
> 
> The information contained in this communication has
> been classified as: 
> 
> [x] General Business Information 
> [ ] Motorola Internal Use Only 
> [ ] Motorola Confidential Proprietary
> 
> 


--------------------------------------
Easy + Joy + Powerful = Yahoo! Bookmarks x Toolbar
http://pr.mail.yahoo.co.jp/toolbar/



reply via email to

[Prev in Thread] Current Thread [Next in Thread]