Re: Binary distributions (was: Re: Release goals for 3.6)

[PhilipNienhuis]

John W. Eaton wrote:
> 
> On  2-Aug-2011, Jordi Gutiérrez Hermoso wrote:
> 
> | On 2 August 2011 12:08, John W. Eaton <address@hidden> wrote:
> | >  so far we don't seem to have even one volunteer who really appears
> | > to want to take on the job of building binaries for Windows.
> | 
> | That's not fair, we have several who have worked a lot on Windows
> | builds. Tatsuro Matsuoka, Benjamin Lindner, Michael Goffioul, to name
> | a few.
> 
> I did not mean that we had never had someone building binaries.  But
> are any of these people actively building and distributing a binary
> installer that typical Windows users expect?
> 

It is my plan (and back on my priority list) to help Tatsuro with this. I
was underway but got distracted.

Once a number of MinGW specific issues are ironed out, it wouldn't be too
hard to transform Tatsuro's binary and his build environment into an
"installer" (AFAIU a mere glorified unzipper with some bells and whistles).
Some days or weeks back I've seen a post from someone who offered help
there.

But to avoid disappointment for the "wider audience", or as you name it:
"typical Windows user" (and potential backfiring to Octave's reputation) I
think more test reports are needed.
Perhaps a chicken-and-egg problem: maybe we first need an "alpha" / "beta"
binary or RC to collect more user reports?

I already have a rough (in the sense of overly complete) "zippable" "binary"
that I use on a daily basis.
Tatsuro has done all the hard work, a beautiful piece of work at that, what
I did was merely adding in Tatsuro's msys and mingw stuff after the 3.2.4
MinGW template so that some octave-forge packages can be compiled (Tatsuro's
binary couldn't compile the java and windows packages, and IIRC the
linear-algebra one, without those). 
MinGW issues solved, some cleaning up, a few tweaks to automagically set the
proper paths, and then the stage is set for the installer sensu stricto to
be made.

Now that Octave can install packages directly from the web, building a
Windows installer became a lot easier. I'd only supply a pre-compiled Java
package in the MinGW binary as that would save users from a.o., having to
install a Java JDK.
The Windows package has issues with 3.4.2; I started debugging this but
dropped it due to lack of time and gdb experience.

So you see, there is /some/ activity. But I expect that coming Octave-3.4.3
might be the basis for a MinGW binary.

Philip


--
View this message in context: 
http://octave.1599824.n4.nabble.com/Release-goals-for-3-6-tp3711420p3713700.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.