Re: GNU Octave 3.4 as a Debian package?

[Jordi Gutiérrez Hermoso]

On 14 July 2011 07:56, Freddy López <address@hidden> wrote:
>>  Is this explained by any major issue?
>
> Was answered the question?

I tried to answer it: we haven't gotten around to it, and a package
will be available when it's ready, and sooner if you help. That being
said, I'm going to make a weak promise to work on Debian packaging
again in the immediate future.

> In my case, I used Linux Mint 32 bits and I had to compile Octave 3.4
> from source and it was successful, and easy, but, when I changed to 64
> bits it was impossible compile.

What was impossible? Can you report the problems?

> Some people here show a lot of hostility when we ignore some technical
> details (it's important to note that not all here have get an PhD in
> computer engineering).

I don't think anyone posting regularly in this mailing list has a PhD
in computer engineering; we're mostly from other backgrounds involving
scientific computation. I certainly don't, and I understand the reason
why packaging is a chore and the quick packages people build for
personal use are not fit for general consumption.

The technical details for Debian are important because a package must
work for all 11 architectures that are currently supported by Debian
plus any that may come in the future. This means you have to make sure
your build scripts follow the proper practices outlined in the Debian
policy manual that enable the autobuilders to properly create binaries
of your package. You have apparently already experienced this
difficulty in being unable to build Octave worrying only about two
architectures, let alone eleven. The package needs to interact
properly with the rest of the Debian system, so if it breaks other
Debian packages, that is also a consideration. In particular for
Octave, this means we have to take special care that Octave-Forge
packages also work properly.

I don't mean to grandstand about my own packaging abilities; I merely
wish to impress upon you the real chore that packaging is and why we
haven't gotten around to it. It's nevertheless of course quite
necessary because making everyone else compile is an even greater
chore overall for the community, so I reiterate my weak promise to
work on packaging soon.

- Jordi G. H.