gnustep-dev
[Top][All Lists]
Advanced

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

Re: GNUstep packaging by TWW tools


From: Richard Frith-Macdonald
Subject: Re: GNUstep packaging by TWW tools
Date: Wed, 27 Dec 2006 10:59:51 +0000


On 26 Dec 2006, at 20:02, T.J. Yang wrote:

My plan
1. prepare gcc-4.1.1 for Solaris 10 intel (U2, 06/2006 version) first and down port lower version solaris and sparc cpu.

This is fairly straightforward ... I've done it for 32 bit and 64bit solaris ... you shouldn't have any trouble.

2. package the gnustep core software(make,base,gui).

I don't know CPAM at all ... but I've packaged these things using solaris native packaging (pkgmk etc) and a cursory scan of the pages at http://en.wikibooks.org/wiki/CPAM_with_TWW/User_Guide suggests that CPAM acts as a layer on top of pkgm, so it ought to be workable.

3. release package sources

You will need to make a copyright assignment to the FSF to get the packaging source incorporated into the projects. See http://mediawiki.gnustep.org/index.php/ Developer_FAQ#How_do_I_assign_my_contribution.3F

4. Lobby gnustep development community to use TWW tool so
  TWW CPAM tool can help GNUstep's CPAD.
This will be hard because asking people to switch to different tools using
  different processes.

I don't know CPAM details, but if it produces a wrapper round a 'native' package format, such that the native package is still available, I expect there would be no resistance as it would allow us to build (and therefore provide easily on the ftp site) both the native packages and the higher-level CPAM packages. If on the other hand, it results in something which can only be installed with the CPAM installer, I expect it would be argued that we should just build packages in the 'native' formats, so that they can be installed without the need to download/use the CPAM tool.

One issue though ... the list of supported operating systems for HPMS does not include ms-windows ... the whole point of this system is to wrap all other systems inside a single toolset, but if it omits what is arguably the second most important target operating system, then it's probably not actually very useful.

I think it's therefore important to find out whether ms-windows support is available, or under development and near enough complete for you to join in and perfect it. If ms-windows support is available then this sounds like a very good idea.






reply via email to

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