[Top] [All Lists]

Re: Bigsur?

To: Laurent GUERBY <>
Subject: Re: Bigsur?
From: Ralf Baechle <>
Date: Tue, 19 May 2009 13:53:10 +0100
Cc: Jon Fraser <>, Andrew Wiley <>, "" <>, "Maciej W. Rozycki" <>
In-reply-to: <1242735440.6098.101.camel@localhost>
Original-recipient: rfc822;
References: <> <> <> <> <1242735440.6098.101.camel@localhost>
User-agent: Mutt/1.5.18 (2008-05-17)
On Tue, May 19, 2009 at 02:17:20PM +0200, Laurent GUERBY wrote:

> On Mon, 2009-05-18 at 23:47 +0100, Maciej W. Rozycki wrote:
> >  Yes, invaluable for native builds and there is a considerable number of 
> > software packages which is not capable of being cross-compiled, or 
> > requires extreme contortions to be built this way, or if buildable with a 
> > reasonable effort, the functionality is limited.  Besides a three-stage 
> > GCC bootstrap is a good way of verifying the quality of the tool, never 
> > mind standard DejaGNU-based regression testing which although possible 
> > using cross-tools and a remote target, is awfully painful to be set up 
> > this way.
> For MIPS in the GCC Compile Farm
> (which is open to all free software, not limited to GCC) we
> have two loongson-2f based netbooks on which a GCC bootstrap
> and check is manageable.
> Right now this farm is more oriented towards upstream userland
> developpers debug/test cycles - they get access to 12 architectures when
> they sign in. It's not really oriented towards porting
> kernel/distributions or building distribution packages which is already
> well covered by existing distribution farms and individual developpers
> and those developpers should get priority on new hardware :).
> This farm project is part of the Free Software Fundation France (a
> french not-for-profit organization) effort to help free software
> development and we accept hardware and hosting donations, and also
> discounts to purchase commercial hardware when donations are not
> possible and there is significant interest in one platform.

That I think is a great solution for people who need short term access
but doesn't really solve the fundamental problem that hardware with
sufficient punch for efficient native development isn't easily available.


<Prev in Thread] Current Thread [Next in Thread>