On Mon, 08 Jun 2009 16:58:46 +0200 hoefle marco <marco.hoefle@nanotronic.ch> wrote: > On Thu, 2009-06-04 at 19:16 +0100, Neil Williams wrote: > > > However, this will only work if you can use the packages from Lenny > > to provide the cross-dependencies. > > > > Crush 1.0 is a proof-of-concept - a test case. What the proof showed > > was that the system to generate Crush required extensive changes to > > make it suitable for Crush 2.0. :-( > > > > The changes are sufficiently complex that there is a chance Crush > > might not make a 2.0 release at all. > > :-( that makes crush not available for a large amount of boards in the > embedded world. I deal a lot with boards having typical flash sizes > around 32 MBytes which will be to less for grip. I don't know what the > actual GRIP:CRUSH installation ratio is but I suppose there is a need > for crush. Crush 1.0 can fit in 24Mb installed. The smallest Grip install is 75Mb and is extremely unlikely to ever be any smaller - if anything, Grip will gradually creep up to a 80 or 90Mb minimum. Crush could get down to 10Mb installed and quite possibly smaller. Quite how small Crush 2.0 will be is largely down to you - if you want it, get involved in making it. It is a *lot* of work. > I am writing to encourage you guys to keep crush :-) Encouragement is not going to change anything. Developers are what is necessary for Crush to be retained. There needs to be a *lot* more people asking for cross-build support in the relevant packages, filing cross-build support bugs, pinging cross-build support bugs that have been ignored for a year already, making noises on mailing lists that cross-building matters to more people than just me. More people making it clear to various people in Debian that Crush is not my personal mission, cross-building requirements are not something I've personally invented for a lack of something better to do, that it matters to more people than just me and that there is more than just one person actively seeking this kind of support in Debian. Not just one-off messages - consistent, persistent and relevant input on lots of mailing lists and lots of bug reports, answering all queries raised with calm, logical and useful input. Real user-case scenarios, real input with relevant data. *Explain* why cross-building matters to you. Read all the existing cross-build support bugs and provide reasons and solutions for issues raised by maintainers. If *you* want Debian to support cross-building, *you* need to speak up across all of Debian and keep on talking until people listen. My single voice is being actively ignored in some quarters - principally because it is a single voice and has been since before Lenny was released with Crush 1.0. Extend the reach of Crush from just GPE that we have already, get other maintainers involved, get more packages to cross-build (like python), give Crush more options and wider support. Make replacements for perl scripts like adduser, update-alternatives and dpkg-divert. You need to create the patches and file the bugs and chase up the bugs. Get Crush on a wider range of architectures, make sure the packages build on those architectures, speak up here as a possible Crush autobuilder maintainer for a particular architecture - once the autobuilders can be re-enabled. (This is a *lot* more work than a Debian autobuilder, you will need to actively get involved with fixing the breakages on your selected architecture and feeding back on issues that arise etc.) Part of the reason for releasing Crush 1.0 was to get more people involved. Involvement doesn't mean just using Crush, it means pushing the goals of Crush within Debian as a whole. I cannot ensure the release of Crush 2.0 myself. If more people do not actively engage (at a level of a few hours a week, every week until the end of the Squeeze release freeze), there is a high chance that Crush 2.0 simply will not release. Encouragement will not change the reality. It is up to you. I'm doing what I can but I do not think that it will be enough. If that upsets to you, good! Now is the time to do something to do something to change things. If all that is forthcoming is encouragement, Crush 2.0 is not likely to arrive. It is as simple as that. -- Neil Williams ============= http://www.data-freedom.org/ http://www.linux.codehelp.co.uk/ http://e-mail.is-not-s.ms/
Attachment:
pgpsISjcZABh0.pgp
Description: PGP signature