<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8"> <META NAME="GENERATOR" CONTENT="GtkHTML/3.26.0"> </HEAD> <BODY> On Wed, 2010-06-23 at 13:01 -0400, Writer, Tim wrote: <BLOCKQUOTE TYPE=CITE> <PRE> On Wed, Jun 23 2010, Gaetan Nadon <<A HREF="mailto:memsize@videotron.ca">memsize@videotron.ca</A>> wrote: > On Wed, 2010-06-23 at 12:54 -0300, Otavio Salvador wrote: > > Just to be sure, I am not advocated to compile under 64 bit *and* to > use the produced driver library. It's to prevent build failure and > allow compile testing for some types of changes. That seems bizarre and perhaps dangerous to me. Imagine an automated packing system running on x86_64, building and packing iX86 binaries. With this arrangement, it will silently fail, producing and packaging binaries that appear to be okay but fail to run on Geode LX. In my view, building on x86_64 for iX86 should either succeed in producing fully functional binaries or fail noisily. Your use case (of testing partial compilation while knowingly producing an incorrect build) seems to be a specialized use rather than the common case. While I understand your rationale for wanting it to work, I'm concerned about making it work at the expense of introducing surprises that may bite a much wider audience. </PRE> </BLOCKQUOTE> I don't disagree with you in principle. It took me a while to figure out the release flow of X.Org. Only source code is produced, so the O/S or distros builders cherry pick the packages. For example, Mac O/S do not use any of the drivers, SUN will only use a subset, Linux will not use drivers that only works on SUN and so on. There is no risk of seeing a geode driver shipped on Solaris. Those who are building the whole 240 packages is a special case indeed. There is a small percentage of build and code maintenance that is done on a batch of packages (e.g. 50 video drivers). The more packages that build on mainstream platforms like Linux/Solaris the better it is. I don't want to force the issue. I am under the assumption that there would be no impact at all for the geode developers. I can live with status quo. <BLOCKQUOTE TYPE=CITE> <PRE> Tim </PRE> </BLOCKQUOTE> </BODY> </HTML>