On Feb 11, 2008 5:58 PM, Michael Homer <[EMAIL PROTECTED]> wrote: > > > + if 'USE' in os.environ: > > > + mergeFlags(flags, os.environ['USE'].split(' ')) > > > > Should we ship an empty/commented Resources/Environment file with > > Compile just to illustrate how one is expected to define it? I think > > it makes sense. > I wasn't really considering people setting them that way - the > environment variable was more for one-time overrides on the command > line, or testing recipes (`USE="-foo -bar +baz" Compile > /F/C/LocalRecipes/MyNewRecipe`). > > My view was of them layered in this fashion: > 1) The "system" flags (the ones that are used in building official > packages, most likely) which would probably be somewhere in > Scripts/Current/Data. That isn't implemented yet, since I'm not sure > if it will turn out to be necessary. > 2) The site-specific flags in /S/S/UseFlags.conf > 3) The one-shot flags set in the environment or on the command line > Each level would overwrite flags set in the previous levels. > > It would still be good to document the variable somehow, but probably > not by implying that's how flags should be set overall.
Ok, I see. Ideally that should be documented in the "man page", which we don't have -- and I don't like too much the idea of polluting Compile's --help dump with environment variables that affect the build. Maybe having that listed in the wiki is enough, just like the NO_UNIONFS var. -- Lucas powered by /dev/dsp _______________________________________________ gobolinux-devel mailing list gobolinux-devel@lists.gobolinux.org http://lists.gobolinux.org/mailman/listinfo/gobolinux-devel