On Tue, 2009-02-17 at 18:12 -0800, Jesse Keating wrote: > On Tue, 2009-02-17 at 20:42 -0500, Jon Masters wrote: > > Er...I might not have thought this through enough but the reason I > > brought this up on IRC the other day was that I'm not convinced this > > won't be equivalent if you're changing the arch and compiler out > > underneath. Where will the i586 versions of those packages be to > > bootstrap with? > > I don't believe any library paths are changing, lib/ isn't suddenly > becoming lib586/. We've had i586 builds going for a day or so now, with > no fallout yet. Of course, if fallout is discovered before the script > starts, we can make some decisions then. Until then... What I meant was that, although on the face of it you can just link against i386 or i586 and the ABI compatibility will see you through, there are probably packages that have build hacks which will break horribly on discovering that i586 has been substituted. Some of those will be due to some binary munging, others will be more along the lines of what was already raised. > > And what will happen when something wants a static > > library file to link against? > > This one is a little bit more interesting, perhaps everything that is > static should get a second rebuild pass. Of course, now I'm going to > want a good programmatic way of discovering what is statically compiled, > preferably without having to look at the binaries themselves. Indeed. This was my *primary* concern (but the above one should not be ignored) and I think gives serious cause to do a second rebuild for good measure. I would at the very least suggest a complete second pass just to be 100% certain that a). everything was caught that might break builds later b). there isn't something built static that is using i386 code when we "promised" i586. You're doing a great job...and I'm sorry to sound like a grumbler :) I'm just trying to get more involved with this side of things. Can I propose a complete second rebuild rather than a piecemeal rebuild as a compromise to doing some kind of full ordering based build process? Unless you think there's not time for a complete second pass...in which case I'll try to think of alternatives. I don't have a good way to search for static linking without examining bins, but I've got a cold and feel like utter poo, so maybe I'm missing something! Jon. -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
