On Wed, Jun 02, 2004 at 04:24:52PM -0500, John Goerzen wrote: > > * it doesn't include unnecessary marketing connotations, and avoids > > the issue whether we even *can* use AMD's name in vain > > I fail to see how mentioning the name of an architecture is using > unnecessary marketing connotations. Perhaps we ought to make these > changes: It is also called ia32e so which should we call it x86-64, amd64, or ia32e... ;) > > * it doesn't *quite* match the others "x86-64" vs. "x86_64" > > Which is also a serious problem that is going to lead to endless > confusion. I don't remember all the reasons given but both _ and - can cause problems in various ways. One reason not to use _ is that it is used to split on in Debian. A reason not to use - is that is used to separate arch-os triplets. I think there were some other reasons given as well. > > The second is due to "_" being used as a filename separator; I'd like to > > investigate what actually *relies* on this and potentially change the > > architecture at a later date (still before archive addition) to x86_64 > > to totally match the others -- we'll see how that plays out. > > Pick something and stick with it. Don't make us change twice. > > What I'm really upset about here is that this is a major decision that > was taken without even attempting to gather input on the lists. If the > consensus was to rename it, fine, but no attempt at gathering info was > made on your part. You should immediately s/x86-64/amd64/ in the dpkg > tree and only change it back after discussion here. For what its worth the archtable entry was incorrect in any case since dpkg --print-gnu-build-architecture was printing out incorrect information. What is up for discussion appears to be whether it should be: x86_64-linux-gnu x86-64 x86_64 x86_64-linux-gnu amd64 x86_64 Chris Cheney BTW - Scott you still forgot to fix the header doc on what the three fields are supposed to mean. Recall you told me field 3 was for --print-gnu-build-architecture not --print-architecture...
Attachment:
signature.asc
Description: Digital signature