Robert Millan <[EMAIL PROTECTED]> writes:
> On Mon, Jan 24, 2005 at 09:08:02AM +0000, Scott James Remnant wrote:
>> On Mon, 2005-01-24 at 04:14 +0100, Robert Millan wrote:
>>
>> > This is the continuation of bug #118910. Since you retitled it as
>> > "[ARCH] clean up archtable", I considered you'd prefer me to file a
>> > separate
>> > bug for this. Feel free to retitle/merge if you think otherwise.
>> >
>> > The changes are pretty straightforwarded, but they should be enough to fix
>> > most of the concerns that porters had with the limitations of dpkg
>> > architecture
>> > handling during the last few years.
>> >
>> This patch looks incomplete? It only touches the dpkg-dev part of dpkg,
>> and doesn't change the handling of dpkg itself.
>
> As I just said to Goswin, my patch doesn't modify the way dpkg handles
> architectures internaly. dpkg-genchanges uses Cpu/System logic to determine
> wether we can build a package, but when generating DEBIAN/control it will add
> an "Architecture" field set to DEB_HOST_ARCH.
Could we automatically define some @linux@ or @any-i386@ variables the
same way shlidbs or other substitutions work?
The split basically just means before looking you translate the
splited value into the old value and then compare. A substitution
variable should be able to do the same without introducing a new
syntax.
> Of course, I'd like a full transition on the long term. But keeping backwards
> compatibility is the most important for me at this point. IMHO, you can
> do this change safely in experimental without disrupting the sarge release.
>
> (Although we won't really start using this feature untill the buildds have
> a similar fix applied and we ammend Policy to require use of this syntax).
MfG
Goswin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
