x86_64 bug

[Jan Engelhardt]

On Mar 4 2008 20:27, Gustavo Niemeyer wrote:
>
>> when there is a package A at x86_64 that is being obsoleted by B, then 
>> smart will install both B at i586 and B at x86_64, which is really wrong IMHO.
>
>Yes, that's wrong, and that's not what the stock Smart should
>normally do in that situation.  Can you please provide the output
>of "smart --explain upgrade" when you see this situation again?

01:10 ccgmbh:../packages/RPMS # rpm -Uhv a-0-0.x86_64.rpm 
Preparing...                ########################################### [100%]
   1:a                      ########################################### [100%]
01:10 ccgmbh:../packages/RPMS # smart install b
Loading cache...
Updating cache...               ######################################## [100%]

Computing transaction...

Upgrading packages (2):
  b-0-0 at i586     b-0-0 at x86_64   

2.3kB of package files are needed. 

Confirm changes? (Y/n): 
Saving cache...

01:10 ccgmbh:../packages/RPMS # smart install --explain b
Loading cache...
Updating cache...               ######################################## [100%]

Computing transaction...

Upgrading packages (2):
  b-0-0 at i586
    Upgrades:
      a-0-0 at x86_64 (upgraded)
    Conflicts:
      a-0-0 at x86_64 (upgraded)
  b-0-0 at x86_64
    Upgrades:
      a-0-0 at x86_64 (upgraded)
    Conflicts:
      a-0-0 at x86_64 (upgraded)

2.3kB of package files are needed. 

Confirm changes? (Y/n): 



BTW, the Conflicts is also wrong I guess, but that may be
an implementation detail of smart.