Comment 2 for bug 1050948

<davmor2> mvo: so looking at subscriptions it say natty
<mvo> davmor2: ok, this subscription dates all the way back to natty? and back then it was i386/amd64 and later it was only i386?
 davmor2: just to ensure I understand this correctly
<davmor2> mvo: correct and now I believe it is being made multiarch now that the i386 version installs correctly on amd64
<mvo> davmor2: aha, so now == precise? and there its multiarch?
<davmor2> mvo: Yes so we tried it recently on precise, I'm not sure if it is multiarch now or whether jpugh was looking to make it multi
 mvo: but until really recently it was i386 only
<mvo> davmor2: it may well be that the multiarch change is the issue
<davmor2> mvo: no I'm pretty sure it is setting it to natty's i386/amd64 version
<mvo> achuni: we may need to talk about the multiarch strategy for the reinstall-previous-purchases as ideally the agent would send a "foo:i386" packagename then for i386-multiarch packages on amd64
<achuni> mvo: ouch
<mvo> davmor2: :) I need to figure out a way to simulate this, I may ask you for the json output of this particular part of the "available_for_me" call - but more on that later
 davmor2: need to go to a call first
<achuni> mvo: ok...
<mvo> achuni: yes :(
 achuni: lets talk after the call, its a bit anoying, maybe in this case-software-center needs to figure it out
<achuni> mvo: if we add an optional arg to the call to specify the architecture, the server could do it in a backwards-compatible way