[Rpm-maint] [Suse patch] Test architecture in refresh

[Panu Matilainen]

Hmm.. I fail to see any behavior change with this patch, but maybe I'm 
testing something else. This would seem to be related to the 
nameversioncompare.diff to some extent - Michael?

 	- Panu -

---

Also test architecture in "refresh" test when not colored. This allows
updates to different architecture possible again.

--- ./lib/psm.c.orig	2005-12-14 18:59:10.000000000 +0000
+++ ./lib/psm.c	2006-02-24 11:46:54.000000000 +0000
@@ -1425,15 +1441,16 @@ rpmRC rpmpsmStage(rpmpsm psm, pkgStage s

  assert(psm->mi == NULL);
  	    psm->mi = rpmtsInitIterator(ts, RPMTAG_NAME, rpmteN(psm->te), 0);
+	    /* this must match rpmNameVersionCompare in depends.c */
  	    xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_EPOCH, RPMMIRE_STRCMP,
  			rpmteE(psm->te));
  	    xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_VERSION, RPMMIRE_STRCMP,
  			rpmteV(psm->te));
  	    xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_RELEASE, RPMMIRE_STRCMP,
  			rpmteR(psm->te));
-	    if (tscolor) {
-		xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_ARCH, RPMMIRE_STRCMP,
+	    xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_ARCH, RPMMIRE_STRCMP,
  			rpmteA(psm->te));
+	    if (tscolor) {
  		xx = rpmdbSetIteratorRE(psm->mi, RPMTAG_OS, RPMMIRE_STRCMP,
  			rpmteO(psm->te));
  	    }