I'll admit up front that I don't know the answer to this and that the current behavior is likely acceptable for v1. I suspect in v2 the querying capabilities will render this a non-issue.

That said, here's what I'm seeing.

I synchronized the i386 and x86_64 Pulp fedora-16 repositories. It's an interesting content set since by NEVRA alone we'll have multiples of the same files. They are obviously different, however, because of the checksums.

The output from the `package search` command doesn't display the checksum. It probably shouldn't as it would just be noise. However, without it, the output looks awkward. Here's a snippet:


[jdob@pulp1 ~]$ pulp-admin package search -n gofer
+------------------------------------------+
             Package Information
+------------------------------------------+
Name         	EVRA                      	Filename                            	Repositories    
gofer        	0:0.65-1.fc16.noarch      	gofer-0.65-1.fc16.noarch.rpm        	pulp-f16-32
gofer        	0:0.65-1.fc16.noarch      	gofer-0.65-1.fc16.noarch.rpm        	pulp-f16


According to that output, it's misleading as it looks like Pulp has a bug that's causing it to show the package twice.

One solution may be to list the beginning of the checksum in this output. On disk, the two versions are separated by a checksum directory[1]. We could display that same bit from the checksum to clue users into the fact that they are in fact two different packages.


[1]

[root@pulp1 noarch]# pwd
/var/lib/pulp/packages/gofer/0.65/1.fc16/noarch
[root@pulp1 noarch]# ll -R
.:
total 8
drwxr-xr-x. 2 apache apache 4096 Feb 13 12:54 31d
drwxr-xr-x. 2 apache apache 4096 Feb 13 12:58 81a

./31d:
total 64
-rw-r--r--. 1 apache apache 64313 Feb 13 12:54 gofer-0.65-1.fc16.noarch.rpm

./81a:
total 64
-rw-r--r--. 1 apache apache 64381 Feb 13 12:58 gofer-0.65-1.fc16.noarch.rpm