Bug 892054

Description of problem:
I am glad a more recent texlive than 2007 is finally making it into fedora 18. I notice some upgrade anomalies whenever upstream versions have "a" or "b" in the components though, where yum upgrade would offer to "downgrade", basically.
Here are a few examples of actual screen scraps:

-----------------
Resolving Dependencies
--> Running transaction check
---> Package texlive-biblatex-fiwi.noarch 1:1.2a.svn28435-10.fc18 will be updated
---> Package texlive-biblatex-fiwi.noarch 1:1.2.svn27096-8.fc18 will be an update
---> Package texlive-biblatex-fiwi-doc.noarch 1:1.2a.svn28435-10.fc18 will be updated
---> Package texlive-biblatex-fiwi-doc.noarch 1:1.2.svn27096-8.fc18 will be an update
---> Package texlive-currfile.noarch 1:0.7a.svn28118-10.fc18 will be updated
---> Package texlive-currfile.noarch 1:0.7.svn26425-8.fc18 will be an update
---> Package texlive-currfile-doc.noarch 1:0.7a.svn28118-10.fc18 will be updated
---> Package texlive-currfile-doc.noarch 1:0.7.svn26425-8.fc18 will be an update
---> Package texlive-libertine.noarch 1:0.svn28432-10.fc18 will be updated
--> Processing Dependency: tex(biolinum-type1.sty) for package: 1:texlive-libertine-0.02.svn27816-8.fc18.noarch
--> Processing Dependency: tex(libertine-type1.sty) for package: 1:texlive-libertine-0.02.svn27816-8.fc18.noarch
---> Package texlive-libertine.noarch 1:0.02.svn27816-8.fc18 will be an update
---> Package texlive-libertine-doc.noarch 1:0.svn28432-10.fc18 will be updated
---> Package texlive-libertine-doc.noarch 1:0.02.svn27816-8.fc18 will be an update
--> Running transaction check
---> Package texlive-biolinum-type1.noarch 1:5.3.0.svn27786-8.fc18 will be installed
---> Package texlive-libertine-type1.noarch 1:5.3.0.svn27835-8.fc18 will be installed
--------------
	package texlive-biblatex-chicago-1:0.9.9.svn27254-10.fc18.noarch (which is newer than texlive-biblatex-chicago-1:0.9.9b.svn28472-
11.fc18.noarch) is already installed
	package texlive-hf-tikz-1:0.1.svn27457-10.fc18.noarch (which is newer than texlive-hf-tikz-1:0.1a.svn28573-11.fc18.noarch) is alr
eady installed
	package texlive-hf-tikz-doc-1:0.1.svn27457-10.fc18.noarch (which is newer than texlive-hf-tikz-doc-1:0.1a.svn28573-11.fc18.noarch
) is already installed
	package texlive-biblatex-chicago-doc-1:0.9.9.svn27254-10.fc18.noarch (which is newer than texlive-biblatex-chicago-doc-1:0.9.9b.s
vn28472-11.fc18.noarch) is already installed
-----------

You can see quite a few components with higher svn revisions are considered old because the newer versions have an extra "a", "b", or in one case, fewer digits in the upstream versions.

Version-Release number of selected component (if applicable):
see above.

How reproducible:
always.

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

I don't see an obvious answer to this problem - bumping epoch up for every svn snaphshot would be undesirable. Perhaps putting the snapshot date as part of the version number, like some of the other packages from snapshots?