On Mon, Jan 20, 2014 at 05:41:04PM -0500, Michael K. Johnson wrote: > On Mon, Jan 20, 2014 at 11:36:13AM -0500, Michael K. Johnson wrote: > > https://opensource.sas.com/its/browse/CNY-3834 filed for this issue. > > > > The good news is that it's unlikely to affect new repository imports, > > so it shouldn't put any work on F20 import on hold. The foresight > > repository is one of the larger repositories out there. > > I built conary with the expected workaround for this issue applied and > put it on the repository server. Let me know how it goes...
The workaround (which worked but involved reverting a Conary patch so was not a good long-term solution) has been replaced with what is expected to be the good long-term fix for this issue. Since the root of the problem was that postgres was planning the query poorly, we cannot be 100% certain that the problem is resolved without confirming in situ. So I would appreciate it if the next person to build something into the foresight repository would watch to see whether the problem is still gone, and post a response to this email with the results either way. Thanks _______________________________________________ Foresight-devel mailing list [email protected] https://lists.foresightlinux.org/mailman/listinfo/foresight-devel
