On Tue, Dec 18, 2012 at 03:31:51AM +0200, Zeeshan Ali (Khattak) wrote: > On Mon, Dec 17, 2012 at 11:07 PM, Christophe Fergeau > <cfergeau@xxxxxxxxxx> wrote: > > Now that OsinfoInstallConfig has access to the > > OsinfoInstallConfigParamList for the OsinfoInstallScript that is > > being configured, > > In the previous patch (06/11) you just added the setter/getter for > config params to OsinfoInstallConfig but nothing yet sets the params. > Either I'm totally confused or the ordering of these patches is > somehow wrong. Let's blame the commit log here, what the series is doing is 1) add the OsinfoInstallConfig:config-parameters property 2) use it in OsinfoInstallConfig _if it is set_ 3) use OS specific values in OsinfoInstallScript 4) set OsinfoInstallConfig:config-parameters where needed This commit is doing 2) even though the commit log can be misleading. I can see 3) being misplaced, it could go last, but apart from this I think the ordering is not that bad. > > we can use the OsinfoDatamap that is optionally > > set on a given parameter to automatically translate a value for > > this parameter from a generic libosinfo value to an OS-specific one. > > Assuming config params in OsinfoInstallScript has (or can has) access > to datamaps, I wonder if there is any need to involve > OsinfoInstallConfig at all here. The changes will be less intrusive > that way AFAICT. I've seen your followup to this email saying to disregard this comment. For what it's worth, I've already detailed in https://www.redhat.com/archives/virt-tools-list/2012-December/msg00288.html why I think adding config params to OsinfoInstallConfig is a good move regardless of the datamaps stuff. Christophe
Attachment:
pgpE2JmAwaAu9.pgp
Description: PGP signature
_______________________________________________ Libosinfo mailing list Libosinfo@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libosinfo
