> > I was not able to split languages out into their own support packages, though. > > The reason for this is that if you boot off a boot.iso without "askmethod" or > > "repo=", we default to the Fedora mirror list. > > A specific hardwired location might be OK for retrieving [say] 1MB of language- > specific data per install. Yeah, we can maybe get away with this. We'd have to prompt for the network in loader, though, because loader's in its own translation domain and we always ship that. It's too bad because I'm really trying to kill loader. > > But we don't know what that is > > until way later in anaconda itself, and that's too late to fetch the language > > support. We may also need to bring up the network to fetch it, which means > > prompting in the user's language. > > In my book such localization is a "[very] nice to have", but not a requirement. > The reason is the cost: many megabytes. Anaconda should quantify this cost > numerically, so that those who insist on localization can pay knowingly > for that preference. It's on the order of 17 MB compressed for all translations. Of course, the uncompressed size (which will be how much memory they consume) is larger. My hope is to first get these patches in, and then work on this language stuff. It's just becoming too burdensome to rebase them every time someone touches something in loader/. - Chris _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
