First, slang doesn't handle switching from non-unicode to unicode in the same process. Rather than find the magic internal variables to frob, just set to a unicode locale before we ever init slang. Second, we need the locale archive on stage1 so that wcwidth and friends actually calculate character width correctly. Third, there is some bizarre interaction going on so that when KMS initializes, some of the console font data is no longer correct. I could start digging into the kernel code to figure out WTF is going on here... or I could just reset the font whenever we change language. Patches 1 & 3 tested in a local build of the loader. Patch 2 not tested directly (just slapped locale-archive in the initrd by hand), but is hopefully either obviously right or obviously wrong. loader/lang.c | 1 + loader/loader.c | 1 + scripts/mk-images | 1 + 3 files changed, 3 insertions(+) _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
