* Denis Barbier [2006-01-15 02:06:24+0100] > On Sat, Jan 14, 2006 at 10:16:47AM +0200, Recai Oktaş wrote: [...] > > I've prepared two separate patches targeting all these issues for your > > convenience. In the first (recommended) patch, which uses a more general > > approach, I've changed the code to handle XkbVariant through an XKBVARIANT > > variable. (To eliminate the code duplication, I modified the 'br' setup a > > little bit, though it could be improved further). The 2nd not-so-invasive > > patch treats Turkish as an exceptional case. > > You are right, XkbVariant has to be added to this code, so your first > patch is definitely better. But an upload is planned very soon and I > do not have the time to check that it does not break anything, so I > applied your second one for now and will apply your first one in few > days. Thanks for your work.
Hi Denis, Thank you very much for your sensitive response. No problem, we can wait. > I tested your code, and have one remark: you change tr_f -> f when > debian-installer/keymap is known, but you may also perform this > change after user selected a variant, he may get used to tr_f and > not know about this name change. You're right, the second patch is problematic. But as far as I can see it, the first one has not a flaw in this respect, since the TR handling code is in the: if [ -n "$RECONFIGURE" ] || [ -n "$FIRSTINST" ]; then body, before the db_inputs. Have I understood it correctly? Regards, -- roktas
