This seems to have been triggered by the fix for https://bugzilla.redhat.com/show_bug.cgi?id=1039185 , see https://bugzilla.redhat.com/show_bug.cgi?id=1039185#c46 . Since anaconda-34.3-1.fc34 landed in Rawhide, containing this commit: https://github.com/rhinstaller/anaconda/commit/e05cc18294e67099bf87076e4f23fe5f031fecb5 if you do an install in Russian, the console layout configured in `/etc/vconsole.conf` is 'us', not 'ru' as it should be. For these languages, the native console layout is "internally switched": it has two levels for inputting Latin characters and two levels for inputting Cyrillic characters. This is necessary as there's no convenient mechanism in kbd for switching between two separate console layouts. So we have to use the native layout, not 'us', or else there is no way to input native characters at the console. I'm not sure yet *why* this change caused this consequence. The code paths are quite complex and ultimately wind up in the libxklavier g-i bindings. The change is clearly only actually intending to touch the X config, but in fact it looks like it winds up affecting the console layout too. My best guess would be that we wind up asking localed to set the console layout after making this change, and if localed sees the X layouts are "us,ru", it will set the console layout to "us", not "ru". Maybe we'll ultimately need to change localed? Proposing as a Final blocker as a violation of "If a particular keyboard layout has been configured for the system, that keyboard layout must be used: When unlocking encrypted storage volumes during boot (but see footnotes) [and] When logging in at a console" - https://fedoraproject.org/wiki/Fedora_34_Final_Release_Criteria#Keyboard_layout_configuration . To me, this situation constitutes "configuring" the ru layout; we definitely intend, when you install in Russian, that the console layout configured should be ru.
Ah, the system log does show this: WARNING org.fedoraproject.Anaconda.Modules.Localization:DEBUG:anaconda.modules.localization.localed:Missing virtual console keymap value us converted from ['us', 'ru'] X layouts which is logged from: if not vc_keymap: vc_keymap = localed_wrapper.convert_layouts(x_layouts) log.debug("Missing virtual console keymap value %s converted from %s X layouts", vc_keymap, x_layouts) in pyanaconda.modules.localization.localed.get_missing_keyboard_configuration() .
Probably, PR: https://github.com/rhinstaller/anaconda/pull/3064 changes this to: [user@localhost user]# localectl System Locale: LANG=ru_RU.UTF-8 VC Keymap: ru X11 Layout: us,ru X11 Variant: , X11 Options: grp:alt_shift_toggle [user@localhost user]# cat /etc/vconsole.conf KEYMAP="ru" FONT="eurlatgr" shall that work?
Yes, it should. Let's mark this as POST. When a new anaconda is built in Rawhide we can see if it's fixed. Thanks.
Thank you!
Unfortunately test still fails in the same way with today's Rawhide, though I thought it should have the fix: https://openqa.fedoraproject.org/tests/754876# I'll dig into it a bit and try to figure out what's going on.
ohh, PR is still open. I thought it had been merged for some reason. So, not a surprise the bug is still there.
PR merged
FEDORA-2021-05c241c540 has been pushed to the Fedora 34 stable repository. If problem still persists, please make note of it in this bug report.
FEDORA-2021-eaae734297 has been pushed to the Fedora ELN stable repository. If problem still persists, please make note of it in this bug report.
bad news: this still isn't working right. good news: it's not because of this bug any more! it's because of https://bugzilla.redhat.com/show_bug.cgi?id=1919483 , a change in kbd. Someone might suggest "fixing" that problem in anaconda, though, so you might wanna read it. More bad news: while testing this I noticed that *displaying* Cyrillic characters on the console is also broken, and in fact was broken at least as far back as Fedora 33 Final :( Filed that as https://bugzilla.redhat.com/show_bug.cgi?id=1919486 .
More bad news: this fix still wasn't *really* right. It works for Russian, but it doesn't work for Bulgarian (or some other cases, probably). See https://bugzilla.redhat.com/show_bug.cgi?id=2239213 .