Description of problem: I have a bunch of init-scripts which detect when key pieces of information are not entered. ie: hostname, sendmail relay passwords, etc. All things that I wouldn't want in a rpm but need to be done before network or other services are active. In the old boot screens if there was a [FAILED] startup it would drop to a text screen where I have the user enter the missing information. Apache when it has a SSL key with a password will also do the same thing. If I just leave it the screen functions as normal the white bar goes across the screen and it just sits there. Doing nothing. When I high ESC nothing happens. The only way I've been able to get around this is to hit ESC immediately when the screen comes up. Then I go to the text startup where I can enter the info. Version-Release number of selected component (if applicable): Fedora 9.93 plymouth.i386 - 0.6.0-0.2008.10.30.4.fc10 How reproducible: Steps to Reproduce: 1. Create a script in /etc/rc5.d/S05test ====================== #!/bin/bash . /etc/rc.d/init.d/functions echo -n "Test" failure echo echo "Need some input" read line echo "you entered $line" ====================== 2. chmod +x /etc/rc5.d/S05test 3. reboot Actual results: Sits at the plymouth screen and ESC doesn't work. Expected results: To drop to the text screen so the input can be entered. Additional info: Normally I'd just change my scripts to work within plymouth but since I know apache could have a similar issue, I thought I'd log a bug. Just in case there are other programs that could cause the same problem.
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle. Changing version to '10'. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping
I'm not really sure of a way around this beside scripts either calling plymouth --hide-splash or plymouth ask-question --text "Prompt here: " I don't think there's a way to detect when the console is waiting at a prompt (especially since plymouth is always listening for key events anyway). One idea would be to switch plymouth from getting keypresses from the tty and instead getting them from /dev/input (evdev), then polling for readable on the tty. That'd be a lot of work and I'm not event sure if it would work. I thinkt he most straightforward move is to have the few scripts that need input to ask plymouth to give it like we do for encrypted disks. Other ideas welcome, though.