[Splashy-devel] Bug#486400: a hack that WORKS !!!

[Luis Mondesi]

On Jul 3, 2008, at 0:38, Eric Doviak <eric at doviak.net> wrote:

> Hi Luis,
>
> After days of struggling with the "resume bug," I have finally found  
> a hack that fixes the problem. It's not ideal, but it squashes the  
> bug and it provides laptop users with a boot splash without  
> compromising the quality that desktop users currently enjoy.
>
> Specifically, I started by adding the following code after line 44  
> of the "/usr/share/initramfs-tools/scripts/init-top/splashy" file:
>
> if [ ! -z "${resume}" ]; then
>         SPLASH=false
> fi
>
> That condition allows me to resume from hibernation, but it causes  
> Splashy to start rather late in the boot sequence. Specifically, it  
> starts when the "init-bottom" scripts are run. Adding such a  
> condition to Splashy would be useful to laptop users, but the late  
> start would degrade Splashy in the eyes of desktop users.
>
> I see no reason why desktop users should be penalized just to make  
> laptop users happy, so I created another condition that only runs  
> the condition above if "laptop" is passed as a kernel argument. The  
> code is below.
>
> By creating a "laptop" kernel argument, desktop users would continue  
> to see a splash screen early in the boot sequence and laptop users  
> would get a user-space boot splash system that works during start- 
> up, shutdown, suspend and resume.
>
> It certainly isn't the ideal solution that I had in mind a few days  
> ago (when I thought we could provide everyone with a splash screen  
> that starts early in the boot sequence), but it squashes the bug and  
> it makes Splashy usable for laptop users. Importantly, it achieves  
> those objectives without degrading Splashy's performance on desktop  
> computers.
>
> Let me know if there's anything else I can do,
> - Eric
>
>
> SPLASH=false
> LAPTOP=false
> SINGLE=false
> FBMODESET=false
>
> for x in $(cat /proc/cmdline); do
>     case $x in
>         single)
>             SINGLE=true
>         ;;
>         splash)
>             SPLASH=true
>         ;;
>         laptop)
>             LAPTOP=true
>         ;;
>         nosplash)
>             SPLASH=false
>         ;;
>         vga=*|video=*)
>             FBMODESET=true
>         ;;
>     esac
> done
>
> test $LAPTOP != "true" || if [ ! -z "${resume}" ]; then  
> SPLASH=false ; fi
> test $SINGLE = "false" || exit
> test $SPLASH = "true" || exit
> test $FBMODESET = "true" || exit

When $resume is not set, does Splashy start from initramfs init-top?  
Why does Splashy starts "late"?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.alioth.debian.org/pipermail/splashy-devel/attachments/20080703/2711cff9/attachment.htm