Hi, I made a patch against the baselayout package adding support for a progressbar, text and animations (only basic support yet) to the silent bootsplash during startup/shutdown. I tried to modify the files as little as possible. Finally affected files are: baselayout: - /sbin/rc - /sbin/functions.sh - /etc/init.d/halt.sh bootsplash: /etc/conf.d/bootsplash.conf How it works: /sbin/functions.sh got 3 new functions - splash_init(), splash_calc() and splash_update(). Additionally the function eend() was extended by a call to splash_update() for those not listed in /etc/runlevels/*/* (e.g. the inline scripts in /sbin/rc). During startup/shutdown splash_init() is called at the beginning of /sbin/rc (before the boot part) which initialises some variables and a counter (this is necessary to show the progress of the inline scripts in /sbin/rc, otherwise the delay between display of the bootsplash and the first step of the progressbar would be too big). After the boot part splash_calc() is called which calculates the number of scripts that have to be executed (without dependecies). Finally the last two extensions to /sbin/rc call splash_update(). The first resides in the for-loop stopping services, the second in the for-loop starting services. splash_update() expects to arguments ("script-name", "start|stop") and passes them to the wrapper script /sbin/splash (the arguments are used to start|stop animations according to the script-name and the appropriate entry in animations.cfg). /sbin/splash first execs the real splash executable (/sbin/splash.bin) paints a message above the progressbar and finally updates the progressbar. The source, additional information and a few screenshots are available at http://www.aichler.net/gentoo/bootsplash/ So, what do you think about it ? micha
While I was working on the animations support I discovered that /sbin/rc was not the best place for updating the progress bar, so I moved the two calls to splash_update() to /sbin/runscript.sh and adapted the functions in /sbin/functions.sh. Some other things have changed, too. Please drop version 0.2 and have a look at version 0.3. (It's still available at http://www.aichler.net/gentoo/bootsplash) Thanks, micha
Stupidly I developed the progressbar with baselayout version 1.8.5.9 and didn't notice that the implementation did not work with version 1.8.6.5 and above as expected because of major changes in the init scripts. I fixed that issue in version 0.4 micha
Bumping to azarah because he is the baselayout guru and I don't want to break anything in there. This may take some time before it makes it into Gentoo, please be patient.
This looks very interresting :)
This works very fine here (except the animated logos)! Go for it!
Hi guys - been a bit busy last month or two (not Gentoo related). I will try to add this in one of the next baselayout releases, as I did promise it to a few other people as well. What is the chance of getting the animated logo's to work in the mean time? Also, anybody know of an effort to get it ported to 2.6? One of the main reasons I have not gotten to this sooner, as this box err ... _cannot_ run a 2.4 kernel currently.
*** Bug 21705 has been marked as a duplicate of this bug. ***
Hi, > What is the chance of getting the animated logo's to work in the mean time? Animations running during default runlevel work well, but fbmngplay seems to be killed in sysinit/boot runlevel which leads to a frozen animation. I haven't had time to figure out what is responsible for it, because I'm currently working at an extension to fbtruetype, so that it's possible to show several lines of text one after another by erasing the framebuffer with a copy of the original background (was a feature request and would allow to show information about the currently executed script ). And - I wondered if it's the right path to play animations script dependend (or at least one by one), because of the parallel execution of start scripts. This might shorten the playtime of an animation to a minimum and IMHO parallel execution will become standard on the long run, won't it ? > Also, anybody know of an effort to get it ported to 2.6? One of the main > reasons I have not gotten to this sooner, as this box err ... _cannot_ run > a 2.4 kernel currently. Don't know anything about it yet.
Right, shout if you need info. On the parallel front - I did add initial for people asking, but it seems to be an overkill for bash. Thus either change things mightely again, or drop it.
ive been workin on moving the patch to 2.6 if anyone cares ;) i'll post it here soonish
well, i was making the patch then i started reading some docs at bootsplash.org which made me stop ;) 4) Will the bootsplash patch be included officially in future kernel versions? We would like to see that happen. But before that can happen, some things have to be cleaned up. The framebuffer interface of kernel 2.5/2.6 is completely different from 2.4, which means that large parts of the bootsplash have to be rewritten. Implementing other bit depths than 16bit, and other picture loaders than the jpg one are things that come to the mind. Also the userspace utility needs nicer config files. There is currently no code that works with 2.5. in other words i give up ;) if they cant handle it quickly i doubt i can
I have sorda 'ported' everything here except for the fbcon-splash16.c, which is the 16bit fbcon driver. That is much more work as it will need a fair large rewrite. Thus if anybody can get fbcon-splash16.c to work (check the sceleton fbcon driver first, and compare), drop me a mail, I can send what I have ...
Michael, I added the rc-scripts bits to baselayout-1.8.6.11 with some changes to suite form, as well as added compat with non bootsplash systems. Can you have a look. On another note - do you think you will have time to modify our bootsplash ebuild to install your scripts, with the animations ? Thanks.
Ok, I see I messed up things a bit - only have 2.4 at work though, so it might be a while to get fixed.
I'm currently quite busy at work, so sorry for the delay ! > Michael, I added the rc-scripts bits to baselayout-1.8.6.11 [...] Great ;) > Can you have a look. I had a look at it yesterday night. At first the progressbar didn't come up and it took me a while to get it working. Here are some issues: 1) the test for existing /proc/splash in splash_init () fails, cause /proc is mounted later on. 2) the boot message does not come up because of the same test in the rc_splash () function. 3) in splash_update() were two escape errors in the dependancy check I'll attach a patch that hopefully solves these parts ;) > On another note - do you think you will have time to modify our bootsplash > ebuild to install your scripts, with the animations ? > > Thanks. I'd like to, of course, but I run at full capacity and don't know when I have time for it. If there's no deadline, I can modify it, though. > Ok, I see I messed up things a bit - only have 2.4 at work though, so it > might > be a while to get fixed. Does the patch help ?
Created attachment 19330 [details, diff] Here's the patch...
I have some other fixes as well - btw, cant we change: -- - if [ -z "$(ls -1 \"${1}/softscripts/\" 2> /dev/null | grep \"${2}\")" ] + if [ -z "$(ls -1 "${1}/softscripts/" 2>/dev/null | grep "${2}")" ] -- to: -- - if [ -z "$(ls -1 \"${1}/softscripts/\" 2> /dev/null | grep \"${2}\")" ] + if [ ! -L "${1}/softscripts/${2}" ] -- Seems the same thing except it do not need the ls/grep magic.
Created attachment 19335 [details, diff] bootsplash.patch Ok, here is a patch with your fixes, as well as some of mine - I once again will only be able to test it tomorrow though. Also, not using SOFTLEVEL and BOOTLEVEL will cause the profile stuff to break, so fixed that, and just use svcdir, as it is defined at this stage. Btw: the ebuild is not that crusial yet.
> Seems the same thing except it do not need the ls/grep magic. Definitely ! I wonder what made me do that ;)
> Ok, here is a patch with your fixes, as well as some of mine - I once again > will only be able to test it tomorrow though. Just applied the patch and tested it. Works all but the boot message.
Created attachment 19344 [details, diff] bootsplash2.patch This should fix the remainging issues, except for the killall5 issue, which I am wondering what to do about. It should go on top of the bootsplash.patch.
Created attachment 19366 [details] /etc/init.d/halt.sh This halt.sh fixes it for me - would appreciate some testing/comments.
> This halt.sh fixes it for me - would appreciate some testing/comments. Just tested it, works fine except the killall issue on system shutdown/reboot. You've introduced a new argument "continue" to /sbin/splash, do you have modified /sbin/splash as well ? Besides that - this part in /sbin/function.sh makes no sense to me, cause the first thing init does on shutdown/reboot is to send these killall signals. I had disabled these lines in the patch and haven't had any problems, so, what is it for ?
*** Bug 31633 has been marked as a duplicate of this bug. ***
Fixed the killall issue btw.
Great news for 2.6 users - a port is out: http://marc.theaimsgroup.com/?l=linux-kernel&m=106681497425277&w=2
Err, on crack there - its the progress patch .. ill have a look if i get time.
Will a fix for this issue be in portage any time soon? It's bretty annoying...
Its actually Tad's domain - I'll ping him and see.
Since the Usage error message seems to be marked as a duplicate of this one. edit /sbin/functions.sh in splash)update (){ change /sbin/splash "${myscript}" "${action}" to read /sbin/splash "${myscript} ${action}" mentioned by relyt in http://forums.gentoo.org/viewtopic.php?t=103591 Also Stephen Ressenar has updated the bootsplash.org patch for kernel 2.6.0-test9 (works with test9, test9-mm1, test9-mm2, test9-mm3 in all tests) ftp://ftp.suse.com/pub/people/stepan/bootsplash/kernel/bootsplash-3.1.3-2.6.0-test9.diff (This is the original for the version which has been updated on bootsplash sourceforge page, bootsplash.org will not be updated til Monday)
Created attachment 21368 [details, diff] patch for bootsplash tarball patch bootsplash tarball to include splash script (gentoo specific), and fix missing -n in initscript.
Created attachment 21369 [details, diff] patch to ebuild to make the patched tarball work This patch makes the previous splash get installed as splash.bin, and installs the gentoo specific splash script. These two patches together make the progressbar work, and get rid of the "usage errors" that people have been complaining about.
Any chance of updating this for baselayout 1.8.6.12? Using it here and it's working, but not sure what it's replaced or how portable it is. I really like this, and think it should become part of the system itself.
Tad, can you give your seal of approval on this?
whoa.... this looks really good... progressbar is awsome and no more usage spam... very nice work.
Brad, I have not looked at this (splash ebuild stuff), and it seems Tad is sorda missing in action. If you want, update the bootsplash ebuild, mark as unstable and depend on >=baselayout-1.8.6.12.
Patch applied and works fine here.
bootsplash-0.6-r3 comitted to CVS. Tyler, thanks for the patches. Sorry for the delay on this guys.
