[Initng] Service xxx already in db

[Peter Zotov]

Hi again!

I have a small question about the message:

  18:35:40 -- WARN:       Service <service> is already in db, please 
look there before parsing.
To flush your cache do ngc -R

I have my own distribution, targetted mainly for embedded devices, so, I 
am writing my own initscripts at least for system initialization. I 
receive these messages for four services: system/initial, 
system/initial/{virtfs,cleartmp,preparelogin} even when starting a fresh 
system in which no cache can ever be. Moreover, I receive same messages 
twice: after the "System is starting up" message and also after some 
services are starting. I looked into source, but have no idea about them.

You can check my disk image at 
http://files.whitequark.ru/t/initng_test_image.gz (3M), and initscripts 
(/etc/initng tree, really) at http://files.whitequark.ru/t/initng_tree.tbz2.

Also, when I mount tmpfs over /dev, I destroy initng socket. Currently, 
I recreate it by signalling initng, but it warns... is there some nice 
way to do the same?

Also, when I type `ngreboot' at console, initng does not SIGTERM 
everything, and I need to exit shell manually. Again, is there a nice 
way to fix this?

And last. NGC outputs many ANSI sequences and replies weirdly on some 
commands (I think that `Command POSITIVE' isn't common reply), and also 
doesn't set exit code. This makes it unsuitable for scripts. Do I need 
to write some patch for it or is it already done?

-- 
   WBR, Peter Zotov