I think packages can have a postinstall script, and so it should be possible to 
adjust on a per package basis to have things auto start and set themselfs to 
enabled. It might makes sense for some packages, but some people probably 
install "everything" and so you do want to do that for every possible service 
just because it's installed... 

On a partially related question about packages... 
The default config for squid doesn't work. You have to run a command to init 
it's directory structure when using disk storage (likely every reboot if it's 
ram backed filesystem), and using a disk based cache doesn't make sense (at 
least on my device a WNDR3800 with 128mb of RAM and 16mb flash). I have more 
appropriate config that tells it to use 24MB RAM (process stabilizes at taking 
about 40MB after overhead) instead of a virtual disk that uses RAM less 
efficiently that then causes a reboot when it fills up... Should a usable 
config be the default, or maybe this is device specific and the default is good 
by default for some other platforms? I have a package that requires squid and 
runs it with a different config file instead of the one the comes with the 
squid package, so that works for me. Not sure if it would be useful to fix up 
and submit a change to the deafult squid package, or maybe submit a 
squid-ramcache package for easy install. If I did submit a package, I would 
need to clean up my package as it also adjusts the firewall rules (to force 
transparent caching) in an ugly way that would have to somehow be done more 
friendly/compatibly with the default rules... 


----- Original Message -----

From: "Jan Lukeš" <johny...@gmail.com> 
To: "OpenWrt Development List" <openwrt-devel@lists.openwrt.org> 
Sent: Saturday, November 10, 2012 2:55:51 PM 
Subject: Re: [OpenWrt-Devel] Autostart of services.. 

It's a matter of coherence among other services. Some services might be 
harmless, but some can pose a security risk. We need to treat them all the 
same, thus as risk. 


Anyway. Persistence is just matter of explicitly enabling it. 


On Sat, Nov 10, 2012 at 6:16 PM, Michael Markstaller < m...@elabnet.de > wrote: 


-----BEGIN PGP SIGNED MESSAGE----- 
Hash: SHA1 

Hi, 

first of all: I know this is "intentionally unwanted" currently; 
still I'd like to start discussing it again: 
If I i.e. install the snmpd-Package (just one example) I'd think it 
should be up&running, no need to manually enable it again so it really 
starts & runs :o 

This behavior (when i start it afterwards manually, it only persists 
until reboot) is IMHO confusing and not very user-friendly.. 
Whats the real point against something the user selected also runs 
without further fiddling to really enable it (once again)? 


best regards 

Michael 
-----BEGIN PGP SIGNATURE----- 
Version: GnuPG v1.4.11 (GNU/Linux) 
Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ 

iEYEARECAAYFAlCei98ACgkQaWRHV2kMuAI4KACgn3F5uyPnQjl+B01kTA+ey1r6 
mZEAn0PZ5WIVYHy2P/pjyx04tBo1DkFB 
=3MIZ 
-----END PGP SIGNATURE----- 
_______________________________________________ 
openwrt-devel mailing list 
openwrt-devel@lists.openwrt.org 
https://lists.openwrt.org/mailman/listinfo/openwrt-devel 






-- 



S pozdravem 


Jan Lukeš 

_______________________________________________ 
openwrt-devel mailing list 
openwrt-devel@lists.openwrt.org 
https://lists.openwrt.org/mailman/listinfo/openwrt-devel 

_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Reply via email to