Re: [leaf-devel] /linuxrc testing - round 2
Charles At 17:26 05.05.2004 -0500, Charles Steinkuehler wrote: Erich Titl wrote: Charles I looked at the new linuxrc, nice, modular, easier to maintain. At 22:14 05.05.2004, Charles Steinkuehler wrote: - /dev/null in root.dev.mk (should be in root.blk.mk even though it's not a block device, as it's used early in the init scripts). Would this not, for consistency's sake, rather go to a root.chr.mk file. Personally, I don't think so... ...if you're worried about a char device showing up in root.blk.mk, I'd either: - Just add a comment in root.blk.mk stating something like /dev/null is here with all the block devices because it's used early in the boot sequence (and in this file!). - If that's too much bloat or doesn't make you happy, you chould rename root.blk.mk something like root.bootdev.mk, root.init.mk or something else that coveys these devices are needed to boot-strap and find the leaf.cfg file No offense intended, I just felt that, if you go to the trouble of naming root.blk.mk explicitly after the type of device it builds, it would be logical to name the one that builds the character device nodes accordingly, e.g. root.chr.mk. cheers Erich THINK Püntenstrasse 39 8143 Stallikon mailto:[EMAIL PROTECTED] PGP Fingerprint: BC9A 25BC 3954 3BC8 C024 8D8A B7D4 FF9D 05B8 0A16 --- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 ___ leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
Re: [leaf-devel] /linuxrc testing - round 2
Erich Titl wrote: snip No offense intended, I just felt that, if you go to the trouble of naming root.blk.mk explicitly after the type of device it builds, it would be logical to name the one that builds the character device nodes accordingly, e.g. root.chr.mk. None taken. The root.blk.mk name was just something I came up with trying to stick with the previous naming convention and keep the name short. At the time, I couldn't think of a better name. It just occured to me that perhaps the file should be named init.dev.mk. Anyone else like that better than root.blk.mk, or got a better suggestion? -- Charles Steinkuehler [EMAIL PROTECTED] --- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 ___ leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
Re: [leaf-devel] /linuxrc testing - round 2
Charles At 13:19 06.05.2004, Charles Steinkuehler wrote: Erich Titl wrote: snip No offense intended, I just felt that, if you go to the trouble of naming root.blk.mk explicitly after the type of device it builds, it would be logical to name the one that builds the character device nodes accordingly, e.g. root.chr.mk. None taken. The root.blk.mk name was just something I came up with trying to stick with the previous naming convention and keep the name short. At the time, I couldn't think of a better name. It just occured to me that perhaps the file should be named init.dev.mk. Anyone else like that better than root.blk.mk, or got a better suggestion? Not really, as you stated before names are... just names BTW. I got 2.4.24 running with modularised iptables/conntrack modules right now with the disabled keyboard patch, but this is just another compile round. I patched and compiled OpenSwan 1.0.3 against it (slink), no packaging yet. cheers Erich THINK Püntenstrasse 39 8143 Stallikon mailto:[EMAIL PROTECTED] PGP Fingerprint: BC9A 25BC 3954 3BC8 C024 8D8A B7D4 FF9D 05B8 0A16 --- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 ___ leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
Re: [leaf-devel] /linuxrc testing - round 2
Charles; I may have found another small problem: firewall# lrpkg -i squid-2.lrp Installing squid-2 ... cat: /var/lib/lrpkg//boot.fstype: No such file or directory Done. I was told: It could be a bug (from the lrpkg script): echo $fn=-t `cat $lrpkgpath/boot.fstype` $d$lrpkgpath/backdisk but boot.fstype is obsoleted with new linuxrc. Any idea how to solve? kp --- This SF.Net email is sponsored by Sleepycat Software Learn developer strategies Cisco, Motorola, Ericsson Lucent use to deliver higher performing products faster, at low TCO. http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3 ___ leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
Re: [leaf-devel] /linuxrc testing - round 2
K.-P. Kirchdörfer wrote:
Charles;
I may have found another small problem:
firewall# lrpkg -i squid-2.lrp
Installing squid-2 ... cat: /var/lib/lrpkg//boot.fstype: No such file or
directory
Done.
I was told:
It could be a bug (from the lrpkg script):
echo $fn=-t `cat $lrpkgpath/boot.fstype` $d$lrpkgpath/backdisk
but boot.fstype is obsoleted with new linuxrc.
Any idea how to solve?
This is a known bug. The error message is effectively harmless, and
only appears when manually installing packages.
The result of this bug is you must manually specify the backup disk when
first backing up a package you manually installed (which you'd probably
have to do anyway), rather than having it default to the (no longer
existing) /dev/boot device.
Assuming using the first entry in pkgpath.disks in place of the missing
boot device/fs info is acceptable, the following will fix the problem:
echo $fn=-t `cat $lrpkgpath/boot.fstype` $d$lrpkgpath/backdisk
Replace with (all one line):
echo $fn=-t `sed -n '\:'$d':{s/^.* //;p;}' $lrpkgpath/pkgpath.disks`
$lrpkgpath/backdisk
A bit above this, the handling of $d needs to be tweaked as well:
Was:
if [ -z $2 ]; then
local d=`sed 's/.*boot=/\1/; s/[: ].*//' /proc/cmdline`
else
Replace with:
if [ -z $2 ]; then
local d=`sed -n '1{s/ .*$//;p;}' $lrpkgpath/pkgpath.disks`
else
The (unused) mount.boot procedure should be removed.
...and finally, the mount.back procedure should be tweaked:
Was:
if [ $dev = ]; then
mount -t `cat /var/lib/lrpkg/boot.fstype` /dev/boot $2
else
Replace with (between if/else is one line):
if [ $dev = ]; then
mount -t `sed -n '1{s/\([^ ].*\) \(.*\)$/\2 \1/;p;}'
$lrpkgpath/pkgpath.disks` $2
else
Of course, it wouldn't hurt to s:/var/lib/lrpkg:$lrpkgpath: on the whole
file, to remove any remaining absolute references to the package directory.
NOTE: Mods above have been briefly tested for correct shell quoting and
functionality, but no guarantee is expressed or implied!
--
Charles Steinkuehler
[EMAIL PROTECTED]
---
This SF.Net email is sponsored by Sleepycat Software
Learn developer strategies Cisco, Motorola, Ericsson Lucent use to
deliver higher performing products faster, at low TCO.
http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3
___
leaf-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-devel
