More nits re 3.1.1b2

View: New views
5 Messages — Rating Filter:   Alert me  

More nits re 3.1.1b2

by groups, freeman :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi, thought that I'd enumerate anything odd or minuscule possible bugs
that occur to me as I work with 3.1.1b2.

- busybox, when reading the /etc/network/interfaces file (via 'ifup
-a'), can't handle a line that says, for example:
    <tab_char>address<tab_char><tab_char>192.168.0.1
however it can read and accept the line if the there's only a *single*
<tab_char> between the end of the token "address" and the provided
numeric ip addy. This contrasts with the behaviour of BB in the 3.0b2
release of BuC, which has no problems with multiple tab chars.

- file /etc/shorewall/zones has different (greater) permissions than any
other shorewall config file in that folder

- created processes are given predictable, incrementing PIDs (versus BuC
3.0b2 which assigns random PID's)

- in the 1680 floppy installation the file from local.lrp:
    /var/lib/lrpkg/local.local
contains no entries to specify inclusion of the "usr/local/bin" and
"usr/local/sbin" directories.

- Problem parsing /etc/modules
There's a new routine (compared to BuC3.0b2) for parsing the
/etc/modules file, that unfortunately breaks the reading of the modules
file if the user has employed any leading tabs or spaces before the name
of the module (indenting). The part of /etc/init.d/modutils that doesn't
work for me is this:
> grep ^[a-z0-9A-Z] /etc/modules |
> while read module args
> do
>      insmod $module $args 2>/dev/null
>      lsmod | grep -n -q ^$module || \
>      logger modutils module $module could not be loaded
> done
which I replaced with
> while read module args; do
>       # Only insmod lines without leading # marks      
>       if [ "${module:0:1}" != "#" ]; then
>             insmod $module $args 2>/dev/null
>             lsmod | grep -n -q ^$module || \
>             logger modutils module $module could not be loaded
>       fi
> done < /etc/modules

(Yeah, I use indenting a lot to make my files more easily read).

-- My workstation=Win XP SP3.
- Problem with formatting option using imaging-1680 exe:
When running the imaging-1680 exe pulled from:
 http://downloads.sourceforge.net/leaf/Bering-uClibc_3.1.1-beta2_img_bering-uclibc-1680.exe
on a blank, 1.44MB floppy I affirm (check) the options for "Writing on
floppy" and "Formatting". On 3 different floppies I got a "Disk error on
track 1, head 0; Address mark not found; the drive cannot find the
sector requested". This seems to be a bug since when I attempt to use an
older imaging-1680 exe (e.g. from
http://easynews.dl.sourceforge.net/sourceforge/leaf/Bering-uClibc_2.4.1_img_bering-uclibc-1680.exe 
) the format goes fine on one of the 'problematic' diskettes (untested
on the others). And when I thereafter use the current 3.1.1b2
imaging-1680 exe but affirm only "Writing on floppy" I am able to have a
1680-sized floppy with the current 3.1.1b2 files.

- Possible WinImage version/licencing problem:
When I run the 3.1.1b2 imaging-1680 exe it gives a pop-up window saying
that the version of WinImage used to create the imaging exe is not
licenced for redistribution of the resultant exe. This contrasts with
the 2.4.1 imaging-1680 exe which does not offer that admonition. This
might be a no-no for a project like LEAF :)

Thanks as always for LEAF, hopefully this is helpful to the project!
scott

groups, freeman wrote:

> Hi, just starting to look at 3.1.1b2 and seem to have noticed that the
> folder
>  \2.4.34.6\kernel\drivers\i2c
> is missing the the modules tarball, that I grabbed from:
>  http://downloads.sourceforge.net/leaf/Bering-uClibc_modules_2.4.34.6.tar.gz?use_mirror=internap
>
> I think this may cause bustage for the sensors (monitoring) functionality?
>
> Thanks as always to everyone for their work on LEAF!
>
> scott


------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today.
Use priority code J9JMT32. http://p.sf.net/sfu/p
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@...
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Re: More nits re 3.1.1b2

by Gordon Bos :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

I think you just need quotes around the first grep argument.

I'd replace the grep part with this though:

   cut -d'#' -f1 /etc/modules | \
   grep -v '^[[:space:]]*$' | \
   while read module args
   ....

Gordon

groups, freeman wrote:

> - Problem parsing /etc/modules
> There's a new routine (compared to BuC3.0b2) for parsing the
> /etc/modules file, that unfortunately breaks the reading of the modules
> file if the user has employed any leading tabs or spaces before the name
> of the module (indenting). The part of /etc/init.d/modutils that doesn't
> work for me is this:
>> grep ^[a-z0-9A-Z] /etc/modules |
>> while read module args
>> do
>>      insmod $module $args 2>/dev/null
>>      lsmod | grep -n -q ^$module || \
>>      logger modutils module $module could not be loaded
>> done
> which I replaced with
>> while read module args; do
>>       # Only insmod lines without leading # marks      
>>       if [ "${module:0:1}" != "#" ]; then
>>             insmod $module $args 2>/dev/null
>>             lsmod | grep -n -q ^$module || \
>>             logger modutils module $module could not be loaded
>>       fi
>> done < /etc/modules


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@...
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Re: More nits re 3.1.1b2

by KP Kirchdoerfer-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

HI;

thx for feedback;

We'll look into the other mentioned issues later, just feedback for two or
three points you raised:

Am Donnerstag, 23. April 2009 02:42:26 schrieb groups, freeman:
> Hi, thought that I'd enumerate anything odd or minuscule possible bugs
> that occur to me as I work with 3.1.1b2.
[...]
>
> - file /etc/shorewall/zones has different (greater) permissions than any
> other shorewall config file in that folder

Fixed in cvs

> - created processes are given predictable, incrementing PIDs (versus BuC
> 3.0b2 which assigns random PID's)

This is "by intention" - the grsecurity maintainer wrote by the time he added
support for the 2.4.34 kernel:
"Removal of randomized PIDs feature, since it provides no useful additional
security [...]"


> - in the 1680 floppy installation the file from local.lrp:
>     /var/lib/lrpkg/local.local
> contains no entries to specify inclusion of the "usr/local/bin" and
> "usr/local/sbin" directories.

This a feature request, not a bug report, correct?

[other stuff deleted]


kp

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@...
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Re: More nits re 3.1.1b2

by Eric Spakman :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi,

>> - in the 1680 floppy installation the file from local.lrp:
>>     /var/lib/lrpkg/local.local
>> contains no entries to specify inclusion of the "usr/local/bin" and
>> "usr/local/sbin" directories.
>>
>
> This a feature request, not a bug report, correct?
>

This is by intention. Local.lrp is meant to provide a way to add custom
config files/scripts to the config database. The config database is not
meant to contain binary files. The best way to have the option to use your
own binary files is create a simple lrp package which contains those
files.


Eric


------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@...
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Re: More nits re 3.1.1b2 (part two)

by KP Kirchdoerfer-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi;

as I said some more feedback:

Am Donnerstag, 23. April 2009 02:42:26 schrieb groups, freeman:

> Hi, thought that I'd enumerate anything odd or minuscule possible bugs
> that occur to me as I work with 3.1.1b2.
>
> - busybox, when reading the /etc/network/interfaces file (via 'ifup
> -a'), can't handle a line that says, for example:
>     <tab_char>address<tab_char><tab_char>192.168.0.1
> however it can read and accept the line if the there's only a *single*
> <tab_char> between the end of the token "address" and the provided
> numeric ip addy. This contrasts with the behaviour of BB in the 3.0b2
> release of BuC, which has no problems with multiple tab chars.


This seems to be a pb with upstream busybox version.
It may help to report it at the busybox bugtracker.
Otherwise just avoid to have more than one tab - I'm sorry.


> - Problem parsing /etc/modules
> There's a new routine (compared to BuC3.0b2) for parsing the
> /etc/modules file, that unfortunately breaks the reading of the modules
> file if the user has employed any leading tabs or spaces before the name
> of the module (indenting). The part of /etc/init.d/modutils that doesn't
>
> work for me is this:
> > grep ^[a-z0-9A-Z] /etc/modules |
> > while read module args
> > do
> >      insmod $module $args 2>/dev/null
> >      lsmod | grep -n -q ^$module || \
> >      logger modutils module $module could not be loaded
> > done
>
> which I replaced with
>
> > while read module args; do
> >       # Only insmod lines without leading # marks
> >       if [ "${module:0:1}" != "#" ]; then
> >             insmod $module $args 2>/dev/null
> >             lsmod | grep -n -q ^$module || \
> >             logger modutils module $module could not be loaded
> >       fi
> > done < /etc/modules
>
> (Yeah, I use indenting a lot to make my files more easily read).

Hopefully fixed in the soon available beta3 version.


> -- My workstation=Win XP SP3.
> - Problem with formatting option using imaging-1680 exe:
> When running the imaging-1680 exe pulled from:

The license issues will be solved with next version, thx for reporting. Other
than that the images should be ok - pls doublecheck once the beta3 is
available.


kp

kp

------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations
Conference from O'Reilly Media. Velocity features a full day of
expert-led, hands-on workshops and two days of sessions from industry
leaders in dedicated Performance & Operations tracks. Use code vel09scf
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@...
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/