ipcop-devel Mailing List for IPCop Firewall

Sorry again.  Forgot the attachment.
Sorry about that.  I've enclosed a tarball of two different runs.

I made a typo, it's the -c option.  MAKEDEV.fail shows that MAKEDEV 
can't handle a microcode device.  I checked the build/etc/makedev.d 
directory, and there's no such animal in it.  I checked the 
/etc/makedev.d directory and there is.  I noticed the call to MAKEDEV 
is correct, but I can't explain how MAKEDEV got to the microcode 
device any other way.  Like I wrote, earlier, I got around this by 
renaming Red Hat's /etc/makedev.d and creating a symlink to IPCop's 
in my first make.sh attempt's build directory.

That got me to the second failure.  GD.fail shows compile errors. 
The first thing I noticed was that the root include and library 
directories were being used.  I kludged the make.sh to change the sed 
that modifies the y/n default responses in Makefile.pl, to also 
change the hardwired -I/usr and -L/usr in the Makefile.pl to point to 
../usr.  Please note, the -m686 in the compiles.

I ran a third attempt, and got pretty much the same errors on the GD 
compile.  I noticed that even though the Makefile.pl showed my sed 
changes, the resulting Makefile did not.  That's what led me to the 
config file out of /usr.

Harry
At 2:07 PM +0100 10/30/02, Mark Wormgoor wrote:
>Hi,
>
>>  I've been trying to install your suggestion for testing.
>>  Unfortunately, I haven't been able to create IPCop 0.1.2 from the CVS
>>  pull on my RH 7.3 system.
>>
>>  I've run into problems with MAKEDEV, it doesn't honor the -d option,
>>  and with GD.pm.
>>
>>  I've kludged around the MAKEDEV problem by renaming /etc/makedev.d
>>  and creating a symlink to the one in the build directory.
>
>How can your MAKEDEV not honor -d or use the makedev.d dir.  From the
>script:
>   $TMPDIR/build/dev/MAKEDEV -m 16 -d $TMPDIR/build/dev -c 
>$BUILDDIR/etc/makedev.d \
>         std console >> $LOGFILE 2>&1
>   $TMPDIR/build/dev/MAKEDEV -m 8 -d $TMPDIR/build/dev -c 
>$BUILDDIR/etc/makedev.d \
>         pty ide usb ipfilter isdn >> $LOGFILE 2>&1
>The MAKEDEV used here comes from the MAKEDEV rpm.  I don't see how your
>distro could influence this.
>
>>  I'm pretty well stumped on how to get GD to compile.  perl is using
>>  the config file out of my /usr... directory tree and not the one in
>>  the build directory.  As a result, GD won't compile and of course the
>>  compile flags specify -mi686.  I've try using sed on the Makefile,
>>  but I am not enough of a perl monger to figure out where the config
>>  file is referenced.  Is another symlink called for.
>
>You're right in that perl will use your config file out of /usr.  But that
>should not be too much of a problem.  The install directories are
>overridden.  Also, the make command overrides the CFLAGS when building GD
>with -march=i386 and -mcpu=i386.  These are both workarounds and I can
>imagine this not working on every machine though.
>   mv Makefile.PL Makefile.PL.bak
>   cat Makefile.PL.bak | \
>         sed "s%,'y'%,'n'%" > Makefile.PL &&
>   yes 'n' | perl Makefile.PL INSTALLDIRS=site \
>         INSTALLSITELIB=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/ \
> 
>INSTALLSITEARCH=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/i386-linux/
>   make CFLAGS="-march=i386 -mcpu=i386" >> $LOGFILE 2>&1
>
>>  What are you guys using for a development machine?  Is there
>>  something basic I'm missing?
>I have used both Mandrake 8.1 and Redhat 7.3 without problems.
>
>It is best if you post relevant parts of your compile logs, so we can
>really see what is going on.

Sorry about that.  I've enclosed a tarball of two different runs.

I made a typo, it's the -c option.  MAKEDEV.fail shows that MAKEDEV 
can't handle a microcode device.  I checked the build/etc/makedev.d 
directory, and there's no such animal in it.  I checked the 
/etc/makedev.d directory and there is.  I noticed the call to MAKEDEV 
is correct, but I can't explain how MAKEDEV got to the microcode 
device any other way.  Like I wrote, earlier, I got around this by 
renaming Red Hat's /etc/makedev.d and creating a symlink to IPCop's 
in my first make.sh attempt's build directory.

That got me to the second failure.  GD.fail shows compile errors. 
The first thing I noticed was that the root include and library 
directories were being used.  I kludged the make.sh to change the sed 
that modifies the y/n default responses in Makefile.pl, to also 
change the hardwired -I/usr and -L/usr in the Makefile.pl to point to 
../usr.  Please note, the -m686 in the compiles.

I ran a third attempt, and got pretty much the same errors on the GD 
compile.  I noticed that even though the Makefile.pl showed my sed 
changes, the resulting Makefile did not.  That's what led me to the 
config file out of /usr.

Harry
At 2:07 PM +0100 10/30/02, Mark Wormgoor wrote:
>Hi,
>
>>  I've been trying to install your suggestion for testing.
>>  Unfortunately, I haven't been able to create IPCop 0.1.2 from the CVS
>>  pull on my RH 7.3 system.
>>
>>  I've run into problems with MAKEDEV, it doesn't honor the -d option,
>>  and with GD.pm.
>>
>>  I've kludged around the MAKEDEV problem by renaming /etc/makedev.d
>>  and creating a symlink to the one in the build directory.
>
>How can your MAKEDEV not honor -d or use the makedev.d dir.  From the
>script:
>   $TMPDIR/build/dev/MAKEDEV -m 16 -d $TMPDIR/build/dev -c 
>$BUILDDIR/etc/makedev.d \
>         std console >> $LOGFILE 2>&1
>   $TMPDIR/build/dev/MAKEDEV -m 8 -d $TMPDIR/build/dev -c 
>$BUILDDIR/etc/makedev.d \
>         pty ide usb ipfilter isdn >> $LOGFILE 2>&1
>The MAKEDEV used here comes from the MAKEDEV rpm.  I don't see how your
>distro could influence this.
>
>>  I'm pretty well stumped on how to get GD to compile.  perl is using
>>  the config file out of my /usr... directory tree and not the one in
>>  the build directory.  As a result, GD won't compile and of course the
>>  compile flags specify -mi686.  I've try using sed on the Makefile,
>>  but I am not enough of a perl monger to figure out where the config
>>  file is referenced.  Is another symlink called for.
>
>You're right in that perl will use your config file out of /usr.  But that
>should not be too much of a problem.  The install directories are
>overridden.  Also, the make command overrides the CFLAGS when building GD
>with -march=i386 and -mcpu=i386.  These are both workarounds and I can
>imagine this not working on every machine though.
>   mv Makefile.PL Makefile.PL.bak
>   cat Makefile.PL.bak | \
>         sed "s%,'y'%,'n'%" > Makefile.PL &&
>   yes 'n' | perl Makefile.PL INSTALLDIRS=site \
>         INSTALLSITELIB=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/ \
> 
>INSTALLSITEARCH=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/i386-linux/
>   make CFLAGS="-march=i386 -mcpu=i386" >> $LOGFILE 2>&1
>
>>  What are you guys using for a development machine?  Is there
>>  something basic I'm missing?
>I have used both Mandrake 8.1 and Redhat 7.3 without problems.
>
>It is best if you post relevant parts of your compile logs, so we can
>really see what is going on.

Hi,

> I've been trying to install your suggestion for testing. 
> Unfortunately, I haven't been able to create IPCop 0.1.2 from the CVS 
> pull on my RH 7.3 system.
> 
> I've run into problems with MAKEDEV, it doesn't honor the -d option, 
> and with GD.pm.
> 
> I've kludged around the MAKEDEV problem by renaming /etc/makedev.d 
> and creating a symlink to the one in the build directory.

How can your MAKEDEV not honor -d or use the makedev.d dir.  From the 
script:
  $TMPDIR/build/dev/MAKEDEV -m 16 -d $TMPDIR/build/dev -c $BUILDDIR/etc/makedev.d \
        std console >> $LOGFILE 2>&1
  $TMPDIR/build/dev/MAKEDEV -m 8 -d $TMPDIR/build/dev -c $BUILDDIR/etc/makedev.d \
        pty ide usb ipfilter isdn >> $LOGFILE 2>&1
The MAKEDEV used here comes from the MAKEDEV rpm.  I don't see how your 
distro could influence this.

> I'm pretty well stumped on how to get GD to compile.  perl is using 
> the config file out of my /usr... directory tree and not the one in 
> the build directory.  As a result, GD won't compile and of course the 
> compile flags specify -mi686.  I've try using sed on the Makefile, 
> but I am not enough of a perl monger to figure out where the config 
> file is referenced.  Is another symlink called for.

You're right in that perl will use your config file out of /usr.  But that 
should not be too much of a problem.  The install directories are 
overridden.  Also, the make command overrides the CFLAGS when building GD 
with -march=i386 and -mcpu=i386.  These are both workarounds and I can 
imagine this not working on every machine though.
  mv Makefile.PL Makefile.PL.bak
  cat Makefile.PL.bak | \
        sed "s%,'y'%,'n'%" > Makefile.PL &&
  yes 'n' | perl Makefile.PL INSTALLDIRS=site \
        INSTALLSITELIB=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/ \
        INSTALLSITEARCH=$TMPDIR/build/usr/lib/perl5/site_perl/5.6.1/i386-linux/
  make CFLAGS="-march=i386 -mcpu=i386" >> $LOGFILE 2>&1

> What are you guys using for a development machine?  Is there 
> something basic I'm missing?
I have used both Mandrake 8.1 and Redhat 7.3 without problems.

It is best if you post relevant parts of your compile logs, so we can 
really see what is going on.

Kind regards,

Mark Wormgoor
-- 
***************************************************************
* |\    /|      |  /|  /       Mark Wormgoor                  *
* | \  / |      | / | /        mailto:ma...@wo...       *
* |  \/  |ark   |/  |/ormgoor  http://www.wormgoor.com/mark/  *
***************************************************************

Mark,

Sorry to bother you.

I've been trying to install your suggestion for testing. 
Unfortunately, I haven't been able to create IPCop 0.1.2 from the CVS 
pull on my RH 7.3 system.

I've run into problems with MAKEDEV, it doesn't honor the -d option, 
and with GD.pm.

I've kludged around the MAKEDEV problem by renaming /etc/makedev.d 
and creating a symlink to the one in the build directory.

I'm pretty well stumped on how to get GD to compile.  perl is using 
the config file out of my /usr... directory tree and not the one in 
the build directory.  As a result, GD won't compile and of course the 
compile flags specify -mi686.  I've try using sed on the Makefile, 
but I am not enough of a perl monger to figure out where the config 
file is referenced.  Is another symlink called for.

What are you guys using for a development machine?  Is there 
something basic I'm missing?

At 10:08 AM +0200 10/24/02, Mark Wormgoor wrote:
>Hi,
>
>>  I just spent an enjoyable few hours debugging a problem getting to an
>>  internal web server on a different node of my VPN.  Basically, it
>>  appears Squid sees the html request, acknowledges the packet back to
>>  the browser and then goes after the page on the Internet, instead of
>>  the VPN.
>
>Hmmm, this suddenly explains some problems I've been having ;)
>Never gave it much thought and tried the internet path, which works just
>fine.
>
>>  I've turned off Squid for now, but is anyone familiar enough with
>>  Squid parameters to tell me how to direct Squid to send VPN packets
>>  through the correct interface or to turn off caching based on server
>>  network address?
>
>The problem is in restartsquid.c:
>/sbin/ipchains -A squid -p tcp --destination-port 80 -j REDIRECT 800
>
>I suggets modifying like this (untested):
>/sbin/ipchains -A squid -p tcp --destination-port 80 -i ! ipsec+ -j 
>REDIRECT 800

Hi Martin,

I'm coordinating the translators for IPCop.

You're on the right track.  The two files you've mentioned are the 
ones needed for the run time.

There are two other portions that need to be translated, too.  These 
are the manuals.  They are in the IPCopDoc sub-directory in CVS.  The 
manuals are in DocBook format.  DocBook allows you to output directly 
to PDF, HTML and HTMLHelp, among many other formats.

Eventually, we will need to write and translate the HTMLHelp files, 
too.  At this point, there are copyright problems with the old 
Smoothwall ones, so we're not distributing them.

The manuals are a big enough project that other translators have 
worked in teams to get them done.

Unfortunately, you've volunteered a few weeks too late for the 0.1.2 
release.  The developers have decided not to add any more languages 
than we already have for this release.  However, there may be a 0.1.3 
release, and there's always 0.2.0.

If you or anyone else is interested in the translation effort, please 
contact me.

Thanks,

Harry Goldschmitt

At 6:24 PM -0300 10/29/02, Mart=EDn Marqu=E9s wrote:
>I would like to contribute with a spanish translation of IPCop, but I'm not
>sure if I'm looking in the right place.
>
>First, I checked 0.1.2pre4 (should I start with snapshots?), and saw in lan=
gs/
>the three languages supported in this version. Now, is it just like getting
>the english copy and translating? I mean, for example:
>
><in en/install/lang_en.c -> es/install/lan_es.c>
>/* TR_CANCEL */
>"Cancel",
>
>would be replaced with
>
>/* TR_CANCEL */
>"Cancelar",
></in>
>
><in en/cg-bin/en.pl -> es/cg-bin/es.pl>
>'back' =3D> 'BACK',
>
>would be replaced with
>
>'back' =3D> 'ATRAS',
></in>
>
>Is this right? And on which version of IPCop should I start looking?
>
>--
>Porqu=E9 usar una base de datos relacional cualquiera,
>si pod=E9s usar PostgreSQL?
>-----------------------------------------------------------------
>Mart=EDn Marqu=E9s                  |        mma...@un...
>Programador, Administrador, DBA |       Centro de Telematica
>                        Universidad Nacional
>                             del Litoral
>-----------------------------------------------------------------
>
>
>
>-------------------------------------------------------
>This sf.net email is sponsored by:ThinkGeek
>Welcome to geek heaven.
>http://thinkgeek.com/sf
>_______________________________________________
>IPCop-devel mailing list
>IPC...@li...
>https://lists.sourceforge.net/lists/listinfo/ipcop-devel

----- Original Message -----
From: "Martín Marqués" <ma...@bu...>
To: <ipc...@li...>
Sent: Tuesday, October 29, 2002 4:24 PM
Subject: [IPCop-devel] translations


> I would like to contribute with a spanish translation of IPCop,
but I'm not
> sure if I'm looking in the right place.
>
> First, I checked 0.1.2pre4 (should I start with snapshots?),
and saw in langs/
> the three languages supported in this version. Now, is it just
like getting
> the english copy and translating? I mean, for example:
>
> <in en/install/lang_en.c -> es/install/lan_es.c>
> /* TR_CANCEL */
> "Cancel",
>
> would be replaced with
>
> /* TR_CANCEL */
> "Cancelar",
> </in>
>
> <in en/cg-bin/en.pl -> es/cg-bin/es.pl>
> 'back' => 'BACK',
>
> would be replaced with
>
> 'back' => 'ATRAS',
> </in>
>
> Is this right? And on which version of IPCop should I start
looking?
>

While I am not the person in charge of this, I can tell you how
it was done on another firewall project with which I helped.

1. The translation team took the en.pl file, and translated that
into their language, and then changed the name to indicate the
new language.  In your case, that would be es.pl.  Also, there
was a file written in C that was the installation program.  That
was translated as well.

2. Then the help files would all be translated into the new
language.  Now, the help files that came from the project that
wqas forked have been removed because it was unclear whether they
fell under the GPL licence or not.  Since then, I have been told
that the help files are going to be written in a completely
different way.

3. Then the FAQs and Configuration files would be translated into
the new language, and then processed into PDF files.

Since I am not the translation manager for this project, I can't
say that the above it how it is being done, but as far as the GUI
interface, that is how it would be done.  It looks like you are
on the right track however.  :)

Rebecca

I would like to contribute with a spanish translation of IPCop, but I'm not 
sure if I'm looking in the right place.

First, I checked 0.1.2pre4 (should I start with snapshots?), and saw in langs/ 
the three languages supported in this version. Now, is it just like getting 
the english copy and translating? I mean, for example:

<in en/install/lang_en.c -> es/install/lan_es.c>
/* TR_CANCEL */
"Cancel",

would be replaced with

/* TR_CANCEL */
"Cancelar",
</in>

<in en/cg-bin/en.pl -> es/cg-bin/es.pl>
'back' => 'BACK',

would be replaced with

'back' => 'ATRAS',
</in>

Is this right? And on which version of IPCop should I start looking?

-- 
Porqué usar una base de datos relacional cualquiera,
si podés usar PostgreSQL?
-----------------------------------------------------------------
Martín Marqués                  |        mma...@un...
Programador, Administrador, DBA |       Centro de Telematica
                       Universidad Nacional
                            del Litoral
-----------------------------------------------------------------

I think having an Add-ons section that works like the Updates one does is a
great idea.

David Martin

-----Original Message-----
From: ipc...@li...
[mailto:ipc...@li...]On Behalf Of Gavin Spurgeon
Sent: Monday, October 28, 2002 2:37 PM
To: ipc...@li...
Subject: Re: [IPCop-devel] Suggestion to help with add-ons


Hi List...

I noticed that I seem to have killed this thread....
I was the last person to reply & then......Nothing...
It went dead....
But....
for those interested.... I have released a new Add-On
(Plz refer 2 my Post on the Devel List
http://marc.theaimsgroup.com/?l=ipcop-devel&m=103583107301847&w=2)
and I think that the technology that we already have in
IPCop is enough to add an Interface for an Add-On
section.....
If you look @ the way I overcame this problem with
my Add-Onz I wanted, I simply butchered the Updates system
Built in to IPCop, and changed it in such a way as to run
with Add-Onz in mind rather than Updates, I still have the
Update system working but now there is a New page called
Add-Ons...
so can we use an Add-Ons system in the same way that
Updates have the 'Info' Link next to them on the Available
list on the Updates page, why not have this same idea but
point the link to the Interface of the Add-On in question ?

Just a suggestion....
what dose the List think ?
Any & all comments welcome...
(Even if to say 'Shut Up' !!)




"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

Hi List...

I noticed that I seem to have killed this thread....
I was the last person to reply & then......Nothing...
It went dead....
But....
for those interested.... I have released a new Add-On
(Plz refer 2 my Post on the Devel List =
http://marc.theaimsgroup.com/?l=3Dipcop-devel&m=3D103583107301847&w=3D2)
and I think that the technology that we already have in
IPCop is enough to add an Interface for an Add-On
section.....
If you look @ the way I overcame this problem with
my Add-Onz I wanted, I simply butchered the Updates system
Built in to IPCop, and changed it in such a way as to run
with Add-Onz in mind rather than Updates, I still have the
Update system working but now there is a New page called
Add-Ons...
so can we use an Add-Ons system in the same way that
Updates have the 'Info' Link next to them on the Available
list on the Updates page, why not have this same idea but
point the link to the Interface of the Add-On in question ?

Just a suggestion....
what dose the List think ?
Any & all comments welcome...
(Even if to say 'Shut Up' !!)


-------------------------------------------------------------------------=
-------

"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

Hi List & Peter...

Just a quick update....
I have done it again !!!!!!!
I have my NEW Add-On ready....
I now I will get a few Bad e-mails from this 1 But here it is.....

I have created a simple system where the IPCop will E-Mail itz
New Red IP address when the Red Interface Comes Up...
Itz very simple and works on the same kind of system as my 'Report'
Add-On, that I made a wile ago....
I have 2 Screen shots on my www site @
http://www.dageek.co.uk/ipcop/addonz/add-on_102_ready.gif
&
http://www.dageek.co.uk/ipcop/addonz/add-on_102_installed.gif

I am going to re-release my Reportz Add-On with a few new trickz
I learned from input from Peter Walker, and am hoping 2 add an
interface to the IPCop Admin Interface to set the E-Mail Address for
each of my Add-Onz that send E-Mail and a few other settingz from
the other Add-Onz I have created for My Costumed IPCopz...
I will also be releasing a few Add-Onz 4 the Modules for thingz like
MSN, H.232 and the other stuff....

If anyone is interested please let me now.... Ill help an any way I =
can....


-------------------------------------------------------------------------=
-------

"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

It's never simple is it?

You need to include the patch binary in your patch file so it is there when setup needs it.
As you are copying a bunch of stuff to the ipcop anyway I don't think this is a problem.

If you really don't want to do this then you need to write a sed or perl script which edits the rc
file to produce the new version - I'd stick with option 1!

Pete.

Gavin Spurgeon said:
> Hi Pete & List
>
> This Looked real good.....
> until I typed patch in an SSH session and got this back....
>
> <Snip>
>  :~ $ patch
>  bash: patch: command not found
> </Snip>
>
> What do I do now......
> This method needz to work on all IPCopz...
> If I am doing something wrong please let me know....
>
>
> --------------------------------------------------------------------------------
>
> "The happiest of people don't necessarily have the best of everything, they just make the most
> of everything that comes along their way.." Gavin Spurgeon.
> AKA Da Geek

Hi Pete & List

<Snip>
 For example:
 1.  Edit rc.updatered and save the new version as rc.updatered.new
 2. diff -c rc.updatered rc.updatered.new > rc.updatered.diff
 3. Copy the rc.updatered.diff file into your patch.
 4. Edit your setup script to call patch < rc.updatered.diff
 man diff/man patch for details.
</Snip>

This Looked real good.....
until I typed patch in an SSH session and got this back....

<Snip>
 :~ $ patch
 bash: patch: command not found
</Snip>

What do I do now......
This method needz to work on all IPCopz...
If I am doing something wrong please let me know....


-------------------------------------------------------------------------=
-------

"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

Lots of ways to do it but what you are probably looking for is patch.

diff reports the differences between 2 files. patch takes an original file and applies the changes
reported by diff to it in order to create the new file.

For example:

1.  Edit rc.updatered and save the new version as rc.updatered.new

2. diff -c rc.updatered rc.updatered.new > rc.updatered.diff

3. Copy the rc.updatered.diff file into your patch.

4. Edit your setup script to call patch < rc.updatered.diff


man diff/man patch for details.

Pete.




Gavin Spurgeon said:
> Hi All.....
>
> I need a bit of help, I have managed to make a 3rd Patch
> for my accustomed Cop Boxes....
> I have come to a point that needs me to add a command
> to a script that is already on the CopBox... So what I need
> to do is Insert a line into the correct place in a existing script... as I understand it the
> 'Diff' command will be able 2 do this (I think)
>
> I am going to be calling the diff command from the file called 'setup' witch is inside the
> patch.tar.gz along with the file 'information' & I need to add a line to the 'rc.updatered' file
> in '/etc/rc.d/' dir
>
> <Snip of rc.updatered>
>
>  /usr/local/bin/setportfw
>  /usr/local/bin/restartsnort
>  /usr/local/bin/setddns.pl
> ###################### add a command here ###################
> else
>  if [ "$DOMAIN_NAME" == "" ]; then
>   /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases
>  else
>   /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases -s "$DOMAIN_NAME"
>  fi
> </Snip of rc.updatered>
>
> so as you can see, I need to add a command @ line 52, but the
> command must be above the 'else' witch by default is on line 53.
>
> can any one shed some light on what I am trying to do ?
> any comments welcome,
> & thx for your time.
>
>
> --------------------------------------------------------------------------------
>
>
> "The happiest of people don't necessarily have the best of everything, they just make the most
> of everything that comes along their way.." Gavin Spurgeon.
> AKA Da Geek

Hi All.....

I need a bit of help, I have managed to make a 3rd Patch
for my accustomed Cop Boxes....
I have come to a point that needs me to add a command
to a script that is already on the CopBox... So what I need
to do is Insert a line into the correct place in a existing script...
as I understand it the 'Diff' command will be able 2 do this (I think)

I am going to be calling the diff command from the file called 'setup'
witch is inside the patch.tar.gz along with the file 'information'
& I need to add a line to the 'rc.updatered' file in '/etc/rc.d/' dir

<Snip of rc.updatered>

 /usr/local/bin/setportfw
 /usr/local/bin/restartsnort
 /usr/local/bin/setddns.pl
###################### add a command here ###################
else
 if [ "$DOMAIN_NAME" == "" ]; then
  /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases
 else 
  /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases -s "$DOMAIN_NAME"
 fi
</Snip of rc.updatered>

so as you can see, I need to add a command @ line 52, but the 
command must be above the 'else' witch by default is on line 53.

can any one shed some light on what I am trying to do ?
any comments welcome,
& thx for your time.


--------------------------------------------------------------------------------


"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

Hi All.....

I need a bit of help, I have managed to make a 3rd Patch
for my accustomed Cop Boxes....
I have come to a point that needs me to add a command
to a script that is already on the CopBox... So what I need
to do is Insert a line into the correct place in a existing script...
as I understand it the 'Diff' command will be able 2 do this (I think)

I am going to be calling the diff command from the file called 'setup'
witch is inside the patch.tar.gz along with the file 'information'
& I need to add a line to the 'rc.updatered' file in '/etc/rc.d/' dir

<Snip of rc.updatered>

 /usr/local/bin/setportfw
 /usr/local/bin/restartsnort
 /usr/local/bin/setddns.pl
###################### add a command here ###################
else
 if [ "$DOMAIN_NAME" == "" ]; then
  /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases
 else 
  /usr/local/bin/dnsmasq -l /var/lib/dhcp/dhcpd.leases -s "$DOMAIN_NAME"
 fi
</Snip of rc.updatered>

so as you can see, I need to add a command @ line 52, but the 
command must be above the 'else' witch by default is on line 53.

can any one shed some light on what I am trying to do ?
any comments welcome,
& thx for your time.


--------------------------------------------------------------------------------


"The happiest of people don't necessarily have the best of everything,
they just make the most of everything that comes along their way.."
Gavin Spurgeon.
AKA Da Geek

For anybody who needs this, and would like to try it out, Tim Dobbelaere has
produced patched versions of /usr/local/bin/updatelists.pl and
/home/httpd/cgi-bin/updates.cgi which allow the available updates file to be
downloaded from behind a proxy.

Download the patched files from here:

http://tim.dobbelaere.com/software/ipcop.php

These patches have been incorporated into CVS, but I think it will only work
with v0.1.1 at the moment, as there isn't an 'available update' file yet for
v0.1.2

Eric

morgad wrote:
> hi,
> 
>  If you have not already done so , could you consider
> appending one (or more) ^G's to the /etc/issue file
> so the box beeps when it has finished booting ?
> 
> (means I don't have switch the monitor on to track
> boot progress)

Dave (and others), go place your wishes at:

http://www.ipcop.org/cgi-bin/twiki/view/IPCop/IPCopWishList

there's no guarantee that anybody will do anything more than look at 
them and go "uh-huh" but at least we can get them recorded.

---eric

hi,

 If you have not already done so , could you consider
appending one (or more) ^G's to the /etc/issue file
so the box beeps when it has finished booting ?

(means I don't have switch the monitor on to track
boot progress)

best regards
Dave Morgan

Harry Goldschmitt wrote:
> Mark gave me a suggestion on fixing my Squid problem, last night. While 
> trying to fix it I did a pull from CVS and tried to do a straight build 
> first.  It looks like MAKEDEV is either ignoring the -d option or 
> somehow make.sh is deciding I'm running on Debian, (no I don't have an 
> /etc/debian-version).
> 
> Has anyone else had this problem?  I know I can kludge my way around 
> this by substituting /etc/makedev.d directories in my real /etc.

I think the current state of CVS is a little unstable.  It should be 
fixed relatively soon.

by the way, anyone fixing bugs, let's have them done by Sunday if at all 
possible.  I would like to release the next image before the end of the 
month.

---eric

Mark gave me a suggestion on fixing my Squid problem, last night. 
While trying to fix it I did a pull from CVS and tried to do a 
straight build first.  It looks like MAKEDEV is either ignoring the 
-d option or somehow make.sh is deciding I'm running on Debian, (no I 
don't have an /etc/debian-version).

Has anyone else had this problem?  I know I can kludge my way around 
this by substituting /etc/makedev.d directories in my real /etc.

Harry

The latest version of GD 2.04 stable has been released, and according to the author it corrects alot of bugs.

Have any of the development team compiled IPCop with it yet?

Darren.

Hi,

> I just spent an enjoyable few hours debugging a problem getting to an 
> internal web server on a different node of my VPN.  Basically, it 
> appears Squid sees the html request, acknowledges the packet back to 
> the browser and then goes after the page on the Internet, instead of 
> the VPN.

Hmmm, this suddenly explains some problems I've been having ;)
Never gave it much thought and tried the internet path, which works just 
fine.

> I've turned off Squid for now, but is anyone familiar enough with 
> Squid parameters to tell me how to direct Squid to send VPN packets 
> through the correct interface or to turn off caching based on server 
> network address?

The problem is in restartsquid.c:
/sbin/ipchains -A squid -p tcp --destination-port 80 -j REDIRECT 800

I suggets modifying like this (untested):
/sbin/ipchains -A squid -p tcp --destination-port 80 -i ! ipsec+ -j REDIRECT 800

Kind regards,

Mark Wormgoor
-- 
***************************************************************
* |\    /|      |  /|  /       Mark Wormgoor                  *
* | \  / |      | / | /        mailto:ma...@wo...       *
* |  \/  |ark   |/  |/ormgoor  http://www.wormgoor.com/mark/  *
***************************************************************

Hi,

I compiled ipcop-0.1.2pre4 with no errors, but when makegraphs runs I =
get the following error "undefined subroutine &DG::FONT::small called at =
lib/perl5/...
any Idea what went wrong

Regards

Emad

At 12:08 AM +0100 10/24/02, Dave Roberts wrote:
>  > From: Harry Goldschmitt Sent: 23 October 2002 23:12
>>  To: ipc...@li...;
>>  ipc...@li...
>>  Subject: [IPCop-devel] Squid Gets In The Way
>>
>>
>>  Hi,
>>
>>  I just spent an enjoyable few hours debugging a problem getting to an
>>  internal web server on a different node of my VPN.  Basically, it
>>  appears Squid sees the html request, acknowledges the packet back to
>>  the browser and then goes after the page on the Internet, instead of
>>  the VPN.
>>
>>  I've turned off Squid for now, but is anyone familiar enough with
>>  Squid parameters to tell me how to direct Squid to send VPN packets
>>  through the correct interface or to turn off caching based on server
>>  network address?
>>
>
>Let me get this right.  You are trying to access an internal web server
>at site2 from site1, with site1 and site2 connected via a VPN?
>
>Turning Squid off will make this work because the connection will then
>be directly from the client machine to the web server.  Squid does not
>work because there is no connectivity from the firewall to the hosts on
>the remote VPN network.
>
>The way Frees/WAN is setup is with a network to network VPN _not_ a
>gateway to gateway VPN.  The effect of this is that you cannot access
>remote VPN hosts from the firewall - you can test this by SSHing to your
>local IPCop and trying to ping one of the remote VPN hosts - you will
>get no reply even though the VPN is up and local machines can ping
>successfully. See the Frees/WAN website for more info - some of my
>terminology is probably wrong.
>
>There should be a way of having a gateway to gateway VPN alongside or
>instead of the network to network VPN.  I did try to set it up once but
>life prevailed.  This would be useful because we could implement some
>VPN monitoring on the IPCops and possibly automatically restart them if
>they die or if a dynamic IP address changes.

Actually, there is a way to get from an IPCop gateway.  You have to 
direct the traffic out through your green address.  Try:

ping -I GREEN_ADDRESS remote_VPN_host

For example, my IPCop's green address is 192.168.1.1 and a host on 
another leg of the VPN is 192.168.33.6.  Just entering "ping 
192.168.33.6" doesn't work, but
entering "ping -I 192.168.1.1 192.168.33.6" works just fine.

If you need to implement the VPN monitoring look at dynchk.  It's 
been checked into CVS, and will be distributed as part of 0.1.2, but 
no one has had the time to set up a cgi-bin to use it.  I've putting 
in cron jobs to monitor all the legs of my VPN, since most have 
dynamic IP addresses.

Harry

> From: Harry Goldschmitt Sent: 23 October 2002 23:12
> To: ipc...@li...;=20
> ipc...@li...
> Subject: [IPCop-devel] Squid Gets In The Way
>=20
>=20
> Hi,
>=20
> I just spent an enjoyable few hours debugging a problem getting to an=20
> internal web server on a different node of my VPN.  Basically, it=20
> appears Squid sees the html request, acknowledges the packet back to=20
> the browser and then goes after the page on the Internet, instead of=20
> the VPN.
>=20
> I've turned off Squid for now, but is anyone familiar enough with=20
> Squid parameters to tell me how to direct Squid to send VPN packets=20
> through the correct interface or to turn off caching based on server=20
> network address?
>=20

Let me get this right.  You are trying to access an internal web server
at site2 from site1, with site1 and site2 connected via a VPN?

Turning Squid off will make this work because the connection will then
be directly from the client machine to the web server.  Squid does not
work because there is no connectivity from the firewall to the hosts on
the remote VPN network.

The way Frees/WAN is setup is with a network to network VPN _not_ a
gateway to gateway VPN.  The effect of this is that you cannot access
remote VPN hosts from the firewall - you can test this by SSHing to your
local IPCop and trying to ping one of the remote VPN hosts - you will
get no reply even though the VPN is up and local machines can ping
successfully. See the Frees/WAN website for more info - some of my
terminology is probably wrong.

There should be a way of having a gateway to gateway VPN alongside or
instead of the network to network VPN.  I did try to set it up once but
life prevailed.  This would be useful because we could implement some
VPN monitoring on the IPCops and possibly automatically restart them if
they die or if a dynamic IP address changes.

Dave