Makefile for man pages of xtables extensions

Hello,

I made this patch to include all available matches and
targets to
iptables.8 and ip6tables.8 man pages. The source file is
lixt_*.c but
the man page is libipt_*.man or libip6t.man as expected.

If there are additional matches and targets (extra, as in
POM), the
variable named "extra" is used which was
unescaped.

But I found that in the generated man pages of the svn
version of
modules the title is
  ... (not supported, see Patch-O-Matic)
for instance helper, length, limit etc. Is it intended?


Regards,
Laszlo Attila Toth

Laszlo Attila Toth wrote:
> Hello,
> 
> I made this patch to include all available matches and
targets to
> iptables.8 and ip6tables.8 man pages. The source file
is lixt_*.c but
> the man page is libipt_*.man or libip6t.man as
expected.


Thanks for taking care of this. Mid term we should probably
also
move to a single source file for the xtables manpages, they
*should*
be identical anyways.

> But I found that in the generated man pages of the svn
version of
> modules the title is
>   ... (not supported, see Patch-O-Matic)
> for instance helper, length, limit etc. Is it
intended?


No, matches and targets available in the kernel should not
be
marked as not supported. Care to send another patch to fix
this?

> --- iptables.orig/extensions/Makefile	2007-08-07
10:03:05.000000000 +0200
> +++ iptables/extensions/Makefile	2007-09-03
14:28:44.000000000 +0200
>  -29,6 +29,11 
>  PF6_EXT_MAN_ALL_MATCHES:=$(foreach
T,$(PF6_EXT_ALL_SLIB),$(shell test -f
extensions/libip6t_$(T).man && grep -q
register_match6 extensions/libip6t_$(T).c  && echo
$(T)))
>  PF6_EXT_MAN_ALL_TARGETS:=$(foreach
T,$(PF6_EXT_ALL_SLIB),$(shell test -f
extensions/libip6t_$(T).man && grep -q
register_target6 extensions/libip6t_$(T).c && echo
$(T)))
>  
> +PF_EXT_MAN_ALL_MATCHES+=$(foreach
T,$(PFX_EXT_ALL_SLIB),$(shell test -f
extensions/libipt_$(T).man && grep -q register_match
extensions/libxt_$(T).c  && echo $(T)))
> +PF_EXT_MAN_ALL_TARGETS+=$(foreach
T,$(PFX_EXT_ALL_SLIB),$(shell test -f
extensions/libipt_$(T).man && grep -q
register_target extensions/libxt_$(T).c && echo
$(T)))
> +PF6_EXT_MAN_ALL_MATCHES+=$(foreach
T,$(PFX_EXT_ALL_SLIB),$(shell test -f
extensions/libip6t_$(T).man && grep -q
register_match6 extensions/libxt_$(T).c  && echo
$(T)))
> +PF6_EXT_MAN_ALL_TARGETS+=$(foreach
T,$(PFX_EXT_ALL_SLIB),$(shell test -f
extensions/libip6t_$(T).man && grep -q
register_target6 extensions/libxt_$(T).c && echo
$(T)))


This doesn't work for IPv6 extensions. We use
xtables_register_match
for both IPv4 and IPv6. I've changed this in your patch and
applied it.
Thanks again.

Patrick McHardy írta:
> Laszlo Attila Toth wrote:
>> But I found that in the generated man pages of the
svn version of
>> modules the title is
>>   ... (not supported, see Patch-O-Matic)
>> for instance helper, length, limit etc. Is it
intended?
> 
> 
> No, matches and targets available in the kernel should
not be
> marked as not supported. Care to send another patch to
fix this?
> 

I made a new patch for connrate: it was not optional (not
tested) and
wasn't in the Makefile's PF_EXT_SLIB: line. It is not in the
kernel
version 2.6.23-rc3 (linus.git)


I made the following changes in extensions/Makefile:

 * no extra target/match by default 
 * man page of fix modules (PF_EXT_SLIB etc.) plus optional
  (...SLIB_OPTS)  modules generated, but not all.
 * because of the previous one I had to rename
PF_EXT_SE_SLIB to
   PF_EXT_SELINUX_SLIB etc. as a non-optional variable,
original
   PF_EXT_SE_SLIB gets the value of PF_EXT_SELINUX_SLIB if
DO_SELINUX is
   set to 1.

Regards,
	Attila

Laszlo Attila Toth wrote:
> I made a new patch for connrate: it was not optional
(not tested) and
> wasn't in the Makefile's PF_EXT_SLIB: line. It is not
in the kernel
> version 2.6.23-rc3 (linus.git)


I missed connrate when removing the other unsupported
extensions,
I removed it now.

> I made the following changes in extensions/Makefile:
> 
>  * no extra target/match by default 
>  * man page of fix modules (PF_EXT_SLIB etc.) plus
optional
>   (...SLIB_OPTS)  modules generated, but not all.
>  * because of the previous one I had to rename
PF_EXT_SE_SLIB to
>    PF_EXT_SELINUX_SLIB etc. as a non-optional variable,
original
>    PF_EXT_SE_SLIB gets the value of PF_EXT_SELINUX_SLIB
if DO_SELINUX is
>    set to 1.


Applied, thanks Attila.