Re: [Bug-gnu-radius] missing .lo files

[Maurice Makaay]

Hi Sergey,

> Wow, that's completely impossible to build a so file with a command like
> that! Please cd to client/pam_radius and try the following:
> 
>    gcc -shared  pam_radius.lo -Wl,--whole-archive \
>        ../../lib/.libs/librad.a ../../lib/.libs/libdebug.a \
>        -Wl,--no-whole-archive -lpam -lcrypt \
>        -Wl,-Bsymbolic -Wl,-soname -Wl,pam_radius.so
>        -o .libs/pam_radius.so -lc
> 
> If it works, I'll try to force libtool behave itself.       

I am able to create an .so file using:

        gcc -shared ../../lib/.libs/librad.a ../../lib/.libs/libdebug.a \
            -lpam -lcrypt -Wl,-Bsymbolic -o .libs/pam_radius.so -lc

Things like -Wl,--no-whole-archive and -Wl,-soname will not work for me,
because those are not options that are supported by the Solaris linker
program. And should gcc use "pam_radius.lo" in your command line?
       
> > * CVS-patch-dropprivileges      run-user patch
> 
> I am currently implementing the similar feature. I'll let you know when
> it is ready.

Great! 
 
> > * CVS-patch-realms_nocase       realms are case-insensitive in the
> > * realms
> 
> This feature will be added soon (the approach taken in the patch is
> not quite right, I'll avoid adding any new members to the REALM structure).

I have been thinking of the right way to do this. For me, expanding the
REALM structure seemed like the best idea, because using this method the
lookups for case-insensitivity would be very fast (the information that is
needed is right within grasp and there would be no need to let the
radiusserver walk through all realm options on each lookup). That's the
reason I choose to implement it the way I did, without using the 
envar_lookup_int() call. But any implementation will do for me. It will keep
about 10.000+ of my users very happy ;-D (this is the number of users that 
use upper case characters in their realm).

> > * CVS-patch-sqlconfig           sqlserver config reading improvements
> >                                 (and SEGFAULT prevention on config errors)
> 
> My apologies, this one has somehow skipped my attention. I'll take a
> look.

Alright, thank you.


Regards,

-- Maurice Makaay