Re: [help-GIFT] cvs gift requires libdl

[Pruet Boonma]

Hi,

Sorry, it doesn't work. I found this error from configure

...
checking if we can embed a Perl interpreter into GIFT... /usr/bin/ld:
cannot find -lperl
collect2: ld returned 1 exit status
./configure: -ldl: command not found
no   
...

Anyway, from your clue, I check the Debian package and found that we
(Debian user) need to install libperl-dev package, which is a optional.
This package is required for perl embeded development, it provides two header
files and a static library required by GIFT. So I thought that this problem is
solved ?

Thanks for your help,
Pruet


On Fri, Jan 11, 2002 at 12:00:39PM +0100, Wolfgang Mueller wrote:
> > I tried to compile gift from latest cvs (11/01/02, 03:00) and found some
> > error about dlopen, dlsym, etc. So I fixed it using the patch attached
> > with this mail. I'm not sure that the problem is in my system or
> > somewhere else. My system configuration is at the end of this mail.
> 
> Hi, Pruet,
> Welcome in this list, and thanks a lot for your contribution. 
> 
> > LIBS= -ldl
> 
> That's a short but useful patch. As explanation: this is only needed, if you 
> cannot embed perl into the GIFT (for there is either no perl, or Perl is 
> broken). This is why all you other guys (myself included) don't have any 
> problem, as SuSE and RedHat provide a reasonable default Perl installation. 
> Apparently the Debian standard Perl installation is still broken (with 
> respect to ExtUtils::Embed), or am I wrong, Pruet?
> 
> I guess I will add the corresponding line rather in configure.in to avoid 
> double inclusion of the library when Perl is present.
> 
> Thanks a lot, and
> Cheers,
> Wolfgang
> 
> (BTW along with the new configure.in I will check in my work towards 
> something that compiles with gcc-3.0.3. It does not yet compile, the current 
> problem is that for the streampos type the postincrement (x++) is not 
> defined, so there's the choice between doing things with x=x+1 and checking 
> that things still work, or defining a postincrement)