develooper Front page | perl.perl5.porters | Postings from September 2011

Re: [rt.cpan.org #69318] EU:MM rejects necessary link library

Thread Previous | Thread Next
From:
Michael G Schwern via RT
Date:
September 29, 2011 01:56
Subject:
Re: [rt.cpan.org #69318] EU:MM rejects necessary link library
Message ID:
rt-3.8.HEAD-3049-1317267032-1199.69318-6-0@rt.cpan.org
<URL: https://rt.cpan.org/Ticket/Display.html?id=69318 >

On 2011.9.28 4:04 AM, Nicholas Clark via RT wrote:
> On Wed, Sep 28, 2011 at 05:56:30AM -0400, Christian Walde via RT wrote:
>> Good question! Why indeed. I personally have no clue why that is 
>> necessary. I can only guess at it being meant to make it possible for 
>> authors to indicate multiple libraries with EUMM choosing the one that 
>> actually exists.
>>
>> But that's a wild guess.
> 
> Which corresponds to the actual documentation:
> 
> =item LIBS
> 
> An anonymous array of alternative library
> specifications to be searched for (in order) until
> at least one library is found. E.g.
> 
>   'LIBS' => ["-lgdbm", "-ldbm -lfoo", "-L/path -ldbm.nfs"]

I'm going to propose that only a very, very small subset of modules need that
functionality.

I'm also going to propose that it would be far simpler and more robust to
compile a small program with each flag.  Somebody that knows Configure could
help with this, maybe Merijn could help.  ExtUtils::CBuilder would come into
play here to do the compiling and linking.

If we go down that route, I'd propose leaving LIBS as a legacy interface for
backwards compatibility.  It will still search for working libraries, but it
would use CBuilder instead of Liblist (a goal being to delete Liblist).  We'd
then create a second key for "just use these libraries" and perhaps provide a
method which can do the searching.



Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About