Front page | perl.perl5.porters |
Postings from September 2011
<URL: https://rt.cpan.org/Ticket/Display.html?id=69318 >
On 2011.9.26 3:45 AM, Christian Walde via RT wrote:
> As schmorp rightly says, that's actually a bit stupid. Due to EUMM
> emulating the compiler actions, there's potential for both false
> positives (when eumm can find the lib, but compiler can't) and false
> negatives (when eumm can't find the lib and removes it from the list,
> but the compiler could).
>
> So, the optimum way of handling this would be to actually use the
> compiler itself to verify the lib list, but i personally have no idea
> how that could be done.
I think the general way to accomplish that is to just try to compile a small
program. Isn't that what Configure does?
I wonder if delegating to ExtUtils::CBuilder would be sufficient. I'd love to
see all that nasty Liblist code go away.
Alternatively... do we need to go through that process at all? Why is
MakeMaker taking responsibility for pruning the library list? Why not just
let the compiler/linker handle it? I suspect the answer might have originally
been "to produce a friendlier error message" but it seems to me not worth all
the trouble and bugs.
Christian, I think of anyone you know Liblist the most. Or if nothing else
you're brave enough to mess with it. What do you think?
> Meanwhile, an option is: Debug through Liblist, figure out why it can't
> find that lib, and change the code so it can find it. Or maybe change
> it so it just warns, but adds the directive anyhow.
That seems to be the usual way.
--
We do what we must because we can.
For the good of all of us,
Except the ones who are dead.
-- Jonathan Coulton, "Still Alive"
Thread Next