rpms/perl-Net-Server/devel .cvsignore, 1.3, 1.4 perl-Net-Server.spec, 1.9, 1.10 sources, 1.3, 1.4

[Nicolas Mailhot]

Le dimanche 08 janvier 2006 à 14:30 +0000, Paul Howarth a écrit :
> On Sun, 2006-01-08 at 07:52 -0500, Nicolas Mailhot wrote:
> > +# IO::Multiplex support is optional, but since it's in FE and not including it 
> > +# causes build problems in some packages...
> > +Requires:      perl(IO::Multiplex)
> 
> If perl(IO::Multiplex) is optional for perl-Net-Server but required for
> some other package, why isn't the fix to add perl(IO::Multiplex) as a
> buildreq/req for the other package? That way, anyone using
> perl-Net-Server but not needing perl(IO::Multiplex) doesn't need to
> install it.

Actually I don't care one way or another, my (feeble) understanding is
perl-Net-Server will enable some code if IO::Multiplex is present.

Some packages clearly depend on IO::Multiplex and won't build if it's
not there -> here a BR is needed as you pointed

Others only depend directly on Net::Server. What happens if
IO::Multiplex is present ? I think the multiplex code will get enabled
too so :
1. they may get some benefit
2. more importantly the package behaviour won't depend on system state
not traced in deps (ie if multiplex breaks them we'll know at once not
when some other package installs it)

But I'll freely admit the main reason I added it is because several
people asked me to, and I'm a nice guy.

If people on the list (or the perl list) feel strongly one way or
another just say so and I'll follow the common view. As I wrote - I
don't care strongly about it.

Regards,

-- 
Nicolas Mailhot
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: Ceci est une partie de message num?riquement sign?e
URL: <http://listman.redhat.com/archives/fedora-extras-list/attachments/20060108/b56d5344/attachment.sig>